I’m wrapping up day 2 of the eMetrics Marketing Optimization Summit down in D.C. If you want full and detailed coverage of the summit, check out my liveblogging posts over at Search Marketing Gurus. This post is all about my own personal reactions to the speakers, announcements and sessions.
First up, were my predictions any good?
Prediction #1 – Google Analytics v3.0 Will Be Released
I was right that Google Analytics announced some new features, specifically detailed tracking of internal site search (maybe Jeff Gillis listened to me during my interview with him), automatic tagging of outbound links and, most notably, tracking of rich internet applications. The internal site search data is a gold mine of qualitative data about what your users want. GA will now hook into any search app that uses parameters and deliver their full suite of site usage and outcomes (ecommerce) data. Most interestingly, they will show you query refinements–an excellent window into the quality of your search results. Naturally, all of these data can be mined for the benefit of your paid search or even planning of products or content to satisfy unmet site needs. They’ll even give you data about who searched and where they initiated their search (did they get stuck on your product page? help page?).
Lots more, so read on.
Second, GA will now autotag outbound links. Before this, you had to do so manually or install a bit of code that might hack clicks on outgoing links into your reports. I did and here’s what it looks like. 
Now that this is out of the box, you’ll get a full set of reports about the data. This is most relevant for sites whose business thrives on driving visitors to partners, affiliates, advertisers, other sites within a brand’s network, etc.
Finally, Google Analytics is pushing forward with rich internet application tracking and advancing their own classification system. It breaks down into 3 pieces
- Object – The RIA you want to track, say a video play
- Action – Control within the object, like a play or pause button
- Label – The name of the RIA, say Leave Britney Alone
Google is trying to determine exactly how the rich media world is measured. Brett Crosby even reached out to the WAA Standards board to discuss this as a standard. Google’s Analytics Evangelist and Commerce360 Advisory Board member Avinash Kaushik is on that board.
They did not, however, address historical data. I was chit-chatting with the very passionate Dennis Mortensen of IndexTools and he noted that at the volume of sites Google Analytics is serving, it would be a mammoth technological undertaking for GA to provide that sort of data. I will confess to being no great computer science mind on this topic, so perhaps my prediction was technologically naive.
Additionally, they did not address the need for additional segmentation. Frankly, this is a sore point for me. Data in aggregate never really tells a story. I’ve coined a phrase, “Averages Lie”. How hard would it really be for GA to offer out of the box segmentation of data by channel? I can do that now if I setup my profiles and filters right, so why can’t they automate it. *end segmentation rant*
I’m realizing more and more that tools are still in their infancy and that you may need to run multiple tools to get the data you want. In the case of exploratory analytics and segmentation, I recommend ClickTracks.
Prediction #2 – Microsoft’s Gatineau Brings “The Who” to the Masses
Ian did indeed show off Gatineau, though the features he highlighted did not expand beyond what’s been mentioned before in the blogosphere. I won’t go into details here, since you can read my article about Microsoft’s web analytics project Gatineau over at SMG. I hope that even more features get rolled into Gatineau as it extends into Beta 2 and it’s hard to really assess its value without being in the tool. Here’s what I think works:
- More emphasis on data discoverability – the tree maps, the calendar of marketing campaigns and the timeshifter all help add context and display data in a way that lends itself to inquiry. This is a key area of analytics that Google really pushed to the forefront and Microsoft is clearly taking to heart. Why is data visualization and data discovery key? First, it helps you translate data into insight into action. The more you act on your data, the more likely you are to use it again and that pays off for the vendor. Second, it rejiggers the world of tables of data into thought provoking and eye catching charts and comparisons. This helps the web analytics newbie run past the data manipulation phase and into the data analysis phase. It also eases the sense of data overload.
Now, I think there’s definitely still room for improvement.
- Demographic Data – Ok, who is coming to your site and what is there age, gender, etc? Gatineau lets you segment your site data with this kind of information BUT it comes from the profiles of people who signed up for Microsoft Live profiles. First, that’s not everyone. Second, the sample is probably not representative of the US population, let alone the population within your vertical. Third, it’s self-reported data (I’m sure plenty of people make $250K). Finally, it did not appear that Gatineau reports this data with a “No data” field to represent the portion of the population not accounted for by Live IDs. I’m afraid there will be a false sense of confidence.
- Attribution - This is the concept of determining which campaign gets credit for a sale or other business outcome. Typicially it’s first campaign, last, all equal or some weighted average. One of the alleged selling points of Gatineau was that it addresses attribution better and helps you see interaction effect better. As far as I can tell, the only way it currently does this is by letting you enter a calendar of your marketing campaigns and plotting the results of those campaigns on one chart. This is great data visualization, really, but it doesn’t solve for attribution.
Microsoft is playing a major game of catchup with Google Analytics. GA went with the buy model with their purchase of Urchin. Microsoft is going with the build model and it’s slow going. If I were them, I’d consider purchasing a bunch of smaller vendors with features, people, processes and infrastructure that you can suck into Gatineau to catch up fast.
More tomorrow…
