Click Insights: Web and Social Media Activation

Game Monetization: Recognize and Reward Your Most Valuable Players

2012-01-12T14:44:00.000-08:00

(First installment a series of player LTV and game monetization articles. The following is an excerpt from Chapter Four of 'Game Loyalty', an upcoming book by the founders of Core Analytics, LLC)

All players were not created equal. This holds true across the entire gaming industry. Video gamers look and act differently from social gamers. Collection-type players of genres of casual games (e.g. Farmville) look different from those who gravitate more frequently to skill-based games (e.g. Angry Birds). Hard-core FPS gamers have different demographics and motivations from regular casino patrons. The gamers around the blackjack table are not identical to the gamers playing the slot machines. There is not only a great variance around demographic and behavioral characteristics of gamers, but also a wide variance in both player profitability and player Lifetime Value (LTV).

From a game developer’s perspective, understanding who is playing and how these players are interacting with the game and with others is the foundation for data-driven decisions about patches, new characters, enhancements, new features, virtual good offerings, pricing, timing of updates, prioritization of bug-fixes, localized version differences, platform, new game development strategy and so on. From a game marketer’s perspective, advertising and creative decisions can also be optimized at a player-micro segment level. This holds reasonably true across the entire game industry. Both video and social game developers can benefit by understanding who their most valuable player segments are, how they are generating revenue (subscription, virtual good consumption, in-game advertising response), and how they are playing the game. In this way developers can optimize game experience for valuable players and even differentiate game versions based on player segment type, optimize virtual good design and introduction, as well as serve up the most relevant advertising for that player type.

In terms of social game monetization and measurement, in-game progress is generally tied to a player's personal profile. One difference with social games is how players interact with their real world networks through social networks, in contrast to traditional video game environments which remain largely virtual (players do meet at game conferences and in real life, and a great many relationships and marriages are generated from initial virtual video game relationships. The comparative longevity of these relationships is discussion for another chapter.) The basic LTV modeling approach for video games and social games is necessarily different, given that currently video games operate on a subscription model, and many social games are based on a freemium model where revenues stem from virtual good purchases and response to advertising. The video game industry is slowly moving towards a more social-game based monetization model, and LTV calculation approaches will likely start to converge across the game space.

How then, to best measure player LTV? Core Analytics (TM) and Game Loyalty have incorporated player LTV metrics into the BrandMeter (TM) Game Module. These metrics have been developed and selected through a rigorous literature review of LTV calculations, with the end result incorporating the best of these approaches for game type, genre and markets. Below is an abbreviated literature review of academic journal articles on the LTV measurement topic. The section following discusses the advantages of each approach and attempts to cherry-pick the best characteristics of traditional LTV calculation approach and translates the outcome according to the nuances specific to the video and social game industry.

Customer LTV has been defined as the ‘total net income expected from customers’. In two related 2002 and 2003 Data Mining and Knowledge Discovery and SIGKDD papers, Rosset and his group of researchers attempted to apply LTV modeling in a telecommunications scenario. They look at LTV as a function of Value over time, Tenure (Length of Service) and a Discounting Factor. This is broken down into three separate models and then combined for an overall LTV measure. This group focuses primarily on modeling the Tenure (expected length of relationship of customer with brand) and experiment with a survival function and hazard function, both of which are estimated from current and historical data. The discounting factor is typically calculated through either an Exponential Decay or Threshold Function approach, and value is constructed with other traditional methods.

Rosset explains that the spectrum of modeling approaches towards tenure (LOS) is wide. A simple method observing an average churn rate of 5% and then setting s(t)=.95’ is functionally as appropriate as a more complex model incorporating individual characteristics, contract differences and other individual and life-event based factors into a Proportional Hazards model, where hundreds of variables are used as predictors. The danger of the latter is either a lack of data to support the method, or missing out on local, current present effects by building based on older data. The approach used in the paper uses only one month of data. Since pure parametric approaches tend to result in non-smooth functions spiking around contract renewal dates, semi-parametric approaches, such as the Cox Proportional Hazard model, seem to be more flexible. Another option would be a neural net where each tenure output has its own node, and tenure is indicated at a monthly or categorical level. In certain data sets, this neural net approach has outperformed the proportional hazard models. It’s also possible to approach this with a survival analysis or standard supervised learning model where tenure is a predictor and churn is the response. The authors suggest that the Kaplan-Meier approach that requires left-censored data is also appropriate but there may not be enough churn events in recent data to reflect current trends and for this reason non parametric estimates of hazard rates are calculated for tenure modeling. In terms of value, this can be done with forecasting, trend analysis and time series, and current average value can also be used at a segment level.

Another 2004 article from the Journal of Managerial Issues by Pfeifer and his co-authors, approached the customer LTV problem by differentiating between customer lifetime value and customer profitability. They assume that customer LTV is important as it is a primary determinant of customer acquisition spending strategy. Pfeifer and team argue that the two terms are not interchangeable, and that confusion exists in the industry when it comes to customer value and customer profitability measures. Most agree that value is a discounted measure of future dollar amounts associated with the customer. Where a more standard definition of customer Lifetime Value revolves around a standard time value of money measure using discounting and traditional present value technique, Pfeifer suggests a finance approach, focusing on customer ‘valuation’ by determining net present value.

January Mountain View Analytics Meetup: Trends in Gamification and Player Personality

2012-01-10T15:17:00.000-08:00

Today the blog is being updated from Mountain View's Red Rock Coffee, which was described by the group in front of me in line as 'ground zero for Mountain View', and a fellow Bay Area Meetup member gave me a tip yesterday about the second floor being where VC's and entrepreneurs frequently meet and sign deals. I can attest that the Ethiopia Yirgacheffe Biloya is outstanding.

Last night the Bay Area Web, Game and Social Media Analytics Group had it's January 2012 kickoff meeting at Cascal last night in Mountain View. There was a small but good turn out, and gaming and gamification was a popular discussion topic among the attendees. One of the members heads up the global gamification efforts for SAP, and described his efforts identifying over 70 separate gamification-related efforts enterprise-wide. He was able to identify each of these through Jive, SAP's community and collaboration platform. One of the largest gamification initiatives at SAP, which has actually been in place for years, is the SAP community network forum. The site uses common gamification techniques such as listing top contributors in a leader board format, and displaying updated thread and message counts. Discussions around gamification involved debate about the usefulness of badges and leader boards, which tend to focus on player/consumer extrinsic rewards rather that intrinsic rewards such as desired information and social connectivity and recognition. With social games, which are increasingly played by consumers and friends and family on Facebook and other social networks, the potential reward of being teased or praised interactively in relation to an in-game achievement provides more intrinsic value and delight than receiving a badge. The site PlayerDNA identifies gamer personality types, and this type of indentification can be useful from a game development perspective as well, to help identify personality of visitors on web sites and in consumer bases, to facilitate relevant game design and relevant virtual good and advertisement offerings. About 1 percent of players/consumers fall into the 'killer' category, those players that derive enjoyment from disrupting others achievement and heckling other players, about 60% fall into the 'socializer' primary player type, and others into 'achiever' and 'explorer'. Each player gets a percentage make up across the four player personality types to understand individual makeup. Players take quiz using their player names, offering a unique opportunity for game companies to tie player DNA to game play data where player names match to playerbase names.

Other members are founders of the recipe discovery site Knapkins, which incorporates social media and gamification user engagement approach into a 'Which looks Yummier?' process where visitors vote on the desert or dish with the most appeal and are presented with an option to access the recipes and share their opinion on Twitter and Facebook.

Game Loyalty and Core Analytics were in attendance, discussing their BrandMeter (TM) offering for games and metrics around player and customer loyalty and LTV for social and games.

Around the Venture Capital topic, the following resources for entrepreneurs were shared: VCTaskForce is a venture community aimed at maximizing communication between the investor community and portfolio companies, The Churchill Club organizes 30-40 events a year aimed at encouraging economic growth and innovation in Silicon Valley, Web 2.0 Summit, and TechCrunch events.

Overview of Angry Birds Monetization and Brand Strategy from Casual Connect Kyiv 2011

2011-12-27T05:01:00.000-08:00

Casual Connect Kyiv was held for it's 6th consecutive year in October 2011. This is an international game conference aimed at covering cutting edge topics related to the social game industry. Many of the conference sessions are available as videos on the site, it's a fantastic resource for talks delivered by industry leaders on current trends related to platforms and monetization, mobile and social mobile trends, and production and design.

One hour-long session in particular was very interesting, VP of Franchise Development for Rovio Mobile Ville Heijari, gave a presentation covering Rovio's approach to Angry Birds branding and monetization strategy. Below is a summary of what was shared.

Popular projections target 1 billion Angry Birds players by 2012. Besides a staggering player-base, what real implications does this have for Rovio Mobile? Ville explained his view on various types of players and stages they pass through. First there are people who have heard of the game and may play it, and discuss it. Then there are players who pay money. The 3rd level of players are the engaged players who play regularly and spread the word. Finally, there is the core player base who love the game. The objective is to move the existing players through the various levels until they reach that final, loyal Angry Birds player level. Then there is another objective of expanding reach, finding and engaging new markets.

Ville explained that Angry Birds took 8 months to build and cost about 100k euro, about twice the cost and development time of other casual games. Rovio released the first Angry Birds as a game developer through a publisher, and later self-published Angry Birds Rio and Angry Birds Seasons as a multi-media entertainment company. Initially there was a focus on mobile, now the objective is to reach as many players as possible across various platforms. Angry Birds was initially developed and published for iOS, then released for Android, and now is available for free for Google Chrome users on a PC. The natural progression from here is likely to release console versions and make the games available through social networks.

The focus has been less on increasing in-game monetezation and engagement, than on creating the most fun game in the world. Players can purchase t-shirts and other merchandise (plush toys on Amazon selling for between $45 and $159 that make up 10 to 20% of business and are selling at a rate of about 1 million toys a month), and there is an in-game virtual good called the mighty eagle, that was introduced over a year ago that involves a can of sardines that is used as bait to summon the mighty eagle that arrives and smashes everything to pieces. This virtual good is not only useful when stuck on a certain level, but it also unlocks new game elements across levels, thereby enhancing the overall game experience for the purchaser. Part of the success of Angry Birds has been a local grassroots marketing approach driving towards virility. Rovio first focused on making the game popular at home in Finland through word-of-mouth, then did the same in Sweden, then Denmark, and then other European countries. With demonstrable and increasing popularity they were able to have their game featured in the UK app store, which led to record growth, and then popularity expanded to the US. The game captures and holds player interest with frequent updates: new levels that are sold in batches and that are provided at no additional cost to players, new characters, new related media and products. For example, one update is the Moon Festival update in Angry Birds Seasons, designed to appeal to global markets, and very popular in China. China, incidentally, is now the fastest growing Angry Birds market, and has Angry Birds retail stores and are represented in a theme park. As of March 2011 the game had made $70 million, which is not unusual for popular casual games and low in comparison to successful video game revenues. A collaborative approach with companies like T-mobile (Angry Birds Live in Barcelona), Nokia, and other large brands has allowed Rovio to benefit from large brand production budgets. They have released a cookbook, with the intention of promoting the practice of parents and grandparents playing the game on mobile devices together, and incorporating the branded theme during another family-gathering events like meals.

Interestingly, this last example was combined with the comment that while of course there is a lot of focus on analytics and metrics, it's not entirely clear who is playing the game. The theory that kids may be playing on their parents mobile phones has been anecdotally supported. One area where semantic technology can be useful to better understand player demographics (age, gender, race, location) is to mine Twitter and Facebook data to identify player characteristics. For instance, Mitre Foundation researchers have identified a reliable method for determining gender of a tweeter based on the 140 characters. Identitext is a Chicago and Israel-based company that mines these type of demographic variables on Tweeters. Core Analytics (and subsidiary Game Loyalty) is a Chicago and Bay-Area based firm that identifies both demographic and behavioral variables from tweets, Facebook and forum posts. AppData provides game- level comparison stats and their premium product appends estimates for age and gender of players. Following the recent trend towards increased emphasis on the applications of player analytics, Angry Birds tapped Seattle-based Medio Systems to take their analytics to the next level with big data and predictive analytics.

Game Analytics: Platform Trends, Monetization and Player Value in Social Games

2011-12-13T20:10:00.000-08:00

The Game Analytics industry is arguably still in it's infancy. A list of niche analytics vendors for social and mobile games continues to expand, with representation by Kontagent, Flurry, Mixpanel, TOTANGO, Claritics and Google Analytics. There are far fewer vendors focusing on the computer and MMO games, and no single analytics provider appears to focus on delivering cross-game platform analytics. It's a daunting task to enter a new industry and attempt to put together a compilation of cross-channel player behavioral analytics research entitled 'Game Loyalty'. The kind of thing only fools rush into, but after taking the plunge into total game industry immersion there's really no going back. The industry is ripe with potential, fascinating from a consumer behavior and emerging brand marketing channel perspective, infinite in terms of data elements to be harnessed by new technology to store and compute the big data, and very promising from a profitability and growth standpoint.

Where to start, when tackling something like measuring player activation and loyalty? There's really no good or bad place to start. Core Analytics happened upon the scene back in late 2010, after delivering our first tech pitch that received somewhat lukewarm press, however soon proved worth the effort when we were contacted by a local game development group about the social media BrandMeter(TM) analytics development as it might relate to mobile games. We did some rapid market research and soon became enchanted with the space, and Game Loyalty and development on the BrandMeter(TM) game module followed thereafter.

There are a certain short list of metrics that are considered standard when measuring social games. Engagement is an indicator of time spent playing a game, DAU is daily active users, and MAU gets slightly less emphasis representing monthly active users, the combined DAU/MAU ratio is popular, K factor is an infection rate measure of viral growth (also the least expensive avenue to increasing core and casual player base), ARPU reflects the average revenue per player. Social games make money in two ways: Virtual good sales and advertising. Lifetime Network Value is another metric that attempts to capture a players total contribution or value to the game, looking at purchases, influence on virality, and net game evangelism.

The social game market is estimated at $1 billion for 2011. To put this into perspective, video games are expected to bring in $74 billion in this same year. Social game developers are viewed a bit as interlopers in the overall video game scene. While the numbers would appear reassuring to the video game industry, social games are offered at a fraction of the price on mobile and social platforms such as Facebook, and allow players to play with friends and family. They represent a fundamental change in the gaming landscape, expanding who, how and why gamers are playing. The console game industry is declining rapidly, as web-based games become the go-to gaming platform.

There are many approaches to traditional measures of customer LTV in offline marketing and digital marketing departments. Here's one that provides a great top-level approach, honing in on a customer LTV metric that is relevant for specific business purposes. Player LTV metrics would benefit from this approach as well. Why is Player Value important?

From a game development perspective, identifying those players most likely to purchase virtual goods, click on ads and convert, champion the game in social forums, and become early adopters of new titles and cross from niche game genres to mainstream are important in terms of designing a game experience most likely to result in longer playtimes, purchases to speed the level up process, promote players to fan and promote the game.

From a marketing perspective it's possible to mine the comment data and identify preferences and consumption patterns and product/brand preferences outside of the game, to target market for in game advertising and branded virtual goods. Modeling Player LTV along the lines of audience and intent with these metrics will refine the measures. A developmental Player LTV will emphasize aspects of player behavior relative to game design, whereas a marketing Player LTV will identify propensities to purchase or suggest optimal treatment strategies. Core Analytics and Game Loyalty have developed a real-time customer brand activation and loyalty process that identifies 9 different types of loyalties, categorized and filterable by business objective. This Player LTV approach might best be approached from a segmentation and business objective approach as well, resulting in a series of objective specific Player LTV metrics.

Game Loyalty: Player Motivations, Player Insights and Game Analytics

2011-12-11T06:21:00.000-08:00

In her fabulous book 'Reality is Broken', Jane McGonigal lists some staggering numbers associated with gaming, such as 183 million active gamers in the US, 100 million in Europe, and 200 million in China. A recent Gartner Report forecasts 2011 video gaming and related hardware revenue at around $74 billion and predicts this explosive growth will continue to reach $115 billion by 2015, with an increasing momentum from the mobile game space. The game industry surpassed the film industry in 2005 and the music industry in 2007. Trends from the October GDC online point to an increased emphasis on the application of consumer behavior metrics, sociological theory and advanced analytics in terms of driving game design and targeted marketing to this consumer base. A recent article in Internet Research is one of the of the first academic publications of it's kind to focus on gratification and flow-related determinants of game loyalty. The focus at Core Analytics and Game Loyalty is on individual consumer and player-level activation and loyalty metrics to guide marketing, virtual good optimization and strategic game design. An upcoming book entitled 'Game Loyalty' is targeted for April 2012 release, with pre-orders starting in late January 2012, that will focus on applied analytics and applications for all (mobile, social, video, PC, console) games, avatar analytics, multicultural, gender and age based micro-segmentation for various types of games, ROI associated with brand loyalty and player activation, and player behavior and psychology that applies across all game channels including casinos and sports.

Game Loyalty and Core Analytics: BrandMeter Loyalty Metrics mapped to a Virtual Recommendation Engine

2011-11-23T05:46:00.000-08:00

Core Analytics, LLC and subsidiary Game Loyalty have added the BrandMeter(TM) Recommendation Engine to the BrandMeter(TM) customer activation and loyalty measurement system. The BrandMeter Recommendation Engine is integrated with the BrandMeter(TM) datamart where real time customer activation and loyalty scores are stored, along with updated channel-specific offerings associated with the highest probability of consumer uptake. The offerings are categorized by channel, and by associated business objective (retention, cross-sell, brand awareness, new product release) and targeted at a customer micro-segment level to optimize enterprise marketing objectives. The BrandMeter(TM) datamart also stores customer response and tracking, to support campaign management initiatives.

BrandMeter(TM) beta client and investor inquiries welcome.

Neural Nets: Implications for Web Design, Error Resolution and Capacity Planning

2011-11-02T16:25:00.000-07:00

Neural network algorithms have long been in use in the forecasting and backcasting toolbox of load forecasters. Such techniques are useful in the energy industry related to short-term load forecasting, fault diagnosis, security analysis and other critical applications related to the reliability, stability and safety of electric generation, transmission and delivery.

There are two distinct kinds of neural networks: biological neural networks (neurons and synapses that make up the chemical processes in the brain) and artificial neural networks (a simplified system of artificial neuron networks mimicking the self-learning processes of the human nervous system, essentially an artificial network based on an underlying mathematical or computational model aimed at adaptive information processing or problem solving). Artificial neural networks (ANNs) are commonly applied across industries in speech recognition solutions, load forecasting, traffic planning, emergency response system development, air traffic control, and many others, and it’s these ANNs that provide the focus of today’s blog topic. Common uses of neural nets for business include classification, prediction and clustering.

The basic steps in deploying a neural net model involve selecting and preparing data, creating the neural network architecture, specifying mathematical models, training the model and then using it to forecast. It sounds complex, but no more so than any common statistical approach to prediction. In fact, it’s possible to build Neural Net Models in Excel!

Many examples exist where neural nets are in play in automated process error detection. These include engine fault diagnostics, facial recognition identification outcomes, digital fault tolerance and dynamic source routing control and link breakage prediction.

A classic tool used in electricity load forecasting is a neural network software called Metrix ND, that is used to forecast deviations in demand for power. While this tool is developed specifically for short term electricity forecasting, the application lends itself neatly to the forecasting of any type of demand load profile based on variance of inputs. In the case of electricity these are variations in temperature, storms, system failures, holidays, sports events; in the case of a high traffic web site like Amazon.com the variations are around holidays, sales, click-through rates, PPC levels, dynamic paid search query term pricing, site stability, traffic, competitive factors, site security, product launch, seasonality related to semester start, news items, and more.

By developing load profiles for clusters of web visitors and customers, Amazon or any other large e-tailer might develop a bottom up forecast by visitor profile (small business, power shoppers, book buyers, college students, information seekers, etc.) and have a high usage, average usage and low usage scenario with expected variance associated with each. Any time actual site traffic fell either above or below the threshold of these traffic and other digital metric forecasts, an auto-alert could be sent to the group responsible for the specific area of the site where the forecast was either too high or low, and by determining automatically when the deviation occurs with auto-triggers, this real-time knowledge of the magnitude of deviation from expected, and also the associated load profile in question, leads to deployment of targeted rapid resolution and adjustment strategy through this bottom-up approach.

Cloud Computing and Demand-based Pricing on the Big Data Grid

2011-11-01T16:07:00.000-07:00

There’s a very useful article in this month’s Harvard Business Review by Andrew McAfee entitled ‘What Every CEO Needs to Know About the Cloud’, that provides an excellent summary of the benefits of cloud computing, and repudiates the assertions of the noisy Cloud-critic community. IBM conducted a CEO survey in 2010, and found that 80% of CEO’s expected to see increasing levels of business complexity (you have to wonder about the 20% who expect otherwise) and over half of the group felt their company was prepared to handle the challenge.

Basically the Cloud computing revolution, as detailed earlier this year, is a paradigm shift from an internal, private and self-contained IT infrastructure, to a lease-based, shared IT extrastructure. The functionality of the IT system does not fundamentally change with a shift to a cloud. The difference is the enhancement in productivity and efficiency, where employees have direct access to resources and documents through a centralized, always-available portal. The Cloud option, not unlike Airline Premier Status, not only bumps an employee and their process straight through security without a hitch and then up to the front of the queue, but then directly onto the plane without a wait. The egalitarian aspect is this: the status upgrade is granted to all employees across an organization operating on a cloud and with access to a device of any sort that connects to the Internet, in companies where cloud-computing usage is set above forecasted usage. Since Cloud computing operations on a pay-as-you-go model and projected to be more cost-effective to rent than to own, this in theory bodes well for employee productivity and innovation.

The main concerns around Cloud computing are cost, security and reliability. McAfee argues that the same concerns exist for traditional on-premise IT infrastructures, and as with most new technologies, the benefits are often understated at the outset. Cloud computing facilitates group and community collaboration, which allows sharing of enterprise templates and case studies, and a standardization of approach and methodology. It also enhances data analytics, with preconfigured data mining applications that circumvent the time-intensive data consolidation and preparation phase and applies algorithms to enterprise data aimed at scoring clusters of customers, incidents and products and their respective propensity to purchase, result in error, or be characterized by high levels of risk or fraud, for immediate counter-strategy development and deployment. Companies no longer have to buy server space to support server space-intensive graphic applications, and can incorporate pre-existing tools for geo-location analysis and visualization. In every instance the solution becomes more scalable and much faster, which makes cloud-based companies more competitive. Some companies are utilizing private cloud and hybrid-cloud solutions.

The cost scenario comparison is unclear, and should be evaluated on an individual company basis, and with the understanding that while the efficiency benefits are immediate, the cost-savings are realized over extended time and usage. Reliability is also unclear, but can be mitigated with internal auto-cloud-management systems, which sit on top of cloud applications and monitor for errors and reroute processes elsewhere or back on-premise to leverage internal cache and processing systems.

It's fair to expect that a new breed of analytics and load forecasters will emerge to manage the reliability and cost risks involved with Cloud Computing. Just as with electricity forecasting, which is supported by a cadre of short-term electric load forecasting professionals to predict system load over an interval ranging from one hour to one week, a similar niche for cloud-based load forecasting (and high-traffic websites) is developing. These traditional forecasts are used for the purchase of energy and generation, load switching, contract evaluation and infrastructure development. Similar load forecasts, utilizing load profiles (groups of users with similar 24-hour load shapes) create a bottom-up forecast that can be aggregated to result in highly accurate short term demand forecasts. Such forecasts applied to cloud computing, identifying similar load shapes, for processes and users, will allow companies to accurately determine short-term cloud computing needs and associated costs. Just as electric-pricing is dynamic, cloud-computing pricing is a good candidate for demand-based depending on the supply and demand environment on the Big Data Grid.

Android vs. iOS: From a Developer's Perspective

2011-07-20T06:38:00.000-07:00

I attended my first MoMo Chicago event this past Monday, held at the VIBES offices at 300 West Adams across from the Willis Tower (they welcome inquiries from the Chicago tech community to use their very cool office space for events, contact Jack@vibes.com with inquiries). Joe Nedumgottil, Manager at Pariveda, delivered and excellent presentation on the topic 'What Can Android Do that iPhone Can't?" Here's a SlideShare copy of the presentation, if you missed it, complete with Android-specific code examples related to the various development categories.

The following is an overview of what I learned at the event, summarizing the take-aways from Joe's very informative talk. While it was presented to a group of developers, it was very interesting from a standpoint of what's ahead on the mobile horizon and how this rapidly changing technology is going to fundamentally change the way we live, and very soon. (This was one of the best Chicago tech events I've attended this year.)

Smartphone usage outpaced PC usage somewhere around the end of 2010. Some statistics mentioned in the presentation introduction include the following: Half of all local searched are done on mobile phones, one third of FaceBook users use FaceBook mobile, 50% of Twitter users use Twitter mobile, and there are 20 million YouTube videos uploaded each day. The importance of rendering a site for mobile was demonstrated by the following example related to Engadget and Gizmodo, two of the leading mobile developer blogs. Users who preferred Engadget who switched to following the blogs on mobile devices soon found that Gizmodo's mobile interface was easier to read, and faithful blog readers shifted to Gizmodo's easier to read mobile news. If a competitor is not focusing on their mobile site yet and has a growing mobile-user base, there's an opportunity to do this first and shift market share.

Just a few more highlights from Joe's presentation I found interesting, since the link to the deck is provided above. The differences in operating systems were grouped into Notifications, Widgets, Navigation, Voice Input, Intents, Multi-tasking and Events. The iPhone notifications are currently not superior, they are pop-ups and pretty bland, and interrupt the user experience when they occur, and Apple is working on improving these. In contrast, Android offers an unobtrusive notification bar with a drop down viewer option, and developers can develop multiple notifications per application. In terms of widgets, most phones offer the ability to arrange widgets similar to arranging applications on a desktop. Some of the most popular mobile widgets are Facebook, Pandera (song play and download), and Beautiful Widgets. The advantage of using widgets include real time Twitter streams, real time bidding on auction sites, real time sales numbers or inventory counts. The downside of widgets, on any device, is high battery drain and high RAM usage. One Android development tip to avoid the battery drain issue was provided by someone in the audience, instead of setting the widget up with http, to TCP push information in a less energy-intense and efficient manner.

Navigation was stressed as a real differentiator. iPhone doesn't currently offer turn-by-turn navigation from code. It's possible with Android to use speech input, and skip to directions without going to the map, and clicking through to directions, and provides live traffic updates as well. Voice input was another very interesting topic. Apple offers plugable text to speech, whereas Android offer TTS and native Speech to Text. In most mobile devices, emails can be read aloud, different voice selections are available. Android's is more flexible with the native option, and developers can create voice recognition with Google Voice Actions or Vlingo. Some tips to voice recognition, which still seems a little inaccurate to most users, is to speak slower, don't use contractions, and set the phone to recognize voice (accent and slang specific). In essence it's possible to train the phone, by correcting it, which is a hassle initially but once programmed and customized works very well. One of the neat options in the Multi-tasking category where developers can set conditions in the Tasker and Locale such as 'automatically turn to vibrate from 8am to 5pm and to ring between 5pm and 10pm" to eliminate the need for the user to manually change this each day, or 'automatically read text messages aloud' when phone is in car dock. Chrome to Phone was cited as another useful application. Just a fantastic presentation overall and very useful, thanks again to MoMo and Joe for a great information sharing session.

See you all at TechWeek soon! Core Analytics and Game Loyalty will be part of the Innovation Showcase at TechCocktail's 5th Birthday Bash, introducing the Game Module of the BrandMeter(TM). Stop by and say hi at the event this Thursday. Thanks to TechCocktail and Trisha Cruz for an awesome GameLoyalty write-up this morning!

Cloud Analytics, Big Data Analytics: The Sky's the Limit for Spatial and Geo-Location Analytics

2011-07-11T08:36:00.000-07:00

Continuing the general Big Data Analytics and Cloud Analytics theme, this TDWI article lists 11 predictions for Big Data. The article points to an increased focus on chips, over hardware, citing a huge gap between chips and business software. Put simply, more chips are needed to process the exponential data growth far more efficiently. Predictions indicate a single modern chip may be able to run up to 512 cores, and that's going to dramatically reduce hardware necessary to manage data networks. The new technology, when it emerges, will offer a more efficient option to traditional massively parallel processing solutions. Other trends include increased focus on analytics, in-house analytical solutions, a hybrid model utilizing open-source and commercial or proprietary analytical tools.

The Summer 2011 Esri ArcNews publication features a special cloud issue. ArcGIS Online is developing a cloud-based system for map and geo-location information sharing. The growing importance of cloud comping for GIS focuses on reduced costs, flexibility, scalability and rapid deployment. ArcGIS Online can be used by anyone to create maps that will integrate with most devices, embed maps into sites and web applications. By moving to a cloud-based system, it's possible to create an information-sharing platform for map-combination, accessible through APIs and free viewers. The system is based on thousands of geo-spatial datasets. The platform is searchable by keywords and maps can be enhanced by users wanting to add new dimensions and views to standard map views. It's all served up in a SaaS application, where users create and share the maps and data and perform sophisticated geospatial analysis in a cloud-based environment.

Here's an interesting geo-location add-on to Google Analytics called InstaVista. It runs on top of Google Analytics and allows real-time or historical visualization of online and multi-channel data in cloud-based dashboard system.

Probiotics, Chicago Weekend Links, Groupon and Data Privacy Changes, McDonald's Cloud Computing, Big Data Analytics and Password Security

2011-07-11T06:56:00.000-07:00

Starting off this steamy, stormy Chicago Monday reading about the health benefits of probiotics, while consuming a bowl of Whole Foods kefir and cranberry granola from Caputo's. Seems like a fair breakfast choice, it's delicious and there's a decent chance it might lower cholesterol and blood pressure, create a low Ph environment where armies of inflammation and infection-fighting antibodies thrive. I acquired a taste for kefir back during my study-abroad days in Krakow, specifically while living at the Przegorzały Castle.

It was a hot, sunny summer weekend in Chicago. Highlights were the Gold Coast Art Fair (complete with outstanding sangria, served at the Glunz Family Winery booth), dinner at Oven Grinders (world class pizza pot pies, highly recommend), checking out the European-style rooftop deck at Gene's Sausage Shop on Lincoln Ave, and the Chicago Folk and Roots Festival. Another high point was waking up yesterday morning to a Smart Data Collective email that the July 9th blog entry on Big Data Analytics had been published in their Sunday edition, a real honor. Groupon is taking the headlines again, introducing new privacy policies that will allow them to collect, store and share customer data with companies like Expedia, for micro-targeting using geo-locational data.

Following up on Saturday's big data analytics focus, the June 20, 2011 issue of eWeek features an article 'Moving to the Cloud: A 'Rational' Choice for McDonald's'. McDonald's has embraced cloud computing as a cost-savings and efficiency-improvement move. McDonald's is utilizing the IBM Rational toolset and hired CloudOne's consulting team to shift Rational to a cloud computing structure. Again, with this cloud-based model, customers pay for usage rather than a flat consulting implementation fee. It allows larger (and smaller) organizations to avoid long term fixed IT infrastructure costs and them to be more nimble in terms of adopting new technology as it becomes available. McDonald's cites increased application speed, flexibility avoiding long queues planning pilots and proof-of-concepts, and of course the zero-cost-of-entry CloudOne provides with their pay-as-you-go pricing models.

Also in the same issue of eWeek is an article 'Password Security Remains the Weakest Link Even After Big Data Breaches'. As companies shift to cloud computing systems and moving their enterprise data to the Big Data Grid, they become more vulnerable to cyber-attack and exposing confidential customer data. Solutions include increasing password strength and two-factor authentication, which includes a follow up verification page entering a six-digit code sent by SMS to a smartphone, and financial institutions are using a multi-layer approach with image verification and hardware token verficiation processes. Some interesting facts about password selection include the following: 93% of passwords are between 6 and 10 characters, 99 percent don't contain a non-alphanumeric character, 64% are random and not found in dictionaries, and 80% are absolutely unique. What about the most common 25 passwords? 'Seinfeld', 'password','winner','123456', 'purple','sweeps','peanut','tigger', and 'cookie' made the cut.

Big Data Analytics and Cloud Analytics: Embracing the Cloud and the Big Data Grid

2011-07-09T10:08:00.000-07:00

Sometimes you have a choice. You either jump in and join the game, or sit on the sidelines and watch other teams play. Core Analytics and Game Loyalty are joining in the Cloud game. Here's a preliminary introduction to what's currently going on in this space.  (Saturday is a good day to go through all the professional magazines and newspapers that have accumulated, before recycling).

I've asked the question many times over the past year: what exactly is cloud computing and how does it differ from current systems. (I'm still under the impression there's a bit of rebranding and productizing something that's been around for a while going on.)  Cloud computing involves using multiple server computers on a digital network, to centralize all enterprise applications. Not using a cloud, you would be using a single web server and with a cloud you are able to use many web servers, by renting extensive shared server space. The advantage of the cloud is that the work load is optimized, space is unlimited, and you pay storage rent based on usage. Essentially, this computing and storage sharing system appears to operate similarly to the energy grid, (which is a series of independently owned and operated plant and transmission lines) where distributed data flow and processing power is analogous to electric power transmission .

Another related term is Big Data. It's appropriately named, these are datasets commonly generated by weblogs, social networks, social data and they comprise large quantities of data that are beyond the ability of most existing software and technology to efficiently store, manage and analyze. New technologies, many of them open source, are emerging to handle these massive amounts of data. Some of the current leaders in the space are MPP databases, data grids, HPCC/ECL, Hadoop, and MapReduce technologies, scalable storage solutions, and cloud computing.

The June 27, 2011 issue of InformationWeek features an article in the analytics section entitled 'Ballmer Defines 'Big Data' in Microsoft's Terms'. To summarize the article, MSFT is making 'Big Data' a top priority, and they are approaching this from a different angle than Oracle (benefiting from their acquisition of SUN Systems and bundling data warehousing) and IBM (that has a grid-based NAS solution). Ballmer is focusing on how Big Data and cloud computing combine in an on-premises data and Microsofts' online data centers. Microsoft is a relatively late entry to the Big Data game, with it's SQL Server R2 Parallel Data Warehouse introduction in late 2010. Ballmer says 'Nobody plays in big data, really, except Microsoft and Google'.  What he's referring to here are the Bing and Google search engines. The search engines rely on infrastructure that easily manages petabytes of data daily. Bing and Google-related processing is far more powerful because it can scale up to petabyte-style deployments. The real value is in leveraging the data for Business Intelligence and Analytics. Nielson and Acxiom are other leaders in the cloud-computing and big data analytics space. Here's an interesting NoSQL conference coming up in San Jose at the end of August related to the Big Data topic.

Chicago Analytics Meetup, Core Analytics, Meta Brown, TenBrain, Cassandra and Hadoop Analytics

2011-06-30T06:28:00.000-07:00

Last night was the June Chicago Web, Social and Game Analytics Meetup at John Barleycorn on Lincoln Ave. This was smaller than many of our past events, maybe in part because we didn't co-list it as a WAW event this month. It was a great event however, a group of six industry professionals with very distinct analytics perspectives and application focus. Nice to sit out in the courtyard and sip wine for a few hours enjoying good conversation with such an accomplished group of folks, something that's more difficult to do with a larger group and structured presentation format. Long time member Meta Brown, Chicago analytics author, speaker and national analytics expert was there, and shared highlights of her recent publication 'Data Mining Fundamentals: Terms You Must Know' on Smart Data Collective (as well as a very entertaining Facebook connection story I'm still giggling about.)

Also in attendance was speaker and author of the 'Historical Women of Leadership', a blog celebrating leadership lessons from notable women in history (and who is also doing some cool things these days in the mobile analytics space). We were very excited to hear from a 2011 Iron Sabre Product of the Year award winner, who has a long list of industry awards for his social media solutions. New to group last night was another fascinating member, one of the co-founders from TenBrain , a Chicago-based lean cloud focused startup, that helps automate the cost of running infrastructure across multiple clouds providers.

One of the conversations centered on the growing popularity of Cassandra databases and Hadoop analytics that sits on top of and facilitates insights from sourced Cassandra unstructured data. IBM is doing some interesting product development in the Cloud Analytics space. Datameer is one company leveraging Hadoop Analytics for the social gaming space.

ITA TechBreeze, Google+ and Google Voice Invite, Player Insights

2011-06-29T05:50:00.000-07:00

Core attended the ITA TechBreeze networking event last evening at Rivers. Lovely night to be out and mingling with the Chicago technology community on the terrace overlooking the river. These are bi-monthly events for ITA members to reconnect and network. (Core is not yet an official ITA member but they were kind enough to let us register.)

The Google+ project grabbed the tech headlines yesterday, with it's (invite only, currently) social networking product that will be built around 'circles' (friends, families, co-workers, neighbors, etc.), that features a faster content sharing option 'sparks', and allows users to group message and share photos uploaded directly through mobile with 'huddle'. Submit your request for more info, and for the invite wait list here. Also, submit your request for an invite to Google Voice, which allows you to make free domestic calls in the US and Canada from gmail, here.

It's a brief blog entry day, wrapping this one up with some GoDaddy auction news. We added playerinsights.com and playerinsight.com to our analytics-domain collection this week. It was a first experience using GoDaddy Auctions, and really a very smooth experience. Simply log on with your GoDaddy account info, and there's a 2-step bid and confirmation process, an opportunity to put in a max bid, and email notification when you win, just like EBay.

I micro-blogged about this book a few weeks ago but adding a link in here. Apparently 'The Internet is a Playground' has been out for some time, and featured on BBC and the Late Show (and others). I found it on the employee recommended shelf at Books Inc. in Mountain View (which by the way is just up the street from Neto, an outstanding Mediterranean Bakery). Some of the really hysterical content from the book can be accessed here. The Magic 8 Ball is a great idea too.

CSL West, SemTech 2011, New Domain Name System, and Pho

2011-06-27T05:44:00.000-07:00

This week begins from the front porch, surrounded by the past few editions of the FT, a book on social media games, and coffee in one of my collection of the blue and white architectural series Starbucks mugs. I'm 'in' New York this morning.

I returned recently from the west coast. Core and Game Loyalty are sponsors of the Bay Area Web, Social Media and Game Analytics meetup, we had a our first event at Cascal on Castro St. in Mountain View. I attended another fantastic meeting of CSL West at the Microsoft offices in Mountain View, built around a discussion on how online and offline communities mix, with specific company examples (HP, EBay, etc.) on strategies and formal programs rewarding power users. This was a really fascinating group of attendees, employees and former employees of MSFT, Google, Ebay, Yahoo, HP, Apple, and other powerhouse tech companies, sharing ideas in an interactive environment, directly on site at one of them. Silicon Valley has this open, shared spirit towards entrepreneurship and facilitating the growth of social communities, also supported by the companies headquartered in the area. In this particular meeting there was an extended discussion on the online/offline mix for the SETI Institute (which recently had to suspend it's search for aliens, due to lack of government funding to keep the Allen Telescope Array running). Mountain View is an extremely interesting place. Not only do you have events like these nearby and in San Francisco to pick from literally every night of the week, 365 days a year, but the culinary selection just along Castro Street alone makes it a destination. After my friend Nelson introduced me to pho at Pho 777 earlier this year in Chicago, I've become a Vietnamese food fan (addict). There are several good ones to choose from along Castro, and we're hosting the upcoming Bay Area Web, Social, Game Analytics Meetup on July 11th at Xanh. No pho discussion would be complete, however, without mentioning Tank Noodle in Chicago, offering arguably the best pho and spring rolls on the planet. But...I digress.

Core had the opportunity to speak on the topic of 'Social Media: Driving Brand Activation and Loyalty' at SemTech 2011 again this year at the San Francisco Hilton. SemTech is the world's largest semantic technology conference that focuses on semantic and text mining applications in industries like health care, finance, publishing, and other industries. Conference buzz for SemTech was overwhelmingly positive, as seem in this SemTech 2011 #SemTech TwitterScript matching tweet volume to conference session. Core is featured in there on June 8th too, thanks!

One more quick comment for the entry today, yesterday the Financial Times had a brief media article 'New Internet Domain System Worries Business' (couldn't find the original link online). ICANN, the governing body overseeing Internet domains, has determined starting in January of 2012 companies can secure and use corporate names (.coke, .apple, etc.) instead of relying on traditional .com (which was introduced 26 years ago!), .net, .org names. They come with a steep price tag though!

Social Game Analytics, Virtual Good Analytics, Gamification and Game Loyalty

2011-05-22T06:03:00.000-07:00

First some news and updates, then on to the first of what will likely be many gamification and game analytics discussions. Core Analytics(TM), LLC has been granted the trademark for the Core Analytics name...exciting stuff. We've also just launched Game Loyalty, a Chicago and Bay Area-based social game advanced analytics start-up that focuses on segmentation, predictive analytics, game design optimization and virtual good analytics that uses the Game Module of the BrandMeter(TM). Core's going to speak at SemTech2011 on June 8th in San Francisco on the topic of 'Social Media: Driving Brand Activation and Loyalty'. This is a longer 30-minute presentation similar to the recent SXSW presentation by Core Analytics(TM). Core Analytics and Game Loyalty are sponsors of the Bay Area Web and Social Media Analytics Meetup. Join the group at Cascal in Mountain View, CA for the June 6th, 2011 Mountain View Web Analytics Wednesday. The Chicago WAW was held at the James Hotel on May 16th, around 30 Chicago-area analytics professionals came out to meet the event sponsors Maxymiser and learn about their very cool multivariate testing platform for page and ad content. Maxymiser delivered a fantastic presentation, which had the group guessing which version of each ad got the most clickthroughs, or drove the most revenue, and they very effectively demonstrated how our assumptions and intuition on best content performance end up in the selection of the least effective versions, and how key multivariate measurement is. Excellent presentation, product suite and team, thanks again to Maxymiser for hosting our May Chicago Web Analytics Wednesday. We'll be posting some more information with a link to their materials in the next couple of weeks. To wrap up the May news, we're really excited to add Glenn Raines of Social Media Moves, a well-known Chicago SEO expert, author and speaker to our analytics meetup membership list. I had the good fortune to be sitting next to him at Starbucks (my new office) and he shared some very effective LinkedIn Search and SEO strategies, and introduced me to some new sites and applications of programs (Google Alerts for SEO and keyword selection, EmpireAvenue, etc.). Last but not least, Core is also sponsoring the new Chicago Area Healthcare Analytics Meetup that will ramp up with events and speakers as soon as our membership grows.

With the launch of Game Loyalty, we're honing in on the social game analytics space. We're currently well into the development of the Game Module of the BrandMeter(TM) now in private beta. (Beta client inquires to info@coreanalytics.com). Given the new focus, today's blog entry is aimed at a general overview of Social Games, Game Analytics, Gamification, Virtual Good Analytics and some mobile and operating system trends that define the ever-changing mobile and social gaming landscape. Gamification has been a buzzword for the last year or longer, it refers to the use of game mechanics (Funware) for non-game consumer and marketing-oriented web and mobile based applications. (Draft1)

Core Analytics at SXSW, Review of 2011 Trends in Social Media from SES, Ira Weiss speaks at TechNexus in Chicago

2011-03-20T20:27:00.000-07:00

Core is back from Austin, what a delightful city to spend five days in for the annual SXSW event, now in it's 25th year. After a long winter in Chicago, relaxing in 80 degree temperatures and exploring the Hill Country and Lake Travis made this a great combination work/vacation trip. This was my first time attending the conference, and it was really interesting. The Future 15 panel discussion moderated by iCrossing on Brand Connectedness was jam packed...somewhere between 400-500 attendees with a line to get in. All of the presenters did a great job, it was a nice mix of presentations and companies. Core's talk on 'Social Media: Driving Brand Activation and Loyalty' went well, several nice comments afterwards and interest from angel investors and practitioners in the analytics space made it feel like a success.

Returning to 30 degree temps in Chicago was a bit of a shock, but the week picked up when I attended the ITA/TechNexus talk by Ira Weiss on Wednesday evening. Ira is an excellent, entertaining speaker, and the event was well-attended. Some of the takeaways from his discussion were helpful...it turns out having a business plan is less necessary than many entrepreneurs assume, it's more useful to have a well assembled deck describing the business, and a well-practiced compelling elevator pitch is key. The longer startups can go without outside investment, the more of the company they will ultimately retain ownership of. Having friends and family invest is a sign of credibility, since these are people who have known the entrepreneur a lot longer than potential investors and this corrects an information asymmetry when evaluating a ventures prospects. Angel investors get less involved in the day to day control of the business than VC investors, and the Hyde Park Angels typically invest 250k-1 million in a company they take on. Valuations for companies on the west coast are higher. Given the nature of the Chicago startup scene, collusion is unavoidable dealing with VC's locally. If you're going after VC, having some west coast exposure to increase valuations is something to consider. Companies that look like they will turn out to be 'lifestyle companies' are not desirable from an angel investment standpoint. PhD's aren't bad, but they're not necessary. The ITA folks also talked about lean startup principles, and introduced the group to a series of classes for entrepreneurs to hone their strategy and offerings, at reasonable rates. There are some helpful sites out there...www.angellist.com is one that connects startups with angel investors. It was an informative, interesting event. The audience was about 95% male. I'm still puzzling over that one. The same was true at the BNC TechPitch where Core presented in October last year. This article addresses the issue, and the shortage of women in tech

The new SES magazine came in the mail this week, it has an interesting list of social media predictions and trends for 2011. Some of the highlights...Facebook will go public and Mark Zuckerberg will remain on as CEO, Social Commerce will ramp up, focus will go to getting a consumer to publicly rate a product, Groupon will regret not taking Google's offer as imitators gain market share, and the main Social Media players (Facebook, Twitter, LinkedIn) will survive but the others will lose traction through consolidation, mobile-specific sites will become commonplace, and number 19 was really interesting...'more people will learn they can plug their laptops into their high def tv's and watch online content, and will ditch their cable tv service and bill'. This last one is something I have yet to try...and as I don't currently have cable, since I'm not a tv-watcher by nature and live by the internet and dvd's for my news and entertainment, I'm going to hook up to my new flat screen and see for myself. The list was written by Erik Qualman, author of 'Socialnomics: How social media transforms the way we live and do business'.

One more item...the review by TechCrunch on the Source Code film preview at SXSW on Facebook. I haven't seen the film, just the Facebook comments and review.

Game Analytics, SXSW Interactive, CapLinked review, SlideRocket and more.

2011-02-24T06:18:00.000-08:00

Core's headed to Austin in a few weeks to speak on the Future 15 Brand Connectedness panel at SXSW Interactive. We got invited in the second selection round, to do a 10 minute talk and then a Q&A follow-up entitled Social Media: Driving Brand Activation and Loyalty'. It's very exciting, this will be the first time at SXSW for both of the Core partners. I'm revising the presentation using a complimentary SlideRocket Pro account extended to conference presenters, guaranteed to make a presentation 'sizzle'. SlideRocket let's you embed photos, video, live twitter feeds and more, it's very sleek, easy to navigate, there's an extensive pre-architected library of presentations to select from. The video insertion feature is amazing, with a click you add a video that is embedded into the slide and as you progress in the slide show the video automatically starts in it's window according to timing and volume specifications.

I'm personally excited about Monday's lineup of sessions at SXSW Interactive, all around gaming and mobile. Here's an example (Future of Mobile gaming) of two sessions (Game On: 7 Design Patterns for User Engagement) that are on my list to check out.

SmartBlog on Social Media had a couple very interesting links yesterday. Here's one entitled 'Exploring the Serious Business of Game-based Marketing'. It talks of gamification as a final state in customer loyalty reward programs, where rewards are virtual goods that are valued by consumers for their personal, virtual currency. (Core Analytics is moving into the game analytics space as an additional digital channel for the BrandMeter(TM) dashboard, capturing user metrics of interest that tie to brand loyalty activation and measurement, we have http://www.virtualgoodanalytics.com/ and http://www.gameloyalty.com/ reserved for future expansion into this space).

The second article of interest in SmartBlog is on a new site featured in the WSJ, called CapLinked. The site is an exchange platform for entrepreneurs and investors. We tried the site out yesterday and posted a profile for Core Analytics and a few discussion topics, and within six hours had a first inquiry email from a potential investor asking for more information on our team, product and funding specifications.

Patient Loyalty towards Healthcare Providers and Facilities

2011-02-10T07:10:00.000-08:00

Insurance brands tend to enjoy some of the highest levels of customer brand loyalty among all consumer product categories. Service industries in general tend to have more loyal customer bases. Older consumers tend to feel most loyal towards auto, home and life insurers, and other financial products and least loyal about consumer electronic brands and food. The article stresses that older consumers are more likely to be loyal to a brand that provides a service that simplifies their lives. All are related to maintaining overall financial security.

Health care brand loyalty, given the nature of employer-sponsored vs. individual and state-sponsored programs, is harder to define. Consumers are limited by choice, subject to auto-enrollment, and as this article points out what appears as loyalty may be simply a habit.

There's a lot of literature on patient loyalty to hospitals, and towards providers. This has implications for health care companies and how their provider networks are structured, since carryover from provider loyalty to a health insurance company that provides access to a preferred provider is one part of the health care insurance loyalty equation. In terms of hospitals, Gartner defines patient loyalty as a likelihood to return to a health care facility. The article gets into micro-segmentation and loyalty triggers based on patient need, and specific instances of care. For example, a sense of urgency and speed of service, and pain management will have higher priority in an Emergency Room setting than in an inpatient setting, where a friendly, competent nurse and clear communication would rank higher. Two things stand out regardless of the nature of service: concern of the staff, and perceived quality of care appear to be the primary overall drivers of patient loyalty.

Certain providers have implemented patient loyalty programs, that include a pleasant office, convenient location, remembering birthdays with cards. Ridgewater College offers one of many available continuing education and/or professional courses in the US on the topic of driving patient loyalty. Their brochure describes a Patient Satisfaction model, where satisfaction is achieved when patients expectations around their health care service are met or exceeded. Key drivers of success include coordination of care, accountability, service-focused vision that comes from the top down. Patient loyalty derived in such a setting is characterized by those who return to a provider facility over time, refer to friends, and comment about providers in a way that reflects the relationship as a partner in care.

How does the patient-loyalty model change, under proposed health reform, for tax-subsidized hospitals that serve the uninsured? With expanded Medicaid eligibility and an insurance mandate, patients will have expanded choice and have the option of private doctors and hospitals for care. What constitutes patient loyalty in such a setting? Most likely tax-subsidized facilities will see their patient base change with reform, and such hospitals will need to learn more about their new patient population and update patient service and satisfaction strategy accordingly. Here's an interesting link about selection by Medicaid members and managed care Medicaid insurers.

Virtual Good Analytics: First in a Series of Research Articles by Core Analytics

2011-01-02T09:25:00.000-08:00

Happy New Year! New year, and new resolutions...one of them is to post more regularly in the Click Insights blog. Just some brief updates..the BrandMeter(TM) development team at Core Analytics has completed a working beta version of the customer brand activation and loyalty BrandMeter dashboard solution and is starting to make this available to clients in private beta. Core Analytics will present at SXSW Interactive on the Future 15 Social Connectedness panel in Austin in March 2011.

Today's topic is the first in an ongoing series around around branded virtual good consumption. Going into this, the hypothesis is that consumers who consume virtual goods are more likely to purchase durable goods by the same brand. That expressed virtual preference should translate into durable branded good sales. As I'm in need of an advanced retail-related independent study topic for this upcoming (and fingers crossed, last) semester at Purdue, the blog will start by taking a look at what academic literature has covered and determined on this topic so far.

Right off the bat I realize I've gone about this in the wrong way. (Side note: Scandinavians appear to be featured at the forefront of this advanced technology research around social and mobile). Folks are already paying (a lot) for virtual goods, virtual good sales are a major source of revenue in their own right, not even considering the indirect effects of demonstrated virtual preference driving durable profits. The virtual good market was estimated to be 2.1 billion in 2007, and more recent numbers slightly lower, at 1.2 billion for the US Market as released in this social game and virtual good monetization report. Here's another one that estimated it would reach 2.5 billion in 2009. The Asian virtual goods market makes up 90% of the global virtual goods market currently, and the total global virtual goods market is around 5.5 billion. Consumers in Asia have been purchasing virtual goods for a decade, consumers in the US have only been doing so since 2007 or 2008. Forecasts expect the US market to continue adoption at even higher growth rate now that payment issues (Facebook credits, etc.) have been introduced. There's an interesting statistic here, Chinese consumers spend 3 times as much on virtual goods now than their US counterparts, however have 1/6 the income. Most of this virtual goods consumption is happening on mobile devices in Asia. Here's a link to an excellent presentation on the topic. More to come on this topic as research continues.

Hospital and Doctor Loyalty, Live/Work Start-up Incubators

2010-12-16T05:50:00.000-08:00

My business partner and I had an interesting conversation the other evening, when driving past the new Elmhurst Memorial Hospital that sits at the corner of Roosevelt and York Roads. The campus is massive, and it's a beautiful series of buildings. I pointed out that it looked like the Four Seasons (see inserts, it really does). The website describes the hospital, set for completion in Summer of 2011, as a new concept in healthcare, serving not only to heal bodies, but renew spirits and energize souls. Aldo made the comment 'Why aren't hospitals pleasant places to stay, with cheerful decor and some luxury? They cost a lot more than luxury hotels'. Good point, and the trend in new hospital development reflects that sentiment. Here's a link to a luxe hospital in Michigan that offers an in-hospital organic grocery store, Caribou Coffee, rooms have flat screens and wi-fi, and live feeds of classical music.

Patient loyalty is a very real issue for hospitals and providers. Luxe hospitals are likely to attract profitable patients, as well as top rated providers using state-of-the-art equipment and techniques. Gallup found that hospitals with the highest patient and physician loyalty metrics had 80% higher earnings. Here's a link to a 'Patient Loyalty Calculator', where patient loyalty is tied directly to patient satisfaction, which is tied to profitability. One study found the top three determinants of patient loyalty to doctors to be treated respectfully, quality of care, and helpfulness of front desk attendant at the office. A similar study found that patient loyalty to healthcare institutions (hospitals) was determined by overall satisfaction with services, the reputation of hospital in treating particular disease patient is facing, and perception of consistent quality of care.

On another note, I'm signed up for the Napa Valley 10k in late March and have been spending more time on AirBNB looking for eclectic hotel alternatives in the Bay Area. Here's a really interesting option, a Live/Work Mission Start-Up Incubator for $65/night. It's a live/work setup for a web development company and their friends and guests. This Miami-based live/work incubator was started up for artists earlier this year. Here's another link that describes the incubator trend.

Core Analytics to Speak at SXSW Interactive on Social Media: Driving Customer Activation and Loyalty

2010-12-15T07:08:00.000-08:00

The blog has been neglected of late...but December turned out to be an exceptionally and unexpectedly busy month. We got the invitation email to speak at SXSW Interactive on the Future 15 and Social Media Connectedness panel in Austin in March...we're thrilled at the opportunity to attend the conference and participate in this way. SXSW is in it's 18th year and brings together enthusiasts from film, music and interactive fields to share, train, and connect. The lineup in terms of company representation, experts in the field, groundbreaking topics covered is truly mind boggling. I've been trying to figure out where to stay at http://www.airbnb.com/ during the conference, some very cool options on here. It's so much fun to participate in these conferences. We're working on our applications for SemTech 2011 and DMA now as well, all around the social media and customer brand activation and loyalty we focus on with Core Analytics and the BrandMeter(TM) system.

Other fun personal news, I have a fantastic committee assembled for my PhD work at Purdue in Consumer Behavior. If all goes well the degree will be completed by end of 2011, with prelims coming up in May. So nice to see a light at the end of the tunnel.

Current research this month includes investigations into propensity for impulse purchase. This becomes relevant as consumers rely on mobile more and more and real-time offers based on geo-location become main stream. It's basically purchase propensity models by product line, then incorporating in psychological variables that capture individual likelihood to purchase without prior deliberation. More tomorrow, stay warm!

Veteran's Day, and the BrandMeter(TM) is Official

2010-11-11T04:27:00.000-08:00

The reminder for Veteran's Day came on Facebook last evening, where one of my friends replaced his profile photo with a poppy, and another friend liked the photo change and mentioned she'd worn a poppy on her lapel walking around London on a trip there this past weekend. Everyone finds their own way to honor and remember Veterans and social media is offering new ways to do this. Here's an interactive map by tbd.com where you can mark the map location to show where the veteran you are honoring served. The official blog of the Coast Guard offers good suggestions on how to honor veterans in our communities.

On a different note, Core Analytics got a delightful notice from the USPTO last evening that the BrandMeter(TM) trademark process is now finalized, and Core owns the mark to use around the 'measurement of customer loyalty' solution sets in development. Small victories are sweet.

Malls, Social Intelligence, L'Occitane Viral Video and a Quick Look at the Nook

2010-11-07T07:06:00.000-08:00

It's been a long time since I wandered from end to end in a mall, as I found myself doing last evening out in Oak Brook. Coming from the east coast, this concept of upscale outdoor malls in a Midwest climate that would seem far better suited to the indoor variety is still somewhat of a puzzle, but the experience yesterday on an early November day went a long way towards explaining it. It was a fun atmosphere, with the holiday decorations and music, loads of people wandering around, lots of retail activity. Kind of a festive market square type experience. The research on the psychology of shopping and retail outcomes based on mall-type is a bit obscure. (It would be interesting to get insights on the type of customer attracted to an outdoor mall vs. interior, or vertical...and if non-holiday sales peak during warmer months and cold temps are not a significant deterrent during holiday shopping.)

I started off at Barnes & Noble, where a surprisingly large amount of retail space is devoted to the promotion of the Nook. Pricing starts at $149, over 2 million books can be downloaded in seconds, thousands of titles are free and most others are under $9.99. There are Nook apps for iPhone, iPad, Android...and the Nook Color starts at $249 and also allows users to surf the web and interact on social networking and gaming sites. Here's a recent comparison on e-book reader options that bring up consumer preferences for features such as weight, backlighting, need for constant wireless connection, pricing, screen size and so forth.

L'Occitane is located at the opposite end of the mall. I became a fan when introduced to their products during my first Four Seasons stay earlier this year (because of a fortunate corporate arrangement and preferred rates comparable to any other business hotel in the area), where the Verbena collection was featured as part of the hotel amenities. The brand represents 'natural beauty and well-being'. Nice, it fit the description of the hotel as well. The fragrance is a mix of Mediterranean verbena, geranium and citrus and since then it's also become my fragrance. I suspect if I'd encountered the collection during a stay at the Holiday Inn I might have continued shopping for a new signature fragrance, just as if the scent had been overly floral. Consumer preference in this case was driven by luxury associations as well as an identification with the fragrance. In any case it was great to find out they had a retail boutique in the mall, and I stopped by and finally asked how to pronounce the brand name (lox-ee-tahn) and this started a discussion about a funny viral L'Occitane video asking people on the street in New York to try to pronounce the brand name. Great use of humor to drive brand recognition.

Social Judgment Theory, Social Media Loyalty and Activation, Core Analytics Competes

2010-11-01T08:12:00.000-07:00

Interesting facts for the day...77% of the US top 500 Fortune companies started without outside capital and were fully self-funded. Here are a few theories about what drives entrepreneurship. The Meyers-Briggs ENTP personality type is typically associated with those attracted to entrepreneurship. And finally, Core Analytics is polishing the Executive Summary due tomorrow by 5pm for the 24th Annual Burton D. Morgan Business Plan Competition at Purdue. This is a $100k competition, first prize winner gets $30,000 towards operationalizing their business idea. Go Core.

Today's topic relates to persuasion theory developed by Sherif and Hovland in 1961. This theory emerged from the socio-psychological tradition relating individual judgement to the content of messaging, with the objective of determining impact on attitude change. SPT is related to anchoring theory, where a standard for comparison impacts individual perceptions. This theory is also based on the assumption that true attitudes are fundamental to self-identify. Since consumers choose brands to signal aspects of their self-identity, this has implications for persistence of loyalty. (i.e. brands with strong self identification and self-esteem ties in their customer base should be poor targets for attitude change targeting unless the brand attempting conversion reflects similar lifestyle and value characteristics). Put another way, brands that incite high ego-involvement have customers who have a more extreme stance on the lifestyle choice or value the brand is representing. Methods of measuring levels ego-involvement is based on decades of experimentation and research. (One componant of the BrandMeter(TM) modifies and applies these concepts to new variables derived from text mining and semantic analysis of raw social media comment data.)

Hedgehog Theory, Variety Seeking Behavior, Persuasion Theory related to Customer Loyalty

2010-10-31T05:07:00.000-07:00

The day started off reading the Newsweek article on candy choice this cold Halloween morning, over a cup of pumpkin spice coffee. As the trick-or-treaters start coming by around 3 this afternoon, there's still time to prepare with more candy...sounds like Tootsie Pops are a sure bet. The article refers to Cornell's Food and Brand Lab and St. Joseph's University Food Marketing program, and scientific research showing that kids (and other consumers) are drawn to what's new and unusual. Which is why given a choice, they will first empty the lollipops from the candy basket before reaching for the full-size chocolate bars.

This is interesting from a brand loyalty perspective. By nature consumers are drawn to the new and unusual, which means a potential brand defection risk exists each time consumer selection occurs. Here's some background into preference theory driving this phenomenon. The Hedgehog theory addresses an individuals preference (or tolerance) for complexity and change. Each consumer has an individual optimal level of new stimulus, or preference for complexity, and a different rate of assimilation to new stimuli that determines individual persistence of preferences. This explains (in part) why brands reintroduce new, rebranded, relabeled, or improved versions regularly in product cycles, and why some consumers tend to be early adopters and others show degrees of variety-seeking behavior. Research into online shopping behavior indicates that customer loyalty may be easier to secure in an interactive, online environment.

Social Media and Healthcare: October Chicago Web Analytics Wednesdays

2010-10-27T05:24:00.000-07:00

Yesterday evening was spent putting together the final deck for this evening's Chicago Web Analytics Wednesdays event, on the topic of Social Media and Healthcare. Essentially the presentation is just a bit of market research about what's out there now, and hopefully promotes a lively discussion with the meetup group attendees. There's a plethora of healthcare-related social media sites and information to sift through and assemble in a way that portrays a comprehensive snapshot of what healthcare representation looks like in social networks and social media. Healthcare as an industry currently has one of the lowest rates of social media representation. Legal and regulatory issues around patient information and pharma marketing prevent hospitals, providers, health insurers and drug companies from full-scale adoption of the social media channels. Most health insurance and drug companies have representation on Facebook and YouTube, and many of the examples found tend to be more non-interactive, one-way corporate-generated informational or marketing related communications. Take for example, this humorous United Health Care commercial or this compelling Blue Cross & Blue Shield customer testimonial video. Others are a venue for dissatisfied customers (in this case http://www.sickforprofit.com/) to express their opinion of the brand and health insurer policies, as in this Cigna tagline contest, which has attracted a distressing (from the brand's perspective) number of views (48,534).

Deloitte's healthcare consulting group put out an excellent white paper on Social Networks in Healthcare a couple months ago that also provides some recent numbers around how individuals are using the web to research conditions, find providers, and so forth. As an overview (this information is slightly outdated and understating the actual numbers) Twitter has over 100 million unique users and is experiencing growth at over 50,000 new enrollments per day. YouTube had almost 100 million unique US viewers a year ago and over 42 million users. Facebook now has over 500 million users. Specific to healthcare related searches, surveys show that over a third of those in the US who go online to research health issues also use social networks (forums, facebook, other social media sites) to find others with their condition to discuss and share experiences before making health care decisions. One in six doctors participate in online physician information-sharing networks such as the popular Sermo site, and 88% use the web to search for medical device and drug related information. Drug companies monitor Facebook condition-related pages and online drug discussion forums to recruit for clinical trials. Forrester put out a report last year about physicians networks being a huge growth opportunity for pharma marketing. Nearly 15% of US Hospitals have a social media presence. Patient information sharing and support sites are popular, such as http://www.inspire.com/ and http://www.patientslikeme.com/. Government agencies such as the FDC, CDC and HHS are keeping pace with the social media curve. The CDC has representation on all of the major social media channels, guidelines for use, and a couple current campaigns related to swine and seasonal flu.

In terms of innovative uses of technology and social media in this space, there's some really interesting stuff out there. For example, the patient-doctor 'speed dating' sites that allow patients and doctors to meet one another and assess mutual doctor-patient compatibility. (Confession...this application was suggested by a doctor friend the other day, and was intriguing enough that it prompted a search and secure mission for 3 new domains for the Core Analytics library...pcpmatch.com, doctoranalytics.com and docinsights.com. Another idea that came up in the discussion about entrepreneurial direction options for http://www.druganaltyics.com/ is an interesting one, and based on some very cursory online research does not seem to be fully addressed yet by market solutions. Often times medical researchers hire research assistants to compile clinical trial results, and a search engine type spidering/scraping tool that assembles all of the clinical trials and outcomes automatically, something like Harzing's Publish or Perish for clinical trials ).

Another great application of technology that's been around for a while is the development of mobile applications for disease management and wellness-related communications, such as this AT&T diabetes app. Telemedine (m-Health) is another area that is seeing high growth...there are now over 6,000 health-related applications for cell phones.

Anyway, that's short overview of the deck and discussion on tap for tonight's WAW, at 6:30 at John Barleycorn...if you're local please come out and join the discussion. We'll post the video later this week and a copy of the deck. Hope to see you there!

Au Contraire: Social Media Measurement is an Exact Science

2010-10-18T05:40:00.000-07:00

This morning's daily iMedia link just arrived, with a link to an interesting 'Tracking Social Media Metrics: Is it worth it?' testimonial video featuring well-known industry marketers discussing their perception of the value of specific social media metrics and limitations surrounding the process. The timing is apt, as Core Analytics prepares for their pitch debut tomorrow evening at the Chicago BNC TechPitch event. We've been working on formalizing our value proposition. (It's one thing to know what differentiates your offering but another to consolidate it into a single concise PowerPoint slide!)

Core Analytics has been invited to speak on the topic of social media metrics at a series of events and conferences this year. We sponsor many of the monthly Chicago Web Analytics Wednesdays that typically feature industry technology providers and marketers sharing their digital marketing and measurement best practices. We've been facilitating, tracking, reporting on and predicting outcomes of multi-channel marketing campaigns for a combined experience total of 35 years. So the (above) video's intro saying 'we all know social media measurement is far from an exact science' stuck out. Our consulting and solution is based on the premise that social media measurement can (and should) be an exact science. It's just that the current commercially available tools, as excellent as many of them are, are still more or less in the first-stages and are basically marketing dashboard reporting-type tools that don't yet incorporate detailed segmentation or industry customization. They provide basic enterprise-wide metrics to show marketers a landscape snapshot of how they are doing in terms of number of comments, followers, fans and general tonality of comments. These tools are useful in the same way that dashboard reporting and business intelligence tools are for any marketing or product development team. But if you want to get into targeted marketing, that calls for advanced analytics, micro segmentation and predictive modeling. Just as IBM acquired Unica and SPSS to build out their advance analytics capabilities, and SAS analytics software is deployed for predictive analytics at organizations on top of database platforms that also feed data into Cognos and Crystal Objects cube type trend identification tools, social media advanced analytics tools that integrate with the existing social listening software solutions will emerge on the scene.

One of these advanced analytics tools is called the BrandMeter (TM) and Core developed the architecture. It makes social media more of an exact measurement process. Specifically around that elusive engagement metric mentioned in the video, except we refer to this as 'customer activation' and 'customer loyalty'. The BrandMeter solution actually makes it possible to solve the issue of tying a customer comment to an actual retail outlet purchase (provided that customer has shared that action in comment form at a prior time on the social media channel). Conceptually it's no different from any other database, where customer transactions over time are stored and mined for customer insights, and this data is used for churn, cross-sell, segmentation and retention purposes. The comments are mined (semantically) for their value relative to actions such as 'information sharing', 'purchase', 'return', 'coupon use', 'information seeking', etc and over time an individual level behavioral pattern emerges, and is stored, and is used from a segmented level to support all the traditional marketing activities but from a digital channel perspective. The data is also available for integration with all of the call center, off line transactional, web navigation and other channel data to get closer to the 360 degree view of the consumer. (The elements that are compiled to capture loyalty and measurement scores from social media comment data can also be applied to comment data from chat, emails, and transcribed call data, and also added to the database.)

Customer brand loyalty, which is what is at the heart of the BrandMeter measurement, is a fluid, non-linear and very individual phenomenon. Just as loyalty in any relationship is both strengthened and weakened by both insignificant and significant endogenous and exogenous events that occur between the participants, so is customer brand loyalty. The events are interpreted to and reacted to differently based on personality, ego, and expectations. Tools available currently take a top-down level approach to measurement of a consumer base in aggregate. It's impossible to capture the individual nuances that form the foundations of customer brand loyalty from aggregate data. To get at this, a bottom-up disaggregated approach is required, and also from a matrixed perspective. From an industry perspective, and then from a brand specific perspective...baseline categorization of types of events (delight, surpassed expectations, motivations, steps in the purchase process, distress and crisis, disappointment, surprise, trust, etc.) are formed, automated text mining and semantic applications accurately categorize the comments, a social media datamart stores the comments at an individual level (reminder: this data is perishable, if companies do not retain this data they lose access to this or will find it prohibitively expensive to repurchase at a later time) from a good source (note: using more than one comment data source is advisable), and then more personality type mining is done. The underlying theory and algorithm is rooted in psychology, sociology theory, and econometrics and predictive modeling produces the final results. Sounds complicated, and it is complex, but it's also what traditional marketing analytics has been focusing on for decades, with the additional emphasis of text mining to extract the variables.

So, we argue that it is possible to get an exact read on how 'active' and how 'loyal' customers are at any given time, and also gather the same for competitors to understand what social media brand loyalty and activation share your brand has over time, and act on that with targeted marketing strategy. And now we're still no closer to having that 1-page value statement done for tomorrow evening, but it's still a fun debate. Come out and support Core's pitch if you're local, and have a great Monday!

Attribution Modeling Techniques and Social Media and Healthcare, Core Analytics at TechPitch

2010-10-13T08:12:00.000-07:00

Finally getting back to the blog on a grey mid-October morning, listening to thunder rumble while leaves fall...an unexpected but appealing combination. The topic lineup for today's blog starts with the upcoming October 27th Web Analytics Wednesday in Chicago. We'll have presenters and take a general look at the Healthcare and Pharmaceutical industry that currently has the lowest rates of adoption of social media as a marketing and customer interaction channel. The event will include presentations, discussions and networking. We'll address some of the privacy and legal restrictions that have caused this lag, and actual case studies where insurers, providers, hospitals and government agencies are effectively using Facebook and Twitter for brand marketing and wellness information communication to individuals. Join the group on the 27th at John Barleycorn at the corner of Belden and Lincoln Ave starting at 6:30pm. Sign up details here. Additional presenters/speakers are welcome. On another note...Core Analytics, LLC will be featured in the business plan pitch in this month's BNC TechPitch...come out to support Core and networking with the growing Chicago start-up community.

I recently attended a fascinating seminar at Krannert, a 'Marketing Models' course that focused on game theory applications for marketing, and journal articles using hierarchical Bayes, conditional logit and survival analysis models to determine relative effectiveness of marketing, advertising, location and product features as drivers of consumer purchase behavior and retail outcomes. Here are some useful takeaways in terms of attribution modeling approaches.

To begin with, software and technology varies widely in their terminology around what constitutes multinomial logit modeling. For example, the IBM SPSS 19 Regression Module provides a 'multinomial logit' option that allows the modeler to select factors and parameters and provides all of the expected Wald, pseudo R Square, log likelihood and classification tables. Users beware though...this is not the option to use for a conditional logit model that allows for choices among the alternatives (as in the famous Nielson yogurt panel data). The appropriate algorithm for the conditional logit (CoxReg) is part of the Advanced Statistics 19 module. This demonstrates that a selected Attribution Modeling approach is also in part determined (and limited) by the technology available in to an organization. Other options include Limdep, probably the most widely used (by academics, less so by marketing analytics practitioners) and most flexible. Stata also provides another alternative. SAS has two options, one is a proportional hazard modeling PHREG procedure, that provides a more limited modeling option. SAS has also rolled out a new MDC procedure that provides a robust new option for users, that looks very promising. Here's an interesting forum to follow and post questions on appropriate modeling approaches for business problems, and Indiana University has a nice statistics site that ranks high in a lot of search engine queries related to marketing models and analytics.

Game Theoretical Semantics and Segmentation, Core Analytics at VatorSplash, BrandMeter Trademark Updates

2010-09-12T06:43:00.000-07:00

It's Fall, and the blog is officially back from summer break. Just a quick blog entry to kick that off and post some updates around Core Analytics, the BrandMeter and add a few interesting links to the mix. Core is competing in the VatorSplash competition and currently holds the #14 spot in a first round selection that may ultimately lead to participation in the Sept 30th Pitch competition. If you haven't had a chance to cast your vote, please take a moment to do so now! Core will speak at the September 22nd 6pm Chicago Web Analytics Wednesday, where other local companies will also present brief overviews of other social media related topics at District Bar with cocktails and networking. Core Analytics now sponsors and organizes the Southern CA Social Media and Analytics Meetup as a west coast counterpart to the original Chicago-based social networking group, and sponsored the first 2010 Santa Barbara Web Analytics Wednesday earlier this month at Blue Agave. Final wrap-up news...Core Analytics is competing in the 2011 Burton D. Morgan Business Plan Competition at Purdue's Entrepreneurship Center (the 3rd oldest in the world) and is currently in the process of expanding their management and development team and advisory board.

More updates specific to advanced consumer behavior loyalty-related variables, semantic and segmentation components of the BrandMeter, as well as two new vertical prototypes (in addition to supermarkets) will be released in early October. Finally, Core's trademark process of the BrandMeter(TM) is moving through a final phase and we expect this to complete by the end of October. More social media-specific blog entries to follow!

Recap of SemTech 2010 and Core Analytics June 2010 Updates

2010-07-04T06:20:00.000-07:00

June was one of those months that seemed to pass in a whirlwind...I had the good fortune to spend about a third of it in various parts of CA. Core Analytics attended and participated on the Social Networking panel at SemTech 2010, and also stopped by the San Francisco WAW hosted by ObservePoint.

SemTech 2010 was an excellent week-long event focused on semantic technology applied across industries. The event kicked off with a welcome speech by Dave McComb of Semantic Arts, Inc. on the topic of 'What Will We Be Saying About Semantics This Year?' It was really interesting, the last time I was in San Francisco attending a technology conference was in 2000 at the Personalization Summit, which featured Geoffrey Moore as a keynote speaking talking about "Crossing The Chasm". McComb referenced the Chasm in his talk, stressing that adoption of semantic techniques was still in the very early stages of public adoption. The recommended approach to incorporating semantics into business processes involved small wins, in the form of individual, specific projects in niche sectors that demonstrated value, and in this step-by-step process an enterprise-wide appreciation for the value of semantics might be won.

Now in it's 6th year, what was traditionally an academic and highly technical conference has become much more mainstream, and the fact that attendance was up by 30% this year demonstrates increasing awareness around semantic applications. There were also very interesting sessions around procuring VC as well as a PitchSlam where 8 tech startups in the semantic space made their business pitch to a panel of judges. Core Analytics, LLC presented in the Marketing and Advertising portion of the conference on the topic of the BrandMeter (TM) social media customer activation and customer loyalty solution, using supermarket data.

Coincidentally, the SF Web Analytics Wednesday was scheduled during the same week, and the partners of Core Analytics stopped by and met with various web practitioners based in the bay area at the B Restaurant and Bar (amazing venue).

SF Web Analytics Wednesdays and SemTech2010 on tap this week

2010-06-20T06:00:00.000-07:00

A very Happy Fathers Day to my Dad, and to all the other great Dads out there!

It's back to California this week, this time to the lovely Bay Area, to attend the Semantic Universe Technology Conference and attend the San Francisco WAW that's scheduled for June 23rd at B Restaurant and Bar and sponsored by ObservePoint.

SemTech2010 is advertised as the largest semantic technology conference in the world, with a five day focus on semantic technologies applied to health care, finance, government, publishing, marketing and advertising, life sciences, military intelligence. I'm honored to have a spot in the Marketing and Advertising panel discussion, specifically on the Semantic Social Networking panel and will do a short presentation on Competitive Social Media Brand Analysis using the BrandMeter (TM) solution.

Semantic technology applied to social networks is essentially text-mining of raw comment data to derive sentiment, influence, behavioral, categorical variables associated with the commenter and brand or topic being mentioned. All of the major players in the social media monitoring solution space are employing semantic technology to provide dashboard KPI's to their users.

It looks like a great week, we'll post an update mid-week from both events, as well as a video of the BrandMeter presentation. Sorry to miss the Mobile TECH Cocktail 2 event in Chicago this week though.

Quantivo, CRM Metrix and Eric Peterson at the June 9th Chicago WAW

2010-06-10T09:00:00.000-07:00

The group met at Hard Rock Hotel for the June Chicago Web Analytics Wednesday, some 50 web and digital professionals turned out to participate in the event that was generously sponsored by CRM Metrix and Quantivo. We had a surprise visit from Web Analytics Demystified author and WAW founder Eric T. Peterson to further add to the evening.

Eric Peterson started the talks off introducing the Analysis Exchange , an effort aimed at providing students and the community at large a broader exposure and opportunity to actively participate and gain expertise in web analysis, while simultaneously providing Non-profits and NGO's cost-free quality web analytics work. I checked it out and signed up as a mentor, that's my profile below. The links are embedded, to register your Non-profit or NGO, to register as a web-analytics expert mentor, or to register as a student wanting to gain hands-on experience with web analytics tools and practices. It sounds like a really a nice way to participate in the community and give back. As a mentor, you can browse open projects based on your availability to donate time.

The next speaker was Paul Patterson from Quantivo who introduced the book "Behavioral Analytics for Dummies". All of the attendees were provided a link to download the e-book, or an email to request a hard copy. The talk focused on Quantivo's approach to behavioral analytics as a way to drive customer satisfaction, growth, retention and profits. One of their focuses is on customer segmentation, and they have a new release of a tool called Quantivo4 that allows users rapid access to a rich supply of perspectives and views of their customer and prospect bases. Quantivo has embraced cloud computing and attributes the speed of web analysis now possible in their offerings in part to the incorporation of this new technology.

The next presenter was Heman Patel from CRM Metrix, who gave a presentation entitled "Is Social Media the Almighty Customer Touchpoint?" The focus of this (sophisticated) presentation was looking at optimal marketing and media mix to maximize ROI through customer targeting. Here's a very interesting example of an approach that CRM Metrix used in the CPG industry to measure brand affinity using their EZ Views platform, where brand affinity was found to be 22% higher with 'email' members. The case study describes the approach CRM Metrix has to leveraging the power of the web to gather qualitative and quantitative data to measure brand attitudes and affinities and develop strategy accordingly.

The turn out was great, as always a fantastic mix of industry professionals from various verticals, MBA grads and students from the area's top schools, lots of audience interaction, Q and A, networking and a good time.

PR+MKTGCamp and Web Analytics Wednesdays Chicago Events

2010-06-09T05:30:00.000-07:00

I had the opportunity to attend, and moderate at a session on metrics and analytics at yesterday's PR+MKTG Campfire event held at the Loyola University Center. (The Chicago Web and Social Media Group is a partner.) It was a really nicely planned and executed event that assembled and sparked conversations between social media practitioners from an amazingly wide array of companies that include Sears Holdings, OfficeMax, Kraft, American Red Cross, Google, Alterian (developers of SM2 social monitoring software), Threadless (this is a really neat e-business where graphic designers send in their submissions, visitors to the website vote each week, and those rated the highest become t-shirts), Chicago Tribune, Crain's, CME, Allstate, MillerLite/Coors, Abbott Labs, Motorola, McDonald's, Boeing, CriticalMass, AON, Ogilvy, Razorfish, and then many of the area's small businesses (who probably represented at least half of the event participation). The event's organizer Dan Greenfield effectively brought together those practicing on the cutting edge of social media from a powerhouse of companies, and created an environment with smaller moderator-directed breakout sessions on topics around 'Social Media and PR', 'Impact and Strategy', 'Metrics and Analytics', and 'Strategy and Execution' and then brought the entire large group back for a recap of what happened in individual sessions. It was a great opportunity to learn, share knowledge and network.

One thing that stood out in our session was the small percentage of smaller-to-midsize companies who are currently utilizing social media monitoring software. The larger companies are using popular commercial solutions, the smaller groups tend to be using manual comment moderation and response methods. ViralHeat (Core Analytics is an Affiliate Partner) was introduced as a low cost (plans start at just under $40/month) solution for small businesses, as well as open source solutions such as SocialMention.com and the Facebook Insights tool. The most popular commercial solutions are Radian6, Buzz Metrics, Scout Labs, Omniture for a web analytics solution to track related tagging, and traditional PR-specific software such as Cision and Vocus. Our in-dev customer loyalty social media measurement tool called BrandMeter, and it's associated social media data mart and dashboard reporting around aggregated individual activation and loyalty metrics was also discussed. Metrics and areas of interest include influence, value, engagement, sentiment, channel profiling and strategies, sentiment accuracy. The conference also included an enjoyable presentation by Tom Yorton (CEO of Second City Communications) on the use of humor in social media and marketing.

This is a big week for social media gatherings, all are welcome and encouraged to attend the Web Analytics Wednesday event at the Hard Rock Hotel starting at 6pm, sponsored this month by CRM Metrix and Quantivo. Drinks, snacks and presentations (and hopefully a TV or two nearby to keep an eye on the Hawks game!) will be featured. Hope to see you there!

Attitude and Persuasion Theory Applications in Social Media, Consumer Loyalty, and the BrandMeter

2010-05-27T05:39:00.000-07:00

The earliest literature around attitude measurement emerged in the 1920's, when Thurstone used an elaborate system of psychometric methods to measure attitude. So elaborate in fact, that it never came into wide usage. Since then social psychologists have been developing various methods to measure attitude, understand how these attitudes developed, then determine appropriate persuasion techniques to direct consumer perception and behavior.

The general objective of advertising and marketing is to influence consumer behavior along a designed set of outcomes: opening an email, clicking on a link, placing an item in a shopping cart, fulfilling a purchase, becoming a repeat purchaser. There are various approaches and views, differing in how much time and messaging is needed to change consumer attitude. The cognitive response approach assumes any attitude change requires a high level of interaction on the part of the consumer to happen, the elaboration likelihood model finds that it can occur with much lower levels of consumer engagement. The literature related to these models discusses the expected persistence of attitudes that are effectively changed through messaging, based on how newly formed the attitude subject to initial change is. This persistence of attitude change, probability of attitude decay over time, measures of resistance to further outside persuasion and likelihood of behavioral consistency are all consumer psychology elements that make up that elusive notion of Customer Loyalty.

The BrandMeter is a consumer measurement loyalty tool that incorporates these aspects of attitude measurement captured and stored through text-mining of comment data combined with web behavior to create a current state, expected and changing over time metric to measure individual and aggregate loyalty shifts around brand. One interesting outcome of this approach is an ability to understand weekly shifts in consumer loyalty across brands in a specific vertical. Another is to identify segments of at-risk consumers who have recently experienced a positive brand-related attitude decay for either consumer loyalty and retention efforts on the part of the primary brand owner of the at-risk segment, or by the competitor with customer conversion targeting.

Since all of the data is freely available for harvesting, the example highlights the importance for each brand to understand their position in the social media space as well as their competitors, and maintain an ongoing social media data mart that measures both their own and their competitors. One interesting point that makes this easier for social media practitioners, is how this data is collected. The social media monitoring programs can be set up by keyword, and if this is set up to collect not only specific brand mentions but also by product category, a company will also capture any newly emerging competitors gaining share in the space and this will be highlighted in updated competitive landscape measurement and monitoring.

This approach can also be used as a measure of social media ROI. In understanding the complete ROI of any marketing campaign, the focus should go beyond the flat dollar amount attributed to the campaign to understand the indirect and future expected value of the response customers (this comes out of a great conversation I had recently around accurately and fully measuring the ROI of a paid search campaign). These measures of persistence and resistance to influence (conversion efforts) are captured in both attitudinal and personality variables. The indirect value of a responder is associated with expected lifetime value, the expected lifetime value of those in the responders sphere of influence, and perhaps some inherent propensity-towards-loyalty score associated with that customer. That leads into some interesting metrics around the spurious or intentional nature of the loyalty itself.

Customer Mood Identification, Competitive Social Media Datamarts and Social Media Advanced Analytics

2010-05-15T08:33:00.000-07:00

This Saturday morning finds the partners of Core Analytics convened in the newly converted office, surrounded by tens of thousands of brand comments, lines of Java code, XML files, text mining classification code, freshly baked blueberry muffins, and coconut coffee. Today's assignment? Pull together a week's worth of social comments from across the blogosphere related to the NY Yankees, Boston RedSox, and Chicago WhiteSox and do a comparison of sports brand comments. How do Yankee, RedSox and WhiteSox comments and sentiment differ? Team rankings, historical outcomes, most recent game scores and team news (from Google Insights) and weather data are added to the analysis file and tied to comment timing. Is it possible to separate out fans vs. not, and hone in on how the profiles of an online Yankee fan differs from a RedSox fan? What implications does this have for sports marketing based on brand? We'll be presenting the results in tomorrow's blog as Teams A, B and C, and ask you to determine which is which in a blog contest, based on the analysis outcomes!

ViralHeat is a pretty nifty social media monitoring offering, provided you are willing to get comfortable with collecting, manipulating, and merging an unwieldy amount of data you use an API to access. Social Media Software comparisons don't seem to be focusing yet on one of the key functionalities these tools either have, don't have, or do have but with prohibitive technical, size or cost limitations...namely raw comment level data extraction capabilities. Without that raw comment data, advanced analytics can't really happen. Without text mining, raw comment data isn't terribly useful. Where it goes from there follows the traditional truth around analytics and statistics: It's still as much of an art as it is a science.

We'll get to the baseball data shortly (publishing this contest tomorrow) and continue along with the Supermarket data theme and competitive analysis from last week. Several new mood and personality type variables have been introduced into the analysis data set through further text mining classification. The general idea and usefulness of these variables is as follows.

Comments generally express some type of opinion, action or emotion. Emotion is related in a general way to what software of this type is labeling as sentiment (positive, negative, neutral or unknown) and further work was done on our file to include emotion-specific variables that identify the presence of various emotions in a comment (happy/delighted, angry/disgusted,doubtful, scared, sad, impatient, etc.). The text mining around identifying these emotions is an extensive process identifying combinations of words that are generally used to express these emotions. Take happy, in the following simplified example. A Tweet or Blog post that expresses happiness or delight will not only include the brand mention that pulled them into the analysis file
but also with specific terms (happy, great, amazing, super, wonderful, thrilled, delighted, jazzed, psyched, etc.) in a cetain structural position relative to pronoun and noun. This identification process uses a rendition of the semiometric approach to classify comment emotion. Comments may be labeled with more than one emotion, so our analysis file has extended flag variables that indicate the presence or lack of each. Eventually gradations of emotions will also be possible for flagging. The final analysis file is aggregated up to the customer level with derived variables around mood (emotion) and timing of comments, and in this way mood and personality related variables emerge over time. It's important to stress that while the data used in this analysis only represents a month of brand data, it represents a process companies would go through in collecting 30-day intervals of data and appending the new month's data onto the data set, rerunning the code and doing this at an aggregate level by customer id by social media site.

Eventually this analysis data set becomes a social media data mart that stores all historical social media comment data and provides a view over time of specific individuals related to brand and shifts in mood and sentiment towards brand can be detected with trending techniques. This data will allow measurement of the growth of social media industry market and mind share, and increases in certain segments of customers and their associated mood, sentiment and preferences, instead of comment level analysis it becomes possible to have individual (by social media site id) level analysis, and determine moving average metrics that capture satisfaction increases and decreases around customer service, products, experience, pricing and so forth. It also allows for the ongoing capture of individual preferences.

For instance, if you posted 'Enjoying a frappucino at Starbucks' 3 months ago when you were at Kroger, Target or Dominick's, you would be flagged as a 'coffee drinker' and 'Starbucks fan' in the database. When individual banner ad or social media marketing campaigns are launched, where appropriate, yours might feature a 'half price frappucino' offer, and someone who extolled the virtues of lime flavored Tostitos will have a highly targeted '1/2 off bag of Tostitos' offer. The social media comment data over time morphs through this collection and derivation to be a wealth of individual mood, personality and preference data that can be used for segmentation, and predictive response/churn/next-best-product modeling. Many of the social media monitoring tools on the market limit back-in-time comment retrieval query responses to 30-days, at best. This suggests that ongoing regular download, as well as monitoring, of social media comments around brand be practiced to build an ever-increasing time series collection of comments that wouldn't be available otherwise. Currently Core Analytics captures and stores 40 major queries across retail, beverage, supermarket, sports team, utility, electronics, on-line dating, telecom, financial, insurance and health insurance brands.

A supermarket wishing to capture market share from competitors also has this public comment data around competitors brands. It's possible to find customers who overlap between two stores and target them in brand conversion campaigns, using this highly individualized data. For example, a traditionally happy individual (Shopper 1) who has always favored Supermarket A has suddenly in the past two weeks shown an increase in the number of angry comments around their supermarket service (identified through the combination of derived emotion variables and content category (service, experience, price, product, other), and Supermarket B knows this because they're not only maintaining a social media data mart on their own brand comments, but have a competitive data mart as well. Shopper 1 is at risk of churn (at least temporarily since upset), and Supermarket B gets an alert that a potential customer target is in a loyalty crisis window and time is of the essence to make a conversion effort. Supermarket B looks at the freely available preference data collected over time by Shopper 1 by social media site, and knows that Shopper 1 likes wine, and is upset about checkout service, so one conversion treatment strategy might be an @ Tweet or message from SuperMarket B saying 'Introducing SuperMarket B's free wine discount program, members enjoy express cashier line access'. (Kind of like priority boarding lines through airport security) Since the data set has so much information, this type of individual targeting by category and specific preferences becomes possible even in an automated alert and recommendation engine type scenario.

Social Media Competitive Analysis: Supermarket Chain Example

2010-05-07T08:51:00.000-07:00

This is the first in a series of advanced analytics applied to real-world social media data examples. The first industry series will revolve around recent (actual) publicly harvested data for 3 large Midwest supermarket chains. The names have been anonymized, in an attempt to be responsible about using a one-month data sample and publishing what may be perceived as general conclusions about corresponding social media activation and competitive profiles of real brands. The findings are based on actual data, but a true analysis used to drive strategy decisions should be based on a larger sample of data to avoid seasonality that may be represented in a single month snapshot of social comment data. The legality of referring directly to brand comparisons is unclear, as the data is publicly available to all, to duplicate this analysis. For the moment, this study focuses Supermarket Chains A, B and C. Data collection is occurring to support similar advanced analytics studies on large US financial, insurance, sports, beverage, electronics, retail brands, and a much larger collection of U.S. brick and mortar and online supermarket data that will provide the basis for a whole-industry study that will be forthcoming as blog topics and vertical examples on the Core Analytics website.

Data was collected over a 30-day time frame to capture all online comments related to three supermarket brands. The supermarket chains in question were selected because of their regional focus, similarity in number of stores, states represented, and annual revenue. The data was then processed through a rigorous text-mining process that identified actual instances such as purchase and returns, and further text mining identified consumer comments that reflected presence of children, price sensitivity, and preference for organics. This data was then aggregated to a customer id level by media type (Twitter, Blogspot, Forum, Generic Blog) to derive engagement, sentiment, and brand related patterns. Reporting of results comes from the individual comment files as well as the aggregate files to offer a variety of representative comparison metrics.

The three supermarkets had fairly dissimilar basic social media representation. Chain A had twice as many total brand comments as Chain B, and Chain C has comparatively no online social media representation. Aside from the difference in volume, Twitter is the media channel representing the largest number of brand comments, with forums and blogs contributing a significant amount as well. YouTube had very little representation and Facebook comments are not generally available in this type of publicly harvest data sample so the handful that filtered in were removed from the data set here. Chains A and B are similar in terms of media representation, however Chain A has a larger percentage of Twitter (light blue), and Chain B has a larger share of 'Other' which includes comments from LiveJournal, FriendFeed, WordPress, Flickr and YouTube. Chain C has over half of brand comments happening in online Forums (red).

The total number of comments that reflected 'Price Sensitivity' was surprisingly high. At least 1/4 of all comments in all Supermarket Chains were related to cost, coupons and prices of items. Chains A and B had similar representation (23 and 26% respectively), while Supermarket Chain C had a much higher percentage (34%). Based on the target markets of these stores, Chain C is perceived to be a 'value shoppers' store. Organic shoppers had similar representation at between 2-3%, which was not surprising as none of these chains are branded as 'Organic' stores.

SEO and SOCIAL MEDIA OPTIMIZATION

2010-05-06T16:29:00.000-07:00

The following article is an abstract of the May 2010 website magazine around SEO and SOM. Social Media Optimization (SMO) is a content based approach, meaning a focus around being proactive and responsive to individuals. Some best-practices around SEO include frequently updating comments on social media sites and focusing on the audience relevant content.

The May 2010 website magazine has an article regarding SEO as it specifically related to a rapidly transforming social media setting. As with every new endeavor, a plan is required for successful execution. In other words, it's necessary to understand the audience, and target potential customers. A research report (by SEO technology and service provider Conductor) about search visibility and optimization effectiveness shows that most companies don’t do that well regarding SEO. Most enterprises have limited natural search visibility into the impact of their targeted keywords. Social media is real-time. Most of the consumers are not aware that social is making a path into search. So what’s the solution of applying SEO for generating 1st page rankings to social media, and vice-versa? It’s all about content and how this is developed and shared.

In order to develop content for users it is essential to become familiar with user interactions (behavior) and how information around brand is shared and published within the social site network and between peers. Using social media monitoring software (like Radian6) allows companies to become familiar with brand conversations and primary brand influencers who initiate those conversations. The Amazon.com website relies heavily on leveraging content generated by the web user, such as product reviews and recommendations.

What about other types of websites, like informational and service-oriented? Adding an information or education section to an existing website and publish linkable content related to company and industry vertical works well. A recent study of 2,168 HubSpot customers showed that businesses which were blogging at least 5 articles in the last 7 days attracted 6.9 times more organic search traffic and 1.12 times more referral traffic than those businesses who are not not blogging.

Creating awareness is another important aspect of social media optimization. SEO is about discovering influence of the social media channels and online environments or content via search. How best to share and distribute brand content? There are several tools out there like ShareThis or AddThis, Facebook Connect and Google Friend Connect. As with SEO, the use of targeted keywords applies as well to social media. Keep in mind, social media content should geared to a specific audience. Including keywords merged in your content that contain links pointing to the company website and/or other social media sites is recommended. Another tip is to anchor text in the links using target keywords. SMO focuses on a strategy built around links, and linking all your social media site contents together and back to the primary site. Finally, the article recommends the best way to generate links is from the primary website (content).

It's Morel Mushroom hunting season in Parke County!

2010-05-06T07:56:00.000-07:00

Off topic from the ongoing segmentation and social media activation presentation project...which is turning out pretty interesting and results will be published shortly both on the blog and on the website, but it's morel mushroom hunting season and that only comes around once a year.

Parke County is located in west-central Indiana, just 3 hours from Chicago. About 2 hours into the ride, depending on your route, the flat prairie vistas and ever-expanding wind farms around Lafayette give way to gentle rolling hills, and trees, state parks such as Shades (with it's eerie name and atmosphere) and Turkey Run (which boasts some of the best hiking trails in the state, with paths winding along the popular canoe/kayak destination of Sugar Creek, as well as a wonderful lodge and restaurant). Parke County is Amish country, it's hilly, green and has a wonderful 100 acre recreational lake called Raccoon Lake. It's also the Covered Bridge Capital of the World. Charming festivals happen throughout the year, the most popular is the Covered Bridge festival every October, but the list also includes the Maple Syrup festival, Watermelon Festival, Cornbread Festival, and the Mansfield Mushroom festival which was held two weeks ago celebrating the annual tradition of morel mushroom hunting. The area is lovely, a visit there is a step back in time, and well worth the trip. Here's a great place to stay if you're planning a trip.

Back to topic, this morel mushroom hunting is taken quite seriously (also by me). Check out this interactive morel mushroom sighting map one avid fan updates and has built a business around. Here's an online club for morel hunters. Here are some online recipes for the lucky finders. Here's a rather hysterical (slow, but that's the nature of the sport) video of morel hunting in Oregon, and here's another Indiana morel hunt set to Stevie Ray Vaughan. Last year, I found 10 of them randomly in the backyard of my lakeside trailer down in Rockville. Here's an important link, if you're going to try your hand at mushroom hunting this year, to identify and avoid false morels, which can be poisonous. Happy hunting!

Advanced Analytics and Social Media: Response Modeling with Sentiment Data

2010-05-02T06:39:00.000-07:00

I'm sure text mining analytics experts have been doing this for years, but it's always fun to approach a data problem with a variation on a classical approach, combining this with new technologies and see if the framework is transferable. The process of building a response model with social media data involves downloading publicly harvested brand related data from the web, text mining of the comments to assign instances where customers purchased, assign their id with a 'customer' flag at the aggregate level, then assigning the others with 'prospect' flags, and running some classification models to see how social media customers differ from prospects, and what social media patterns and behaviors differentiate those who purchase vs. those who do not. Presumably, such a model could then be used to score the brand file to identify look-alike commenters for targeting, as well as prioritize customer response to social media events based on customer lifetime value and loyalty metrics.

Another hypothesis that is expected out of this analysis is a snapshot of how two brands social media customer bases differ. That is, how is Brand A's online social media customer (as defined by 1 or more prior documented purchases identified through text mining, and eventually this can be extended down to include product category level purchase) different from Brand B's social media customer. How does this compare to offline target market and historical customer segmentation. Does the typical active social media customer for Brand A resemble the typical Brand A customer? Is Brand B still a competitor in the social media channel, or is there an unexpected Brand C that is competing for Brand A's social media customer share?

Initial work on this process, starting with the data collection, has already led to a few interesting learnings. Freeware vendors the Archivist and SocialMention.com are real-time search focused. If one is aiming at historical mining of comment data, they are not suitable data collection tools unless the analyst is willing to do daily manual search and save brand extracts. ViralHeat (specifically the business plan) offers a lot of data that is categorized into micro blogs, Facebook, and other types of online forums with data extraction capabilities, but it is necessary to plan ahead with the data collection as files and custom queries are captured from the point at which the profile is designed, and does not go back in time. Radian6 does allow for a 30-day rear view mirror of comment data history, as well as extraction of the raw data, however the extraction process is prone to occasional failure and it sometimes takes a few attempts to get the file. In the version I last worked with, there was a 5000 comment limit with extraction files, so some creative time windows were needed to get an entire 30-day window of brand comment data with the extraction restrictions.

Here are a couple of unrelated but interesting links on mobile applications that were covered at a recent predictive analytics summit. Andreas Weigend, prior chief scientist at Amazon.com, had a presentation (see midway down in the schedule) on the New Data Economy and mentioned RedLaser, an iPhone app that allows consumers to scan barcodes while shopping and compare prices elsewhere on the item, and Boy Ahoy that combines online dating with geolocation to score compatibility with those in close proximity. This looks like it was a great predictive analytics summit. There's another related event coming up in Boston this month, the 6th annual Text Mining Conference.

Open Source Apache Software: Hadoop, Chuckwa and Pig

2010-04-25T06:29:00.000-07:00

Open source software seems to be gaining a lot of traction. When faced with an inquiry about what types of open source software one is familiar with, the standard answers for a data mining, predictive modeling and analytics professional used to be R and RapidMiner. R, or 'The R Project for Statistical Computing' is essentially a Unix-based collection of free integrated software that supports data collection, merging, aggregation, manipulation, analysis, regression and time-series, classification and clustering, graphing and reporting through a processing language 'R'. In this sense, it operates similarly to SAS. RapidMiner, generally considered easier to use than R, is an open source data mining toolkit that provides many of the same functions (different algorithms) as well as a text-mining capability.

With the proliferation of user-generated comment data in social media and micro-blogging, an increased emphasis on map-reduce style open source applications has developed. Apache Software Systems offers Hadoop, designed initially by Doug Cutting who was working on a search engine project, and named it for his son's stuffed elephant. Hadoop has been in use at a lineup of search, social networking and digital publishing powerhouses such as Yahoo, Facebook, Twitter, New York Times, eHarmony, Ning, IBM, and LinkedIn since 2008.

Hadoop offers users the ability to store extremely large datasets through a distributed file node system, that replicates data and leverages cloud computing. The system supports click stream analysis, marketing analytics, data mining, image processing, XML message analysis, serves as a web crawler, and text analysis.

Exploring the Google Public Data Explorer Labs

2010-04-06T03:44:00.000-07:00

An interesting Google Labs project provides access to data and a visualization tool for data exploration. The datasets currently available include data from the World Bank, OECD, and other public data realted to uemployment, consumer prices, retail, wages, personal income, education, mortality, and disease.

The charts and output of the visualization tool can be embedded in presentations and automatically refresh with updated data. The Retail and Food Services industry, according to this data, started to plunge in months leading up to December of 2008, where it reached bottom and started it's growth pattern again. A related economics blog cites similar monthly economic data trends in February of 2009 that indicate the economy might have already hit bottom.

Green Parking in Chicago and Pricing Optimization Processes

2010-04-01T04:29:00.000-07:00

Today's posting is a departure from the web and social media focus, just some commentary about a green-minded parking garage in Chicago I park in and some interesting price and discount variations they seem to be randomly testing.

The Greenway Self Park Garage was developed by Friedman Properties, Ltd., to incorporate several green features. For instance, all of the exterior lighting is powered by the six Aerotecture wind turbines, and any excess electricity is donated to the grid. These turbines are expected to generate between 10 and 15 thousand kWh annually.

Energy-powered ventilation has been replaced by natural ventilation engineered glass panel spacing, and energy-efficient glass used in the construction will reduce energy demand for the lower retail level spaces. The parking garage has rain cisterns to collect water to be used for maintenance on the garage. They also provide a bike room, plug in stations for electric vehicles, and have the lower levels reserved for hybrid vehicles. Recycling containers are found throughout the structure. In addition to all this, they're offering real parking deals in this expensive River North business location.

Back in mid-January I fortuitously happened upon a waving man wearing a placard advertising $5 all day early-bird parking to celebrate the Grand Opening of the Greenway Self Park Garage. All that first week they provided $5 parking and complimentary coffee and bagels. Their website also offers the opportunity to sign up for discounts. The second week the introductory price went up a couple dollars, and then attendants could be found standing next to the parking payment machines passing out instant discount coupons. If you asked for more, they cheerfully provided a handful of $5-off single use coupons. The new $3 off coupons are good for 30 uses at a bright, modern, centrally-located garage that is setting a fine example in their efforts to sustain the environment. Word to the wise, Early Bird parking ends with 8:30 entry. I found out from experience that 'one minute matters' when I arrived at 8:31. But, as the attendant charmingly pointed out, 'Oh, but you get lots of coupons'. I wonder if hybrid owners get better discount coupons, or perhaps none at all since a highly green-minded segment of customers who have already demonstrated their support to green initiatives with a hybrid car purchase might also be willing to pay a premium to support a green parking structure. As the coupons are handed out by attendants on an individual basis, individual personalization can be easily managed. It's a nice garage and I recommend it.

Here's another article that talks about the garage's sustainability and design.

Micro-review of the Facebook Era: A Good (Strategy) Read

2010-03-28T06:30:00.000-07:00

The Facebook Era, was written by Clara Shih (who wrote Faceconnector, the first business Facebook application, and leads the salesforce.com partnership with Facebook) and has an associated blog where you can sign up for daily email tips on social media marketing. The book is required reading for a Global Entrepreneurial Marketing course at Stanford and a social media course at Harvard Business School.

It's a comprehensive how-to manual for social media marketers at both small and large companies, with an emphasis on Facebook and Twitter. I found Chapter 8 particularly useful, as it includes a step-by-step guide into developing a social media strategy. The recommended process starts with identifying objectives. Does your brand want to conduct market research, improve customer satisfaction, cross-sell products, increase positive WOM, build and expand brand presence? Identifying the objective at the outset and a prioritization for this (and others) list of social media goals will determine the next step in the process. There's a focus on measurement and a reminder to always consider the interaction from the customer point of view, and what they are hoping to gain in the interaction (express emotion, provide constructive feedback, associate their identify with your brand, derive a sense of utility through helping others, expand their network, indirect financial benefits, bond over product experience with others, etc.) The next step is to locate 'unsanctioned' communities related to your brand. Even if your corporate strategy is in its early stages, chances are there are brand communities that already exist related to your brand and it's important to understand their nature and focus, and incorporate these into the overall plan. In general, it's recommended to let these communities continue without interference as regardless of the nature of the conversations, there is information to be mined from these interactions. The rest of the chapter gives specific examples of identifying existing social mention of your brand and deriving insights through the use of data visualization tools. This follows with a stepwise guide on setting up a Facebook page and customizing pages for different audiences, then connecting the new official community to exisiting brand communities and through offline channels.

A Consumer Psych Approach to Evaluation of Consumer Social Media Engagement

2010-03-27T07:26:00.000-07:00

The blog theme today continues with the proposal for a segmentation-based approach in next-generation social media monitoring tools, to incorporate an additional customer stage measure related to the concepts of 'online brand loyalty' and 'social media adoption'.

Consumer psychology views consumer socialization from a perspective of cognitive and social stages, and this approach may easily translate into a tool for identifying stages of consumer affiliation with a brand across social media forums. One popular socialization framework, Piaget’s theory, separates the process into four agerange-based stages of cognitive development. The first stage 'sensorimotor' refers to birth through the age of two, before moving into a 'preoperational' stage from two to seven years of age and characterized by response to easily observed aspects in a single dimension. The third stage, 'concrete operational', is where individuals can assess several dimensions and is capable of abstract reasoning, before entering the fourth operational stage where complex assimilation of the concrete and abstract occur.

Distinct stages of development occur in all consumer socialization research. Another approach includes just three steps of information processing: strategic, cued and limited processors. Other more social development-related research focuses on non-cognitive aspects such as morality, altruism, and impression formation. A scale that identifies site-specific and aggregate development stage variables, possibly substituting a tenure-volume interaction variable for the age indicator used in the models above, will separate customers into one of four specific cognitive stages on each site. This categorization will reflect levels of individual sophistication and provide a framework for marketers to target customers for assistance and education related to the more complex site features as well as different social media dimensions.

Social Media Segmentation, Social Media Loyalty as Two More Dimensions in Social Media Activation Measurement

2010-03-20T13:22:00.000-07:00

Proponants of market segmentation typically hold the view that business analysis improves by first segementing the group in question into natural segments, and then performing analysis. Otherwise so many of the nuances of how each segment buys, behaves, interacts, responds, votes, expresses dissatisfaction, and exhibits preference are overlooked when only focusing on consumer events from a top-down perspective. Social Media takes the need for segmentation to a whole new level, since so many variables are at play. These refinements to social media management solutions will allow for greater flexibility around measurement and management at the segment and social network level. The Core BrandMeter is one solution that attempts to pull it all together.

Historically, the primary marketing channels include direct mail, call center (inbound and outbound), web, search, email, catalog, and seminars/conferences, depending on what industry you're looking at. Recently a new channel has emerged, social media. Eventually, because consumers interact so diversely across the major social networks, each will likely be considered a channel in its own right.

The Groundswell blog authors have segmented social media users into six distinct groupings: Creators, Critics, Collectors, Joiners, Spectators and Inactives. Paul Bannon has an interesting blog article that links to Forrester's Technographic Ladder. Real-time views that capture who's mentioning your brand and what category they fit in are one of many dimensions that later stage social media software will help identify.

Segmentation around social media is best viewed as a multi-dimensional endeavor. Other dimensions of interest including personality and navigation patterns, current stage of social media adoption, real-time buying and crisis signals, social network-specific activity and social media loyalty metrics, as well as moving average windows of influence and intent, at a social network-specific level would make the insights delivered by current offerings even more actionable.

SEO and Social Media, Social Media Literature, Social Moves to Mobile

2010-03-15T07:23:00.000-07:00

The Monday morning iMedia connection updates have arrived, with two of particular interest. Kevin Ryan wrote a nice piece on the 5 new can't-miss SEO trends which emphasize the reality of real-time search, and it adheres to a different set of SEO rules. Some tips to stay on top include using keywords in tweets, make the tweet text relevant, and web address abbreviators matter (this means when tinyurl or bit.ly offers the option of creating a personalized address over using the long auto generated version, take advantage of renaming the address related to keyword and message, and exercise the advanced bit.ly tracking options on the new url). People are focusing more on social networking sites than browsing the web, so incorporate social media data and sites in SEO efforts. The article also reminds us that what we're working to optimize is now also competing with content around video, images, and apps.

According to MediaPost.com today (this is an amazing resource), a new Comscore study indicates large growth of social networkers accessing social sites with their mobile and smart phones. Some stats around this: Facebook mobile access grew by 112% in the last year, and Twitter saw a 347% increase in the same period. In January (this year) over 25 million mobile users visited Facebook.

This weekend found me browsing Borders checking out the current social media literature landscape, the picture above was just the small stack that made it off the shelves for perusing. The 3 that made it home from this batch (selected from an analytics focus) were Groundswell, the Facebook Era, and Social Media Marketing in an Hour a Day.

Personality Psych Applications for Web and Social Media

2010-03-07T08:05:00.000-08:00

Sunday's to-do list includes updating the blog with a new entry, so I've gathered a bunch of things here for inspiration. One of them is a lit review I did a few months back, related to a current area of interest in applied psychology related to social media and online communities. I started research into this topic during a consumer behavior class at Purdue last semester, and put together the attached flow chart approach that is recommended as a visual aid in putting together a problem, hypothesis, course of research, construct development, viable data to prove or disprove theories, and the areas of focus for a literature review to see what other researchers have already found that's related to the topic.

Experience working on a few consumer activation projects related to health care in terms of patient activation to self-manage wellness, disease and insurance plans, suggests that this concept of consumer activation would seemingly apply across many industries. With this in mind, I began some research into scale development approaches to developing a tool to efficiently measure the level of customer social media activation related to brand on social media and other social network sites, that would result in measurement of information that is relevant and actionable. The value of such a scale rests on the assumption that companies will be able to employ it as a tool to understand their current and ongoing social media position relative to competitors, to plan how to best build positive brand recognition, acquire and retain high value customers, deliver cost efficient customer service and increase customer loyalty and value.

The scale is developed from the hypothesis that specific types of customer activation on social media sites can be regularly harvested from publicly available comment data on the Internet, then aggregated and categorized resulting in specific and relevant measures of customer activation useful in understanding general trends as well as individual target marketing selection metrics.

This proposed usage of the scale is closely related to the traditional field of attitude persuasion. Attitude theory and persuasion research focuses on involvement as a measure of consumer engagement, and related outcomes on advertising persuasion. The types of interactions consumers have on social media sites can be interpreted as a reflection of their level of involvement and attitude towards products and brand. The scale development is based on the premise that customers with higher volumes of positive sentiment comment and interaction, with higher reflected levels of involvement in products and brand, will be more likely to exhibit positive consumer actions such as responding to advertising and increased purchase outcomes. Studies have shown that differing levels of product involvement result in different persuasion outcomes, depending on the quality of advertising accompanying the differing levels of involvement. (More on this later).

To introduce a different topic, I'm cleaning up old newspapers and have an FT article of interest for comment. The Thursday, Feb 25th FT has an article 'A future alongside Facebook' that discusses leading European social network sites such as the French site Skyrock, which has lost 1/3 if it's dedicated user base to U.S. social networking-based Facebook. Many other European social networking sites have lost share to FB, including StudiVZ (Germany), Tuenti (Spain), Hyves (Holland). Facebook is syphoning share from these community bases in a way MySpace did not. Managers of online European online community sites are subject to strict privacy guidelines that Facebook is not required to adhere to, and they are unable to address this advantage. While Facebook polices are technically illegal in Germany, for example, the site is exempt from these restrictive rules because of safe harbour agreements between the US and the EU.

BrandMeter: A Social Media Activation Measurement Study

2010-02-22T04:56:00.000-08:00

It was difficult to select a title for the blog entry, since there are so many topics of interest to comment on. To start with, in my spare time I'm involved with an academic development project on a social media activation project, called the BrandMeter, and will chat more about this at the upcoming Feb 24th Chicago Booth Alumni Club show-and-tell Marketing Roundtable. (If you're not familiar with the alumni Roundtables, these are industry-specific regular knowledge sharing, discussion and networking meetings and one need not necessarily be a Booth alumni to participate in all sessions.) The BrandMeter is a social media activation measurement tool that serves as a multi-tiered scale to identify benchmark customer/prospect engagement levels along various customizable categories of social media interaction (brand sentiment, product information, customer service, product development, brand awareness, WOM, market research feedback, etc.) to start with a base level of activity and evolutionary progress at the outset of a social media strategy project, and perform ongoing and relevant measurement matched to strategic goals along the customer social media adoption evolutionary curve, from a segmentation perspective. The tool not only differentiates itself from a segmentation approach, it also looks at consumer activity and comment from a psychological and behavioral perspective to measure levels of 'activation' in addition to sentiment classification. Eventually the tool will also have attribution-analysis components developed through comment text mining and machine learning, in a later version. The BrandMeter is in beta now, therefore companies are currently being solicited for development purposes....specifically exchanging transactional data under confidentiality agreements in exchange for insights, measurement and reporting. Email if interested in participating or for more details.

Here's a great HBR article link that talks about regional popularity (see above map, a screenshot from the article) of social network sites, and implications for marketers to develop social network specific strategy to account for the demographics of their customer base and target market, incorporating these regional trends. For instance, MySpace is 'over indexed' in the South and some parts of the Midwest, and a high concentration of Facebook users are from New England, MN, VA, SD, and NJ. A case study of a Houston-based cafe chain profitably leveraging this business intelligence is featured in the article.

On another note, meetup.com just launched their new live-streaming video capability and this is something we hope to utilize for our next meeting. Last Wednesday we co-hosted the Chicago Web Analytics Wednesday at Wilde in Lakeview, and took some video footage of the event. Member presenters were on hand to talk about the Business Intelligence Social Media tool Sysamos, freeware options such as SocialMention.com, the Archivist, and Hootsuite. Other members shared web analytics measurement and strategy impacting case studies, and advanced bit.ly search features. Once we have member approval on the video we'll be posting that on the blog and meetup site to test out the standard site video functionality.

Marketing Applications of Biometrics and Eyetracking/ Attribution of Social Media through Textmining

2010-02-05T04:28:00.000-08:00

Sitting in a web meeting on a new laptop yesterday, clicking through the screens to get familiar with the technology, led me to the viewer video screen where I was surprised to see myself on the screen ('not sending'). Ideally there would be some kind of warning, indicating a 'you are appearing live on others screens, via the built-into-your-laptop-webcam' notification (and there most likely is).

That led to an examination of the unobtrusive webcam, and consideration of what type of online market research and general applications that a built-in retina tracking device on laptops and temperature/pulse receptors on keyboards or mobile devices might have, as well as face-recognition and expression identification where video is present in live chat, and what might already be in development to gauge and test physiological responses to messaging and offers. (Been doing interesting reading of others research on the general neuroscience aspect of reactions related to marketing recently that connected the two). Interesting links here about eyetracking technology and how webcams include eyetracking devices. Here's a link to a downloadable api file to install EyeTrack on your computer and settings for face and eye detection and pupil settings. The keyboard and mobile pulse and temperature detection technology is already popular in health monitoring applications.

Biometric technology is popular in security applications , and speech technology has been in use for a while to identify patterns, changes in tone pitch, as well as text mining of call data for segmention of customers based on personality type, in-conversation emotional response This data is used to improve outcomes of customer service scenarios. Similar technology likely exists for online chat customer service channels, where speed and keystroke force (related to normal individual patterns), increased pulse, even narrowing pupils could indicate increased agitation and signal the rep to employ mitigation strategies. Interesting, but invasive, unless conducted in an opt-in experimental environment as much of consumer research data is collected. Are there vendors out there that offer this capability?

The second musing is related to a recent sales attribution technology vendor discussion, many of these tools currently appear to be using market research and survey scenarios to hone measurement of various types of media combinations and frequency for optimal media mix. Combining channel data, incorporating web and search data, and tying in data from a social media monitoring tool, then applying text mining to assign events and sequence in the product awareness process will provide an alternative transaction attribution source. Online purchases and intent to purchase on the web need not be based only on transactional shopping basket or purchase data, or survey response. A separate social media datamart sourced from various social media sites would be valuable, to identify Twitter, blog, Facebook, etc. comments related to brand, purchase and product sentiment, reaction to a piece of direct mail, thanking someone for a product suggestion, commentary on a radio or tv spot. This data could then be processed with text mining technology, and assembled into a consolidated separate database to provide another method to assess sequence and attribution. and be tied to various social media-specific user names. For a segment of these, it should be possible to link contributors across various sites, to provide an online view of the sales sequence and brand awareness evolution of the customer. This data may also be used to validate existing measurement methods.

Virtual Store Simulations for Retail Research and Business Intelligence

2010-02-01T06:16:00.000-08:00

Cutting edge retailers are employing visualization software that uses 3-D technology to simulate retail stores and product placement options to optimize their merchandising efforts. Four vendors in this space are Decision Insight, Fifth Dimension, Red Dot Square Solutions and Vision Critical. The software provides computer-based store configuration simulation to perform market research, with associated advantages of lower setup costs, greater flexibility, better accuracy of consumer behavioral tracking and data collection.

This WSJ article describes how the technology's retina tracking device and pressure sensitive shopping basket handle can be pushed to navigate the virtual store, while the software captures the virtual shopping trip and logs what captures attention and holds it.

Aside from the market research applications, a virtual store/ Business Intelligence tool that allows retailers to filter on specific customer segments with product placement and sales data, producing heat maps to show trends in seasonality and product popularity, would allow retailers greater flexibility in terms of optimizing store display. Information might not be actionable in a real time store rearrangement scenario, however could be utilized on a weekly, holiday or monthly basis where a store's most profitable customer segment shifts based on season, product launch, and other factors.

Mobile Tech Cocktail and Applications of Mobile Analytics

2010-01-29T07:17:00.000-08:00

I was able to stop by the inaugural Mobile Cocktail event in Chicago last evening for a while. Several tech firms in the mobile space were on hand to showcase their applications, one of which was an interesting WalkJogRun itunes application for the iPhone and IPod Touch that helps locate exercise routes from WalkJogRun.net. The 'limes' were well-represented as either vendors or visitors at the event, iLime is a solution for integrating Apple Push Notification Service, KeyLimeTie also does iPhone application development, and another called FizzyLime.

These events always seem to lead to fascinating conversation with someone passionate about the web and social media analytics space. This evening one was with another of the Chicago Web and Social Media Analytics Meetup members, a producer working on a series portraying the growing technology presence in Chicago, was discussing his interest in mobile analytics and how it can best be leveraged by companies. Currently only about 22% (if you have a link to a specific percentage please post) have a dedicated mobile website for their customers. With the explosive growth in mobile, companies who apply analytics to their mobile sites and tailor them specifically to optimize information differently on their mobile sites to reflect information research on products and patterns that are associated with the highest customer LTV, loyalty and brand awareness stand to gain a competitive edge with their mobile customer segment.

Facebook Fellowship available for PhD Students

2010-01-24T06:40:00.000-08:00

Facebook announced it's first Fellowship program on January 8th of this year, to support social media and web related research done by PhD students related to e-commerce, cloud computing, data mining of social media and web data, and other related projects. Five fellowships will be granted in late March, providing a stipend and tuition/fee remission. More details can be found here.

Meetup.com now offers Google Analytics for traffic monitoring!

2010-01-23T10:45:00.000-08:00

As a huge fan of meetup.com and an organizer of a couple current and past groups, it's very exciting to see the news about the recent capabilities created by a union between Meetup.com and Google Analytics. Instructions on how to set this up are available here. The service provides information on page views, mapping of visitor location down to the city level, pages viewed, keywords and referring sites, if discussion postings or group emailings generate signups, and the parts of the site that have interest but aren't being developed. For instance, if organizers aren't using the poll option but visitors are navigating to this page, it might be useful to test these. Does group enrollment and member participation increase when event pictures are posted, and is there a correlation between photo viewing and attendance at future events? All you need is a unique Google Analytics account set up for the meetup group url, and the associated 'UA Code' to set this up for your group.

Social Media Activation vs. Social Media Optimization

2009-11-22T05:47:00.000-08:00

I came across a LinkedIn reference to 'social media optimization' yesterday on a recommendation update by Avinash Kaushik , which is a natural term for the process related to SEO, defined as 'a set of methods designed to lead visitors to website content by promoting and publicizing it through social media'. I've been embroiled in a literature review on the separate topic of 'social media activation' for the past few weeks for a deliverable due in a Consumer Behavior class I'm taking at Purdue, which has led to about 3600 (at least) journal articles on various topics related to marketing scales, social-psychological net rituals in online communities, the elaboration likelihood model, anchoring as activation, social media software and related data harvesting options and variables.

This late 2008 WSJ article talks about Web 2.0 and how most companies have yet to make in-roads into properly understanding the potential and correct approach to deal with this new generation of customer interaction. The primary steps they suggest are to walk the customer through the marketing process, interactively communicating with customers about product features and promotions. It's also important to somehow provide an incentive for customers to participate, either through rebates or recognition. Be sure to listen to what's being said and occasionally use the opportunity to respond, and don't focus on social media for selling, also be open to experimentation. Web 2.0 is evolving rapidly and nimble companies are finding new and innovative ways to seize the opportunity to enhance traditional market research, customer satisfaction, and brand promotion efforts.

A quick search on a handy academic journal freeware search tool called Hartzing's Publish or Perish on the term of 'social media optimization' returned 71 papers with 22 citations over 4 years. Browsing down the list, none of them appear in the academic journals. A term search on 'social media' returns more, but just a smattering of journal articles related to computer systems or psychology. Apparently it's still a wide open publishing field for researchers.

The aim of the 'customer social media activation' study is to develop a scale companies can first use to develop a benchmark to understand where they are in the social media evolution process - that is, how many customers and prospects they have using Web 2.0 tools to interact about their brand, how these interactions are categorized (are they complaint related, product feature comments, customer service issues, potentially damaging brand mentions, or complements that should be magnified?) Who's making these comments, how often and across what social media clusters? Do certain industries, such as supermarket chains, have a different clustering of interaction, volume and nature of comments than telecommunications companies? The social media activation scale will provide a tool by which to measure starting process for a company moving into the social media management space as well a strategic tool to determine if current activity and growth matches corporate goals, and an ongoing customer social media activation measurement tool. Such a scale allows companies to set a baseline, set a direction and measure against this as the consumer Web 2.0 activity evolves along the social media life cycle, as consumers move along a customer evolution life cycle. The term activation indicates a guided process promoting the types of online interaction that benefit both consumer and brand, and steering away from negative interaction. The measurement scale will be customizable by industry and company size and use applied statistical techniques for clustering and probability prediction.

GeeYee - a promising addition to the social media analytics scene

2009-09-10T05:44:00.000-07:00

I first heard about GeeYee's Opinion Observer Social Media Listening Platform as a reply to a blog posting about sentiment classification tools on Twitter. Since then I've had the opportunity to view the solution, and learn how the company differentiates from other leaders in the space such as Radian6, Umbria, BuzzMetrics, SM2 and Google Analytics.

The GeeYee solution offers a compelling combination of great graphics, innovative sentiment classification system that goes beyond text-mining to capture contextual information, light and rapid deployment without heavy consultant support needed, and an attractive starting implementation price. Users are able to view brands, competitors, customer care concepts on a heat map that facilitates quick identification of where brands stand in relation to others in terms of customer satisfaction with packaging, taste, price. Bar graphs show relative ratings based on positive and negative sentiment on main product/brand category items, with drill-down and data exporting capabilities throughout the interface, even to the individual link level.

One interesting application of this software might be in a customer care center, with dedicated reps using Geeyee to monitor brand sentiment and responding rapidly to specific customer comments and issues in the same forum the comment was made. The solution sources the best-in-class 3rd party data to encompass social media conversations across all publicly accessible social media sites, blogs, and web pages. This looks like another solution to watch in the emerging social media analytics arena.

Tableau...BI software vendors have a serious contender

2009-08-19T06:12:00.000-07:00

Companies have another Business Intelligence reporting software option that is gaining traction since it was first released in 2005 (thanks for the retweets correcting the original post!), one that at first trial appears to do all it promises and more. Non-technical users really can create a file, connect to almost any database or data source, generate a template with drag and drop options, and publish directly to a PDF or dashboard. Regular reports that need development? Simply update the data source, refresh, and the report automatically takes care of the rest. And the price point is almost absurdly low, for desktop, viewer and enterprise additions. An informal survey of dashboard reporting users (ok, on this blog) familiar with Cognos, Business Objects and SSRS found this excellent in terms of ease of use, efficiency and quality of reports.

Measuring Twitter Sentiment

2009-07-09T08:22:00.001-07:00

I came across a very interesting online tool using Twitter Search yesterday, that categorizes tweets into 'positive', 'negative', and 'neutral' for keywords and brands. It's reasonably accurate with the categorization, and viewers have the option of correcting the classification so that presumably the algorithm becomes more refined over time. I'd like to find a sentiment classification tool that provides summary data and date, with twitter user name and other available info in a downloadable file format that can be used for predictive analysis and segmentation in SPSS or SAS. Any ideas?

Social Networking Analytics and Behavioral Targeting

2009-07-01T14:28:00.001-07:00

A June 29, 2009 Adweek Media article 'Connect the Thoughts' by Brian Morrissey explores the promise of data mining the plethora of social network data, and discusses the uncertainty involved in sophisticated behavioral targeting using social media data.

Advertisers recognize the targeting potential that lies in the new data sources, where members of online communities are interacting together in a way that demonstrates their connections, influence and relative taste for technology, that should, at least in theory, provide highly identifiable traits among segments of users who tend to purchase online, and who not only become brand fans but recommend products to others. Advancements in social media have changed how consumers fundamentally connect.

The online world presents a new challenge, old methods of content-focused advertising are shifting to audience-centric targeting. Online data is overwhelming in it's volume though, and mining through this to create intelligence requires collection of one's own customer and prospect data, and supplementing it with 3rd party data.

Currently advertisers are having the most success with the click mining of this type of data by focusing on re-targeting. That is, targeting existing customers who have previously made a purchase, with another item they have a higher affinity for or have scored high in terms of probability of uptake. Not only is a customer targeted with the 'best product' recommendation, but so are 5-6 of this customers closest connections. So far, as demonstrated with AT&T Labs Research with AT&T data, as well as practiced by industry leaders such as iCrossing, Media6 and Havas, this re-targeting approach is found to be effective by extending advertising reach and response among cohorts with a high propensity of response. If one customer in the network is likely to buy, it turns out that their connections also have a higher likelihood of response.

Loving the Groupon Idea

2009-07-01T07:21:00.000-07:00

Bill Tancer talks about the creative applications of Hitwise data to derive actionable business intelligence related to online web efforts in his popular book Click. The Hitwise site has a nice blog link, Hitwise Intelligence. One of the more recent links related to the increasing rise (though the rate is now leveling) of the online discount coupon. Or rather, as they say in the UK (where this study was focused), 'vouchers'.

The search for 'online voucher', 'discount voucher' and other general variations on the voucher theme peaked in the months leading up with budget-minded recession shoppers searching on 'voucher' in the weeks preceding Christmas. There was a dramatic drop in UK 'voucher' searches after the holiday, which is to be expected. Recent search studies have indicated an overall drop (28%) in voucher searches, but closer examination of search behavior in this market indicates not so much a relaxation of discount searches so much as an increasingly sophisticated search pattern among consumers. More specific voucher related queries are coming in, related to brand and product, so the search traffic is redirected, not seeing as steep of a decline as a cursory look at general search queries might indicate.

Related to the online coupon theme, here's one I signed up for on Facebook a few weeks ago and it's a great concept so far, and has a lot of appeal. I've purchased one ($25 coupon at Manny's for $10!) so far, and sent it around to friends who have done the same. Meet the 'Groupon'. They're on Facebook, on Twitter...and send a daily email with a generally enticing local Chicago offer at steep savings, which is only valid provided a minimum threshold of responders sign up and buy. Here's how they describe the process on their site:

'Each day we feature something cool to do at an unbeatable price.
You only get it if enough people join that day… so invite your friends!
Check back the next day for another awesome Groupon!'

Today's feature is 1/2 off Nona's pizza, named the 'Best Chicago-Style Deep Dish', and there's also a bonus ad for $89 Cubs roof top tickets.

Google gives partners analytics on YouTube pages

2009-06-23T16:41:00.000-07:00

Twitter Helped Dell Increase Their Sales

2009-06-23T16:39:00.000-07:00

Reuter published that Twitter has helped Dell increase their sales in millions of dollars. According to Dell, it has raked more than $3 million from Twitter followers who clicked through its posts to its Web sites to make purchases. Dell uses proprietary software to track their sales. Of course $3 million is nothing compared to their complete sales revenue per year, however it shows that companies can profit and leverage from Twitter.

Dell the PC Maker has about 600,000 followers and is one of the top 100 most-followed accounts on Twitter according to private trackers “TwitterCount” and “TwitterHolic”. Small companies are also finding financial success on Twitter. Co-founder J. Leach of Naked Pizza ($1,000,000 yearly revenue sales), sees a sustained 20% sales from its Twitter presence.

Question remains for Twitter, should they charge for such benefits?

Another Twitter Provocation: Obsolete within a year?

2009-06-15T05:53:00.000-07:00

Jason Clark has borrowed the iMedia megaphone, and is predicting the demise of exponential Twitter growth within a year. It's a well-formulated argument. He provides the attached Social Network timeline and predicts that Google Wave will take the forerunner spot as a central consolidation portal to manage social network communication.

Google Earth and Democratized Intelligence

2009-06-15T04:32:00.000-07:00

After hearing the NPR version on the radio yesterday, I went online to check out the featured website 'North Korea Economy Watch' and also found extensive links across the Internet following up a May 23rd Wall Street Journal article that showed how a small group of citizens are slowly uncovering North Korea's vi sable secrets with the combined use of Google Earth and a growing network of website visitors that supply pieces of information.

A George Mason PhD Economist candidate started the project two years ago, that focuses on an online map of North Korea with tags indicating a wide variety of labels such as 'nuclear issues', 'dams', and ;restaurants'...essentially all economic, military, political and cultural infrastructures. The tabs themselves provide additional information about the location. The online map indicates rocket launch sites, prisons and upscale beach resorts. Mr. Mason visited North Korea himself twice on two group tours, and started posting locations online related to what he recognized from his visit. Soon his site became a popular online destination for those with interests and personal information about the country, and people would send in tips and information about other locations. Over the course of two years the information assembled in wikipedia-fashion through a web-site portal, using the technological mapping advancements of Google Earth, supplemented through anonymous global information sharing. Information and updates come in from contributors describing themselves as former US military professionals assigned to study North Korea, and others who prefer to remain anonymous. The information is all cross-checked wherever possible. The result challenges North Korea's old Soviet-style of secrecy and exposes information about one of the the worlds least open countries.

The Tweety Wheel gets the Grease

2009-06-09T04:14:00.000-07:00

We're still in the early stages of development in terms of finding ways to get customer generated chatter, pull it into a customer/prospect database, match it up somehow to the rest of the data we have on customers, and even then just what to do with it. Monitoring for brand sentiment and building an opt-in fan base, mitigating negative online brand chatter, using the various social media sites like Twitter and Facebook as sales channels are all examples of early forays into the relatively uncharted depths of social media. Analytics is never far from mind, and new and creative ways of using the data for predictive modeling and segmentation is of particular interest.

It's not really new to think of customer influence, and prioritize campaign targets based not only on propensity to respond as identified with logistic regression and other probability modeling but also on a customers potential reach. One company that was onto this early in the game is Idiro Technologies, applying analytics to call data to segment customers into different segments of influence, based on volume, reach, types of calls (business vs social). Using these measures for predictive modeling and customer segmentation still seems somehow new though.

If I were putting together a wish list of data derived from social media sites, I'd include customer influence metrics. This could be developed by collecting social network log on information during customer registration...asking for Twitter user name, Facebook name, YouTube and LinkedIn presence, blog names and creating an in-house measure of social reach. That is, a customer with none of these filled out either doesn't want to provide the info so goes into a privacy-conscious segment, or isn't a social media butterfly. A customer that fills out this information and has an id or a check for all of them is in one of the more potentially influential buckets. I'd take that one step further and gather a monthly Twitter influence metric by putting their user name into Twitalyzer and getting point in time metrics on social clout, generosity, relevance and so forth. I'd go onto Power Twitter and get another metric that shows volume of tweets, numbers of followers, and each one of these would be components in a data set used for predictive modeling and customer profiling to incorporate their online reach. Still not sure how to get a quality rating on the number of followers and friends, perhaps there's a comment crawling software out there that would go through all of the tweets, retweets and responses with a text mining tool to identify topic affinities related to brand and purchase behaviour.

Another interesting behavioral metric other than influence is demeanor. What categories to the majority of this customers tweets or blog comments fall into? Are they a generally positive contributor, or a chronic complainer? Are they news link providers more than anything, relaying information rather than opinion generators? Again all of these metrics can be identified through text mining, and even be managed somewhat manually at first in something like base SPSS or SAS with tweet coding (i.e. does the tweet, comment, blog contain the any of the following words (compile a list of them based on a manual review, and build on this continuously) and do a bunch of logic coding where 'if word 'great' is present in tweet then 'positive' or '1', else '0' and flag coding and summation to come up with a measure of how positive, negative or neutral the quote is. It won't be perfect, but over time with refinement it will get better. And there's enough free text mining ware out there that the described method to identify demeanor can probably be side-stepped. More to come on the freeware text mining options and additional customer data set variable wish list!

Super Chirp and paid Tweets

2009-06-08T09:34:00.000-07:00

New service from Super Chirp allows Twitter users to get paid for their content. Also by 83 Degrees is Power Twitter, that allows you to see not only all tweets generated from a twitter account, but every time that twitter user name is referenced in Twitterland.

Bing and the Archivist

2009-06-08T08:08:00.001-07:00

Some random Monday thoughts related to the web and what's new and interesting, to jump start the blog this week. In a recent search to find applications that work with social network sites and facilitate the mining of Tweets and comments, I came across The Archivist... it's a genuinely helpful tool. Twitter only provides about a month's worth of data in it's Advanced Search tool, after that the Tweets are no longer available to the general public. This tool allows a manual download of search results. If you're monitoring brand sentiment for a company, a monthly exercise with this tool will help assemble an updated database of brand sentiment over time.

The big news in late May was the launch of a $100 million branding campaign for Microsoft's new search engine Bing. A recent BtoB article 'Bing or Bust?' speculates on how the new launch will impact the paid search landscape. Most advertisers are spreading their online search marketing dollars between Google and Yahoo since MSN only has about 8% of the share of searches, and as Bing garners part of that search market advertising dollars will shift to Bing paid search as well. More to come this week on Bing, and new uses of analytics and segmentation to boost sales.

Social Media Marketers the problem, not Social Media?

2009-05-27T07:10:00.000-07:00

Online Spin has a piece featured today by Joe Marchese, one marketing guru who does not buy the recent conclusion that Twitter is mostly hype, and the assertion that social media has failed as a medium. He puts forth a good argument that the issue is that marketers are not marketing effectively or present en masse...mainly he encounters ads for whiter teeth, online dating and work from home jobs..all products and services he shouldn't be targeted for based on his low propensity to respond to, regardless of channel and creative. There's a flaw in the assumption made in the article (see blog entry below for link to article), stating that research shows folks don't go to social media sites to solicit information about products therefore social media is failing as a channel.

What influences consumers purchase decisions? The answer is fluid, it's based on individual need, personality, circumstance, how likely they are to do pre-purchase research online, in stores, talking to peers, based on industry rankings and surveys, price point, promotion, how compelling they find the copy, individual mood, the weather, season, and employment status, how fast their Internet connection is, who else at home needs the computer, just to start.

Twitterers are still ramping up the adoption curve and figuring out how to gain the most benefit from it. Many online are still going to Google and the other engines with product and rating questions, but will continue to shift to Twitter for more real-time customer experience testimonials, however brief. Some measure of social media site tenure and volume would be useful to factor in to form a proxy snapshot of time-in-time location along the social media adoption curve. Whereas historical multi-channel analysis was relatively stable, once a customer was tagged with a channel priority preference, social medial channel targeting requires more frequent update of channel preference score as adoption and expansion of use increases.

Following that line of thought, channel effectiveness is individual. Customer segmentation is the solution to understanding the effectiveness question, the more granular the better. Social Media is certainly failing in some segments. Consumers with no interaction in social media and low Internet usage would be fruitless targets via these channels. Certain segments of Generation Y consumers would be ideal candidates for social media targeting, but poor candidates for direct mail or outbound call center targeting. Consumers are doing word-of-mouth research via social media sites, and for companies who are selling a product where a profitable current or prospective target market does have regular interaction on social sites it is advised to start monitoring traffic and participation on sites, to figure out which social media channels are important. Perhaps there's a high i-sentiment brand rating on Facebook, but a lower one on Twitter. Social media channel brand and word-of-mouth strategy to mitigate negative i-sentiment is one solution. Identifying action points can be done through segmentation and text mining, overlayed with probability modeling. WebTrends has a new tool that acts as a brand meter for the popular social media sites, and serves up the dashboard reporting and auto generated custom reports based on customer configuration, it's based on the previously mentioned Radian6 engine, as is Cision. Omniture is developing one as well, as is Google.

Once the critical site and trend is identified, here's an article with a few suggestions on how to improve word-of-mouth sentiment.

Twitter mostly hype?

2009-05-22T04:50:00.000-07:00

Thought-provoking article in MediaPost about social media failure to develop into an effective advertising medium. The statistics are compelling. About 83% of all internet users are on social network sites and half of these are regular social media site visitors. A new report 'How People Use Social Media Sites' finds that they aren't much of a marketing channel. People benefit from the connections and interaction with friends, but don't use these sites for brand research. Only 5% solicit purchase related information from friends and contacts.

The numbers on Twitter are even more out of proportion to the media attention it's getting now. Only 1% of the population uses Twitter weekly. It's importance is expected to expand in the realm of promoting positive brand sentiment, and less as a channel for direct connection to consumers.

Commute and the Coffee Wars

2009-05-21T07:33:00.000-07:00

The commute this beautiful Spring morning from my home in Forest Park to the blue line was made even more pleasant by my new iPod Shuffle (hazard of working 3 blocks from the Michigan Ave Apple store), listening to a great new free Starbucks Tuesday download 'Fitz and the Dizzyspells' by Andrew Bird, thinking about playlists and the wealth of consumer behavior data and predictive segmentation opportunity in individual iTunes playlists. (Also in the Blockbuster.com queues, where members can share movies and ratings with friends and recommend titles, and also the Facebook movie comparison sharing tool.)

There are indications of economic recovery around Chicagoland, for the past few months huge billboards along the 290 have had vacant real estate, with 'call this # to advertise here', but the Chicago beverage advertisers are out in force now. Pepsi dominates the Chicago highways this month, and McDonald's has a really delicious looking McLatte ad.

McDonald's and Starbucks are in a battle for the upscale coffee market mind-share. McDonald's launched it's McCafe advertising blitz earlier this month in an aggressive attempt to capture some of Starbucks' floundering specialty coffee market, representing their largest product launch since the 1970's breakfast promotions. Their mocha and iced coffee promotions are all a part of a national branding effort for McCafe. Last year, McDonald's spent $820 million in US advertising, they are the number one US fast-food advertiser (Subway is in the 2nd slot).

Starbucks has the edge taste-wise, and have long been the darling of mainstream specialty coffee drinkers. Their recent ad launch mentioned in this blog a couple days ago is directly attacking McDonalds on quality, as well as incorporating social media into the advertising channel mix. McDonalds has the price advantage.

It will be interesting to see the impacts of the May coffee advertising wars. A quick search on Google Trends shows something thing interesting happening since the beginning of the year. McDonald's is garnering the majority of Google search share in Australia, the UK and Germany since the beginning of 2009. This view is really interesting at the Illinois level which does indicate a recent increase at the state level as McDonald's gains more search share. Take a look at the city level data on this, Evanston and Northbrook favor Starbucks and Bloomington and Lake Forest are spending more time searching for McDonald's related links. Twitter search trends site 'Twist' shows a different pattern, but only has a 7-day window on tweets, and this coincides with the new Starbucks social media advertising Twitter and FB contest out this week.

Measuring the effects of Social Media

2009-05-19T11:26:00.000-07:00

Thanks to a LinkedIn connection post about an interesting NY Times article, I'm reading about the new Starbucks advertising campaign in six major cities aimed at leveraging opportunity in online social networking sites by launching a contest to find the ad posters, take a photo and be the first to post this on Twitter. It's a strategy aimed at Generation Y, and the idea mimics existing online Starbucks brand-related social behavior patterns that were developed during photo contests, involving a race to post stores holiday-decor, or one on Flickr where photogs search out perspectives that capture 2 or more Starbucks in the same frame.

The article cites other upcoming social media branding campaigns for YouTube. Measuring the effectiveness of the 1.5 million Facebook fans and 183,000 Twitter followers is still a conundrum.
Chris Lake at eConsultancy posted this March 2009 list of 10 ways to measure Social Media Effectiveness. Here's a summary: Launch it and let it develop, and take an aggregate approach to measurement. Instead of focusing on statistics like numbers of fans or tweets, focus on the correlation between increases in these related to increases in key KPI's of sales, profits, customer statisfaction and loyalty. The key is to let it age and look at performance over time. Start with a snapshot at the outset, before you launch. What kind of numbers are you starting with (number of fans, followers, links, referrans, website trends, SEO rankings, and ALSO ROI KPI's and business data tied to the same snapshot date)? Things to focus on are traffic (but don't mistake quantity for quality), level of customer engagement and trends over time, tracking sales from all social media sites, not just paid search and FB. Twitter is now a channel in it's own right. (The article cites how Dell did a bit of research and found out that $1 million in sales came directly from Twitter in the past year and a half!)

Social media sites are a hotbed of customer leads. SEO and search marketing are key, as there is an exponential factor involved in posts and links, and retweets and referrals, and all of these elements translate into higher organic search rankings on Google and the other major engines. Monitoring brand-related sentiment is key.

While the article doesn't go into the text-mining applications to refine sentiment and categorize it through machine learning, it looks inevitable that these text insights will soon gain momentum as an integral part of online analytics efforts and eventually the major brands will employ text mining experts in their Social Media Analytics departments. This data combined with call center speech analytics and tied back to the customer level, and measures of individual i-fluence may soon make the traditional offline purchase behavior data analysis many companies do look like analytics from the dark ages.

Pioneer brands that start leveraging text data through text mining of social media, blog, customer email, search behavior and tie it back at the individual level to call center data translated into text, to form a consolidated text insights-based segmentation strategy, should enjoy a huge competitve advantage in time and relevancy consumer targeting.

Avatar Analytics?

2009-05-12T10:05:00.000-07:00

A friend brought secondlife.com to my attention yesterday, when I was talking about a general interest in applying social network data to customer probability and loyalty modeling. An August 2008 WSJ article describes the 8 million individuals worldwide who join and create an alternative virtual fantasy life, forming relationships, purchasing homes, sharing a mortgage with their online spouse, and going into shops and exhibiting online consumer behavior.

Avatars are generally the antithesis of who an individual is in the real world. They tend to have vastly different networth, salary, location and circumstances. However, all of their preferences and purchase behavior are captured, as well as IP address. Just how useful is this data in terms of building online customer behavioral profiles? Do preferences translate across the virtual devide and are they reliable sources of data for product location and packaging, customer loyalty and price sensitivity?

This is the first in a few articles on the application and testing using the data for bevioral and social studies. Here's a related pyschological conference on virtual life and social behavior, which features a semiar on advertising effectiveness to avatars. Here's a digital ad company that hails advertising with avatars as one of the most effective types of advertising available.

Along that thread, many of the SN sites offer the option of avatar creation. In terms of using avatars for marketing, a flag for Facebook_Avatar_Usage and Twitter_Avatar, and so forth, could provide a simple measure for the probability of avatar advertising effectivness at the individual level based on count, so a different type of targeted copy opportunity.

Social Media Analytics: Probability modeling to reduce churn and build loyalty

2009-05-07T13:05:00.000-07:00

One area we will soon begin to see emerge on the web analytics front is advanced social media analytics. Paid search has been seeing a year-long shift towards advanced analytics, in terms of identifying optimal bid strategy through the use of regression, classifcation trees and time series.

Social Media content, such as blogs, videos, Twitter postings, Facebook updates all contain some degree of minable text (videos also have titles and descriptions). Companies that collect and house social media data in a warehouse related to their brand and customer base, who collect the main social network site logon names (we're all accustomed to providing email addresses, why not collect Twitter id and Facebook signon along with other customer information efforts?) and are able to tie the data back to a single customer record, and append this information to analysis data sets will take multichannel marketing to a new level. By applying probability modeling to large datasets, companies will be able to identify the top 5% of customers most critical to contact to retain valuable customers, minimize negative brand buzz, and reduce customer churn. The social media data therefore becomes actionable and impactful with the analytics refinement.

Intrinsic to understanding social media content are text insights. The content is first run through a text mining tool and generated into 'positive', 'negative', 'neutral', 'unidentified', sentiment categories, then broken down further into 'product-related', 'customer-service-related', 'price' and other relevent categories.

Smaller companies don't have the need for text mining, but larger national brands are unable to process all of the Tweets and comments and blog entries and separate them into actionable categories for quick response. That's where text mining comes in, and the reason it's critical for larger companies to start collecting, housing and analyzing brand-relevant social media data.

Human identification can initially be used on the text to identify top-level 'positive' and 'negative' sentiment rules that are used for initial bucketing. These rules will soon be replaced with machine-learning identification of key terms and combinations of terms that have historically resulted in a positive or negative outcome. A worrisome negative outcome, for example, is customer churn, or a negative comment that is retweeted over a certain threshold and begins to increase the negative buzzshare. To carry that idea just a bit further, 'negative buzzshare' is a separate concept and should be maintained as a daily metric. Google Insights and BuzzMetrics and Hitwise are tools that can be used to maintain a few aggregate 'brand meter'-related metrics. The combination of an increased volume or high threshold of negative retweets (in scenario two) that coincides with an increase in negative buzzshare based on these 'brand meter' metrics should automatically trigger an alert email with the details of the original negative content generator as well as retweeters, to the social media monitoring and PR department for mitigating strategy action.

Customer churn associated with social media content generation at the customer level offers a great opportunity to append customer content to customer loyalty data, and use as predictive variables in churn models and customer lifetime value models. Collecting and housing this data, collecting customer logon information (at least until a clear standard to tie all this together such as ping.fm is in place) to the most popular social media sites is the first step in moving towards a truly integrated multi-channel marketing analytics environment embracing the social media channel opportunities.

The Rising Advent of the Brand Meter

2009-05-07T08:21:00.000-07:00

Along with the explosive growth of user-generated content on the web via blogs, social networks and social media comes the necessity for companies to monitor what is being said about their products and service, in order to protect and build brand. One company that stands out in this arena is Radian 6, by offering a subscription to a dashboard monitoring service that allows the selection of channels for monitoring and six metrics related to brand: on topic posts, total comment count, total unique commenters, average engagement, on topic inbound links, and vote count.

It also provides information on which channel, site and user provided content related to the search content specified, and delivers up a constant steam of information related to these alerts. It also offers the ability to track online buzzshare related to company and brand.

Tweet Insights on Earth Day...free coffee for all with travel mugs at Starbucks today

2009-04-22T03:35:00.000-07:00

In the media news this morning, Pizza Hut is searching for a paid twintern, to post deals and tweet about new menu items (pasta and sandwich selections aimed at swinging the veto group+ lunch vote not in the mood for pizza), while monitoring pop culture news and negative buzz on the brand. The article discusses the positions other responsibilities related to Facebook, the Pizza Hute FB page has more than a million fans as well as a FB ordering app, and other social media outlets. Digital brands will use the social media connections as a way to interact freely with customers in real time. It's an emerging channel that facilitates forthright real-time conversation.

In honor of Earthday, here are the search results for the hashtag #earthday on Twitter Search right now. Thanks to Rightsleeve for this one...head into Starbucks today with your travel mug and get a free coffee!

SpotLight: GPS dog finder debuts next month

2009-04-21T05:54:00.000-07:00

Pet lovers have something to cheer about...this new device will become available next month. Information Week discusses the American Kennel Club's SpotLight product to help locate lost dogs (and cats?) with a tiny GPS device that attaches to the pets collar.

Owners can use mobile, smartphone and pc to access mapping software and tracking for up to a week on location and path, a dedicated website and staff that helps locate and return the pet. Alerts can be set if the dog leaves the yard or an otherwise designated 'safe spot'. This device/service is priced at $250.

Recession hits Mother's Day Spending Forecasts

2009-04-20T02:20:00.000-07:00

Interesting Monday article in MediaPost....Mother's Day expected spend is $14.1 billion lower this year, after Valentine's Day and Christmas this day represents the 3rd most popular US holiday for cards. This holiday ranks #1 for the hispanic population. The article points out that traditionally this holiday is used to celebrate relationships with not only mothers, but with daughters, friends, grandmothers, stepmothers and sisters. The expected decrease in spend this year is within the ancillary relationships listed.

Where's the analytics industry heading?

2009-04-19T04:39:00.000-07:00

Interesting interview on Lars Johanssons blog with Joe Davis, CEO of CoreMetrics on the competitive view of enhancements to 'complementary software' such as Google Analytics and where the focus will be in the next year for analytics. Joe takes the generally accepted industry view that Google Analytics serves the analytics industry as a whole by raising awareness of web analytics and making software accessible to small companies and analytics shops that otherwise would find the task prohibitively expensive. Coremetrics and other software providers in the space have a different target market that require more sophisticated solutions that aren't offered with freeware options.

In terms of where the industry is going, Joe suggests greater personalization, ironing out the ROI attribution question by moving past clicks and impressions to presumably a wider multi-channel integrated equation. More refined individual online targeting, improving the ability to combine online and offline data, and incorporation of social media data. Along the topic of greater personalization, Yahoo now offers more demographic targeting.

Greater emphasis and targeting of mobile customers is a certainty. Facebook now has over 25 million mobile users, 4 million of which access the site via a mobile device such as iPhone, Blackberry Bold and other smartphones on a daily basis. Mobile targeting may differ in that it will be more location based and address-book centric, facilitating real time connections based on the two.

A sudden shift to Audience-centric buying

2009-03-30T05:46:00.000-07:00

iMedia Conection sends out daily emails with digital industry updates, today's contains an interesting piece related to the ongoing struggle to allocate advertising budget to optimize ROI in an environment where ad spending is declining across the board or remains flat at best.

The industry is seeing an explosive recent shift (seeing growth rates even greater than paid search) from traditional site-centric advertising to audience-centric tactics...advertising approaches that take into consideration the behavioral and individual preferences of the visitors and website audience and customize the ad experience through the use of contextual advertising. The author of this article predicts the dramatic increase towards audience-centric ads will flatten and eventually there will be a 50/50 split on site vs audience-centric.

Banking on mobile and cable

2009-03-16T05:57:00.000-07:00

The Advertiser Optimism reports, the online equivalent of the Consumer Confidence Index, indicates that popularity and high expectations continue to surround online media, but momentum may be slowing. Also ranking highest in the index is television spend, specifically cable which is seeing growth during the recession where other mediums are seeing decline. Online ad optimism has seen a 72% year over year drop, however mobile growth matches cable in terms of current growth rates and industry expectations

Web Analytics Wednesdays a success!

2009-03-05T07:34:00.001-08:00

Last evening was the March Web Analytics Wednesdays Chicago event, 14 web-minded professionals attended and shared experiences with combining online and offline data, discussed software vendors, how their companies are addressing the rising importance of social media data into the mix. Some vendors that came up in discussion that manage online and offline data were Convio (specific to non-profits) and Conexa. Other topics dicussed include the well-executed Obama web strategy campaign. Special thanks to Eric Peterson and Web Analytics Wednesdays Global Fund for co-hosting this great event!

Time for Hispanalytics?

2009-03-04T05:46:00.000-08:00

The Hispanic and Latino population is exploding in the US, currently represents 15% and the largest minority, and projected to reach 30% by 2050. About 30% of this demographic group is currently under the age of 18. This shifting consumer population landscape represents great opportunity for a marketing shift, to better serve a community that has developed it's own network of commercial and media channels. Consumer preferences for this group will effect product development and advertising strategy as this explosive demographic group continues to gain population share and greater corporate attention.

Here's a good overview to Hispanic marketing, that details dramatic effectiveness of Spanish ads over English, prevalence of radio (on average a US Latino household listens to 26-30 hours of radio a week), and local newspapers that serve as a mainstay of current news to many in this market segment. Hispanic consumers are 3.5 times more likely to respond to direct marketing and event marketing tends to be effective. Univision is the nations 5th largest network.

This February 2009 market report talks about increasing adoption of the internet and Hispanic trends that include higher blog and social network interaction than non-Hispanic groups, preference for Spanish language sites, different navigation patterns and focus on culturally relevant content, and finally the need for segmentation among a diverse cultural group increasing online activity, and expected to grow as a market even as the global recession continues, at a rate of 48% over the next four years.

Another explosive population group, while on the topic, are the Amish who have increased in population by 50% over the past 16 years. While Amish leave school after 8th grade, they are finding the need to have separate practices at home and at work, as an economic shift from farms to small-business takes place. Many are moving online at work, have computers and cell phones. They are ideal small business loan candidates, with a 5% default rate. More than 50% of the US Amish popluation is under the age of 21.

Web Analytics Wednesdays in Chicago Tomorrow Night

2009-03-03T18:57:00.000-08:00

Come out and join the Web Analytics Wednesdays group tomorrow evening at the Tasting Room in the west loop at 7pm, to discuss 'Combining online and offline data' and other topics near and dear to the hearts of online marketers and web analytics aficionados.

Join the Chicago Web Analytics Meetup and see you tomorrow evening!

Sweet to Tweet

2009-02-27T08:19:00.000-08:00

I found an interesting article in today’s FT analysis section.

I think everyone is familiar these days with Twitter. Twitter comes with its own quirky language and the messages are known as tweets. People who read your messages are known as ‘followers’. Twitter found a way for people to connect with friends, express themselves and find information that stands a chance of one day becoming as popular as other mass trends such as blogging and social networking. Twitter users post short messages, either from a PC or mobile phone.

The growing interest explains why Silicon Valley’s financiers sense that this could be the next big thing on the internet. Twitter’s founders created a way for other developers to write internet applications that ride on top of Twitter.

Six secrets of an online success…

Immediacy: Real-time flow of comments and adaptability to mobile handsets makes it more immediate than blogging.

Brevity: Limiting messages to 140 characters makes them easier to produce and easier to digest.

‘Pull’, not ‘Push’: The ability of users to choose whose tweets they follow makes it less random than email.

Searchability: Messages can be searched, making the content move accessible than the comments on a social network.

Mixing the public and the personal: A user’s personal contacts are on an equal footing with public figures.

‘Retweeting’: By copying and retransmitting messages, users can turn the network into a giant echo chamber.

If Twitter succeeds in attracting a mass audience, there will be all sorts of revenue opportunities. But for now, Twitter has plenty of cash in the bank, it can afford to build its audience for ‘years’ before worrying about how to make money.

Check out your Social Media Clout on Twitalyzer

2009-02-27T03:31:00.001-08:00

The author of Web Analytics Demystified has just launched Twitalyzer, where a Twitterer can type in their user name and generate a personalized report on key metrics of relative influence, signal-to-noise ratio, generosity, velocity, and clout. Very neat!

Facebook, on the other hand, doesn't currently have available applications to measure Facebook Influence. Here's a blog with a link to Sixty Second View, which proposes a detailed methodology and weights attached to various metrics to calculate a blogger's influence. Unfortunately there's no fun link to have this auto-generated, however, great concept.

This was kind of nice, while I may only have .1 clout on Twitalyzer, by some fluke ended up #10 on this 'suggested folks to follow' list related to SEO with a 75 out of 100 grade. So I have a solid 'C' but with the curve I get an honorable mention. Thanks cjweb!! @kchulis

Spheres of I-fluence Metrics Adapter needed

2009-02-25T04:38:00.000-08:00

The article yesterday about integration of Twitter data into SiteCatalyst prompted thought about a need for an application that acts something like one of those universal international travel converter/adapters and plugs into the various API's of YouTube, LinkedIn, FaceBook and Twitter and collects the information (presumably by IP address, since the magic identity-protecting ID key to link all this data at the individual level remains elusive).

Here's a blog entry from last year that offers some potential SN influence variables in calculating ROI. Here's an article from this morning's Marketing Daily that indicates that only 48% of marketers are looking at online metrics.

A quick search on API release for some of the big SN sites indicate various stages of readiness to integrate with other systems. LinkedIn did a pre-announce and seems to have backtracked, and only one article on TechCrunch integration with LinkedIn mid-2008.

FaceBook is ahead of the sharing game, with their Facebook developer site and local user groups. YouTube is open for integration as well.

Here's SocialComputing article that talks about how to use all this SN i-fluence data to promote product marketing, and another that discusses why 2009 will be the year of database marketing integration.

Venture Capital Growth in the Silicon Prairie

2009-02-24T12:24:00.000-08:00

While California still garnered the bulk of VC funds (51%) in 2008, the Midwest saw strong growth in the midst of an overall drop in VC funding for the year. The Midwest had 267 deals and $4.9 million in investment dollars, while Texas, the Southeast and New England all saw double-digit declines. Illinois, unfortunately, showed a -12% decrease while Iowa saw a 538% increase!

Integrating Twitter with Omniture

2009-02-24T07:14:00.000-08:00

At the risk of turning this blog into TwitterInsights (it's available), here's another interesting piece about how to integrate Twitter into Omniture data that was introduced at last week's Omniture Summit in proof of concept format.

The premise rests on utilizing a data insertion in SiteCatalyst that uses an API to push non-website data into the database, and since Twitter has an API linked to it's search.twitter.com site, it's possible to connect the two data sources. What could be done with the data? The following business cases are proposed. Suppose you work for a company that wants to tie web analytics together with what customers are tweeting about the product and brand. Metrics of interest include the number of times the company is mentioned, patterns in company-name mentions by hour, day, week, season, and what primary topic focus trends are emerging, what products are mentioned and criticisms or positive comments are emerging, what new features are customers requesting that can be passed on to product development?

In terms of monitoring brand tweets, the search.twitter.com API can be set to collect the results on Twitter search on the company name and link it to an Omniture 'Success Event'. The results can be viewed in chart format in the desired time range, and alerts can be set up once the number of tweets passes or drops above/below a pre-determined level. In terms of monitoring who the most active messages are coming from (thereby identifying the customers on Twitter with the highest I-fluence), both author and recipient (@reply name) were captured, and that not only allows a ranking of the most prolific tweeters on your brand but also interconnections on those who communicate together about your brand (I-fluence clusters, let's call them). These I-fluence communities could be contacted with surveys or identified for other promotion strategies. In the same way that users can be ranked by volume, keywords in the text can be mined and ranked to identify what products and themes are top of mind for customers. Data can be collected and tied to website conversion data, providing another rich and potentially predictive source of customer sentiment data.

SEO goes Mobile

2009-02-23T07:13:00.000-08:00

Resolution Media's Bryson Meunier is doing something well with SEO, each time I do a search on mobile search analytics... there's a top link to an informative and relevant blog post on the topic associated with his blog. This one is a SEO guide from last year to mobile analytics that introduces mobile-specific tracking software that are options to enhance web analytics. While Google Analytics and other leading applications will track mobile, they do not offer the customized mobile metrics that marketers and advertisers will want to report on and use in predictive modeling. The link above discusses new technology better suited to the task, such as Amethon, Bango, GetMobile Mobilizer and Mobilytics (I have a fondness for this last one purely based on name, I have the Hispanalytics domain in my own collection). Here's a short run down on the pros and cons: Amethon has extensive metrics and is priced for larger clients, Bango is a good option for small to mid-size however has a re-direct in capturing the information that may prove tricky based on the integration configured at your company. Mobilytics is also hosted however offers free analytics to anyone willing to put a small text on their site, and offers loads of metrics. Looking forward to a 2009 version of this helpful guide!

Also of note in the mobile analytics landscape is the unveiling of Buzzwire's Mobile social site, recognizing the wireless growth via cell phones and offering content sharing on the Web via email. Users text message 28999 (Buzz9) or click a Buzzie button to push content to the web. The application is ad-supported and has a couple initial advertisers signed up. Last year Buzzwire co-launched a mobile channel related to coverage of the 2008 Democratic National Convention. It is reportly hoping to position itself as the go-to movile social news service.

How to use Social Web for your Business' Free Marketing

2009-02-22T13:13:00.000-08:00

A recent post on the Huffington Post by Lawrence Coburn talks about of how to use the social web for free marketing. Here are 5 easy ways of how to drive your revenue, attract customers and expose your business to the web.

1. Start with creating a Facebook page for your business. It’s a great way of engaging your customers in a conversation.
2. Setup a Twitter account for your business. One simple Twitter strategy is to automatically follow any Twitter user who mentions your business or business category. This can be done through a service called Twollow.com
3. Let your fans / customer review your business. Online reviews are a critical component of your business’ reputation by attracting new customers.
4. Hire a social media optimizer. These people are experts in driving traffic through popular social media sites like Reddit, Facebook, and Twitter.
5. Lastly, Search Engine Optimization (SEO). Make full use of Google Webmaster Tool to optimize Google’s indexing of your site.

The rising importance of I-fluence in the Digital Landscape

2009-02-22T05:34:00.000-08:00

Website Magazine's Feburary 2009 issue devotes it's monthly Top50 to Social Media sites. Citing a Sept 2008 Universal McCann study, the article describes a new technical landscape where three dominant trends are on the rise: increased growth in social media, recognition that digital relationships are valuable, and an explosion of what are referred to as 'influencer channels'. The concept of influence as measured by one's digital relationships is just a shift in channel. Our blog discussed this a few years ago in this old post that I keep referring back to lately (a bit of internal linking for SEO never hurt anyone) because it's relevant, talks about how if a 'company builds a great product or service, people will spread the word. The more contacts a person has, the more potential new customers they will tell, and the key is to identify those customers in the company database with the largest social reach, and create incentives to facilitate connections between them. Amazon's Jeff Bezos says "If you do build a great experience, customers tell each other about that. Word of mouth is very powerful." This has only become more true in recent years, with speed and interconnectivity of communication proliferating.

The Top 50 lists YouTube.com as the number one social media source (measured by Monthly Traffic Rank), as researched on Ranking.com. MySpace and Facebook come in number 2 and 3. Bebo is at 4 and BlackPlanet at 5. My personal favorite, LinkedIn is at number 11, and Twitter at 17. Others such as viddler and Sphinn and grono and revver and Badoo made the grade. Tribe.net has experienced a YOY drop, Yelp.com is experiencing a steep and steady incline YOY, and StumbleUpon appears to be plateauing out. Twitter started skyrocketing in May 2008 and appears to have the steepest growth curve slope at the moment, and a quick check on today's rankings shows that it's moved up from 17th place to 6th in a space of six months. Go Twitter!

The same pattern plays out using Google Trends to compare a few of the top social media frontrunners. This graph shows the global rise of FaceBook and YouTube, the trend for Twitter is easier to see isolated in the second graph. MySpace appears to be flattening in growth currently. A search for what happened in July to instigate the tipping point for Twitters popularity inflection produced this article about Twitter's acquisition of Summarize (to facilitate search on the site) and speaks of $15 million in venture capital acquired two months before.

I used Twitter a few weeks ago looking for blog trends and social influence data, and came up with BlogPulse Trends and Rapleaf, so far. Here's a blog entry from Compound J that has some nice representations of the concept and distinction between real and imagined, good tips on using Google Trends to measure and grow influence (add tags each time you join a SN) and who provided the next blog topic of #hashtags.

Here's another 3rd party tool that measures the reach of your tweets but makes no promises on the privacy of your data. More to come on the I-fluence topic soon.