September brought my annual pilgrimage to California where I have been attending the X Change Web Analytics Conference since 2007 (except for 2009 when I had to go to Paris to launch my book). If you ever heard of that conference, but never attended, you probably already know how so very well participants are treated. As usual, the event was superbly organized under Grace Angel’s auspices, and tireless care. X Change have always been events with much class, and you know what? class is a very nice thing.
Gary Angel, CEO of Semphonic, the Web Analytics consulting firm organizing the event, made sure again this year that the conference content would be of the highest quality, and boy did he do a good job! With around 50 Huddles (the now famous X Change session format), the range of topics was just astonishing. This is always the first tough thing an X Changist has to do a couple of weeks before the event: choosing which 6, out of those 50 deep 90-minute discussions, s/he will attend. Needless to say that having to make that kind of choice is borderline cruel, when one reads the long list of fascinating topics. And at the same time, it’s the beauty of choice; which ones are really the most important topics to oneself this year. This is why I don’t think Semphonic should go through the very cumbersome task of producing a document with all session content. Choose, then live with it.
Pr. Elea McDonnell Feit of the Customer Analytics Initiative at Wharton gave the keynote presentation on the first morning. Pr. Feit did very impressive exposé of the research conducted at Wharton. Several of my consulting colleagues and myself were green with envy; imagine getting data sets from a companies, and having close to a year to analyze it. That’s the kind of heaven in which the CAI group lives in. During her presentation, Pr. Feit shared several key learning points generated by various projects. I for one will certainly spend the coming months digging in the several papers available on the CAI site.
My first huddle was “Analytics in the Cloud” (led by David McBride of Comcast) a session that attracted me because of current reflections on how I will evolve WAO/MARKETING’s services in the coming year or two. However, it seems that, for the near future at least, “clouding” analytics is beneficial to organizations who have to invest very heavily in infrastructure, computational capabilities, etc. Fascinating challenges for sure, but something I think still on the edge of analytics, and not accessible to small businesses.
The second session brought me in a room with several other people interested in “Campaign Attribution Models” (led by Jared Waxman of Adobe Systems ) . Most participants agreed that the last-click attribution model was by far the most biased, and usually skewed toward search. We didn’t come up with an ideal model, of course, and there seemed to be an agreement that each company, or at least industry, has to figure it out. I think that question evidently brings the technological one, i.e. how well our current Web Analytics products are geared to handle it (I will soon blog about GA new multi-channel function, and I will have a white paper on attribution model sponsored by AT Internet in France).
My first day ended with what probably was the most fascinating huddle. Fascinating for someone like me, i.e. involved in Web Analytics consulting for the last 10 years. Led by Katie Reagan of Marriott International, “BI Tools & Web Analytics” was a session where I had the opportunity to hear large, analytically sophisticated businesses, say that the most valuable analyses they performed were now happening outside the traditional Web Analytics tools. It was quite a shocker to me (and I believe to Bill Gassman too, who was sitting next to me); in short, those companies are now using those applications as just one step in their ETL process! Sure, that discussion had been around for some time (see here, and here). However, never had I before heard companies clearly saying they were now beyond WA tools, and building their own suites of web analytical powerhouse. I remember one participant saying they were even considering building their own data collection system! The only vendor present, IBM-Coremetrics, seemed to be jubilating; I guess they are quite well positioned to support where those large accounts are going. When I asked participants if they would keep investing hundreds of thousands of dollars in Web Analytics solutions that are destined to just preprocess data, most of them said no. See also Gary Angel comments at the end of this post.
The second day started with “Predictive Analytics – Low Hanging Fruits” session led by Kiele Cauble of American Airlines. Kiele gave us very interesting examples of how American Airlines uses predictive analysis of Web site data to serve other sides of their business. Predictive Analytics are all the rage these days, and recently I even came across a new buzzword in an IBM commercial, “Prescriptive Analytics“. This is obviously not done everyday by every business, at least with the Web arm of their operations. Most participants were rather on the curiosity side, wondering what those performing that magic were doing and how. Of course, there were a lot of talks about products, sophisticated ones mind you, and how not everyone could afford implementing them. I offered that sometimes, predictive analytics, or let’s say more humbly, finding correlations that seem to be predictive of something, is quite possible without having to invest in expensive systems and statistics PhDs. I gave the example of a software company that uses recency/frequency analysis of support ticket creation to predict which clients are more likely to renew their license.
Finally, “Web Measurement Processes” was a great opportunity to debate under the guidance of Ian Gruber of Walmart. Ian did a great job at covering the various aspects of data governance, and there were a couple of participants who seemed to be investing considerable efforts and budgets in making sure that data is handled properly in their organizations. A good part of the session was spent talking about how difficult it is to service other divisions in big businesses when demands keep growing. Having everyone do correctly what needs to be done to guarantee reliability is still very hard. I brought up the question of what I call Results Sharing Governance, which is I believe still quite neglected by many companies, so much attention being put on data and tools. In fact, I think this is the second wall businesses hit during the long Web Analytics implementation process: having not establish clear processes and directions regarding results sharing, and decision making, many analysts realized that managers end up plainly ignoring their work.
All in all, a great event. As with all conferences, a good part of benefits comes from meeting people between sessions, and X Change offers amazing opportunities in that regard. It is a great time to meet people from companies doing incredible stuff, stuff that gets more sophisticated year after year.
Semphonic hasn’t announced yet where the next edition will be in 2012, but you will certainly see me there!