A great friend and mentor, Avinash Kaushik, presented at Adchemy last week. Avinash is an analytics evangelist and industry visionary. He has two books on Analytics and speaks frequently. I worked with Avinash a few years ago at Intuit where we had a very prolific and fruitful working relationship. The whole company turned out to see him do his thing.
Avinash’s style is pretty unique. He is exciting, entertaining and utterly authentic. His stories stem from real experience and his point of view has been crafted over years of hard work. He has a great sense of humor, although it skews towards the bawdy. The great thing about Avinash’s talk was that he didn’t devolve into mathematics functions. There isn’t a magical formula you need to do what he is asking. You just need to be thoughtful, skeptical and objective. Too often “data” is anything the analyst says it is. You could make a decision on “data” that is ill-informed and doesn’t actually help you with your goal. Other times data is just noise which obscures the truth of a situation. Avinash is excellent at drawing attention to the underlying truths and not just the numbers. If I could sum it up what I learned from the presentation it was “Figuring out which numbers matter is the most important and least understood thing about analytics.”
To illustrate one of his points, Avinash had a case study about Barack Obama’s presidential campaign site. He had several different possibilities with videos and different images plus calls to action. He said, “You probably won’t guess the winner.” I answered immediately in my head, “Soft fuzzy family shot, and ‘learn more’ button.” I thought the question was simple. (Video loses to static image. Family shot loses to loner shot, and Join Now loses to Learn More). There wasn’t even a slight question in my mind. This made me recall my first meeting with Avinash.
I was sitting in his office with Florence Tang, a graphic designer. The first thing he said was, “Make sure it’s clear: You guys don’t know anything. Testing, that is how we know things.” I asked the same question that I would ask today, “Who gets to decide what to test?” You can’t test everything. Who designs the tests? At Intuit, the answer was often (unfortunately), “Business people and marketers.”
In his presentation, Avinash did call out WHY as an important factor in testing and analytics. Just testing “what won?” is insufficient to learn anything. Why did one test win over the other? What process in test creation/production contributed (if at all) to the winner? Is the testing just improving our situation on local max island or is it potentially moving to the big island? WHY something is better allows you to learn and incorporate the learning into future tests. Just a “winner” and a “loser” doesn’t give that insight.
On that topic, Avinash had one slide with concentric circles.
This slide kept itching my mind. I think my problem with it is that it shouldn’t be displayed this way. Competitive Analysis is not subsumed by clickstream. Maybe if it was like this:
Am I nitpicking Avinash’s presentation here? Yes, undoubtedly. I think Avinash was a engaging and educational speaker. Pointing the obvious is often extremely hard. I am hoping that his message of thoughtfulness, skepticism and objectivity is taken seriously by all those who see him. He isn’t just entertaining!