Analytics 2.0 with Avinash Kaushik, Google

I’m at Darden today, speaking at an executive education course.  Great presentation from Avinash Kaushik, Analytics Evangelist at Google.  The big idea of the morning?

Analytics 2.0:
– [start here] Clickstream data (the what)
– Multiple outcomes analysis (the how much)
– Experimentation & testing (the why)
– Voice of the customer (the why)
– Competitive intelligence (the what else)
– Insights (the gold) [end here]

Can your company answer the question: "why does your website exist?" in 15 words or less.  Craft this statement around business outcomes.  Conversions aren’t always sales – they can be interim goals, e.g. could be brand referral or propensity to visit a store.

The job of the online marketer is "learn to be wrong, quickly."  The web allows experimentation (A/B testing) at scale previously very difficult.  Google did this on Picasa.  Current page showed 30% lift.  "Free download" text doesn’t work as well as a button that says "try now." A picture showing functionality didn’t work at all (i.e. trying to show how much better than Photoshop).

Three key questions to figure out site intentions.  Most sites have conversion of 2%.  Why?  Not everyone came to your site
to buy something.  Goal:  to find segments of discontent – and respond
to them at scale.  So answer these:

  1. Why are you here?  (aka primary purpose)
  2. Were you able to complete your tasks today?
  3. If you were not able to complete your task today, then why not?


  • Insight One:  the 10/90 rule.  Tools and Professional Services:  $10 vs. Investment in intelligent resources:  $90
  • Insight Two:  Data quality sucks, get over it.  The web is the most perfect imperfect medium in the world
  • Insight Three:  Reporting is not analysis.  Empower your analysts.  The critical few is the secret – on any dashboard you should have max 6 – 8 metrics.

Yahoo! up next.