eMetrics was the usual whirlwind of meetings and presentations – plus a very enjoyable lunch with a number of local or semi-local Semphonic clients. I think it was the best eMetrics I’ve attended in the last few years. I didn’t get the chance to see as many presentations as I would have liked since I had quite a few meetings and some real work that had to get done, but there were some excellent keynotes.
James Robinson of the New York Times did a really outstanding presentation on some of the ways they are using web analytics data “outside-the-box” to improve their business. There are powerful uses for web analytics data in the broader market research world – and this presentation has some great, thought-provoking examples. My favorite (using late-afternoon traffic to help set print-runs for tomorrow’s edition) was simple, obvious and beautiful. Everything a really good analysis should be.
In my experience, deeply impactful analysis is often simple; birthed not from fancy tools or complex methods but from a fertile imagination able to see uses for the data that the rest of us have overlooked.
Joe Megibow’s presentation on Voice of Customer and CEM analysis at Hotels.com was also fantastic. I don’t think any presentation I’ve seen illustrated (completely implicitly) the degree to which the success of analytics is a function of culture. Hotels.com seems relentless in their pursuit of customer service and their use of analytics to build a zero-defect web site. As with the New York Times presentation, what was impressive wasn’t the quality or complexity of the analysis but their commitment to using the data they had to make the business better.
These guys take customer complaints seriously (“if even a small number of customers say there’s a problem, believe them – there is a problem”) – even to the point of regularly picking up the phone and talking directly to those customers. Even better, I thought, was the insistence (as exemplified in their registration/loyalty program) of giving customers a reason to do what you want them to.
No amount of process engineering is going to make registration attractive to visitors or significantly reduce fall-out effects unless you actually give people a good reason to register. In our business, the temptation is to think there are always tactical solutions to web site problems and that multivariate testing of copy and colors can somehow make up for a bad deal to the customer. But the best solutions are often both deeper and simpler. If you lose visitors when you force them to register, it likely isn’t UI design but simple cost-benefits that are responsible.
It made me think back to my recent Geico post – what a refreshing difference to hear a presentation on analytics so firmly rooted in the idea of providing a truly superior customer experience.
I enjoyed giving both of my own presentations. It was great to present with Mark Ruzomberka of Traffic.com on taking SEM programs in-house. In a way, this presentation turned out quite a bit broader that the immediate topic – since it really dealt with ways to understand where your real optimization opportunities are and how Traffic has changed their PPC program to reflect those opportunities.
I was worried about attendance for my second presentation (after lunch on the last day of eMetrics can be deadly) but I had a great turn-out and I enjoyed the presentation on behavioral segmentation. As I wrote that last sentence, I realized it might sound a bit odd. But just as I enjoy hearing some presentations more than others, giving them turns out to be similar. Since I rarely give the same presentation, I often find myself surprised to be enjoying or suffering through my own presentations! This was a fun one.
The questions and comments afterward made me realize that one of the most interesting aspects of the topic was the process of getting web analytics data from a SaaS tool and using it in a data warehouse. So while I touched on some of this in my previous series on behavioral segmentation, I’m going to embark on a new series that delves into aggregating and warehousing web data. Sophisticated organizations are increasingly focused on this topic and it is far richer than a casual acquaintance might make it appear.