Subscribe Bookmark



May 21, 2014

5 Questions with Avinash Kaushik of Google

Avinash Kaushik, Analytics Evangelist for Google, talked with John Jones of JMP while both were at a conference in June. Here’s part of their 20-minute conversation, which actually included more than five questions and answers.

Kaushik’s answers in the Q&A below are just the beginning of his engaging and informative discussions on his job, the importance of Web analytics, his blog and the need for "a true data democracy." The full transcript of their chat is available, as is the podcast.

Kaushik will be a keynote speaker at the Innovators’ Summit this September in San Francisco, where he’ll present on a panel titled “Extending the Reach of Analytic Visualization.”

John: What does it mean to be an Analytics Evangelist?

Avinash: I think the title “Analytics Evangelist” doesn’t mean anything, which means I can do anything I want. But usually I tell people I’m the author of the book Web Analytics: An Hour a Day; I love blogging; and then the third part is Analytics Evangelist. Interestingly, my title is an evangelist, which might imply an external focus, but most of my role at Google is inwardly focused towards the company….

John: Why are Web analytics important for an organization today?

Avinash: There is a fundamental shift in how consumers are being marketed to, how consumers consume content, how they listen to messages, and how they behave. And increasingly, online is a part of it. My friend, David Hughes in the UK, has this term he calls nonline marketing, and I love the term because we used to have offline (what we now call offline marketing), and of course then we had online marketing for the longest time. And we still call it online marketing, but in reality, I think our persona has moved very fluidly between on and offline depending on the task that we’re trying to complete….

John: In your book Web Analytics an Hour a Day, you discuss an approach you call the Trinity Framework. Can you elaborate on that?

Avinash: Web analytics used to be only clickstream analysis: Collect the log files, parse them, report hits out, and that used to be the world for the longest time. But I found that clickstream data by itself are not very good at identifying actionable insights, and they were also not very closely tied to outcomes.

So while at my prior job at Intuit, I created this new framework of thinking of mind, new mind set to Web analytics, and they call it Trinity. And one part of it was understanding customer Behavior; it’s what do people do on our Web sites? But the second element of the Trinity I very quickly went to was Outcomes. So, all this behavior on our Web site happened; what was the outcome? What was the outcome in terms of revenue and order size and numbers of leads?

And then it took me a little while to get to the third element of the Trinity, which is understanding the customer Experience. The primary thing I wanted was, when John came to my Web site to buy this product as an entrenched customer, was he satisfied? Was it a pleasant experience?…

John: Your blog is called Occam’s Razor. Why did you choose that name?

Avinash: William Occam was a 14th century monk in England, and I really love his quote. In English, it loosely translates as: “Of all the explanations for something, often the simplest one is true.” It connects with my heart a lot because my experience when coming into Web analytics was that we tend to make it way too complicated. Unlike any other medium in the world, we’re collecting literally megabytes and terabytes of data a day, every single thing about a customer. And there is this tendency of all Web analysts and Web applications to essentially become data pukes. Paralysis by analysis is the norm. Often in the middle of cookies and sessions and URL parameters, timeouts and exits, we forget that people use our Web sites. People are not that complicated. People are simple, and they’re trying to do simple things….

John: You will be speaking at the Innovators’ Summit this September in San Francisco. The topic for your panel discussion will be “Extending the Reach of Analytic Visualization.” With regard to Web analytics, how does the combination of deep analytics with visualization empower innovative thinking in an organization?

Avinash: The traditional BI world, and the world of statistics and predictive analytics, is ruled by the few. You have some few brilliant statisticians, mathematicians, analysts – very few people in the company who have the kind of skills set to use the kind of tools that are necessary to find the answers that the business needs. And so everything goes to these people. They really work hard to find really golden, awesome nuggets of insights that can then get back to the business that some point will get actioned. It was OK for this process to be run by the few and for the process to be a little bit slow.

The thing is, the Web really blows that apart. It changes with every passing minute, and you have to react at a pace of light. I’m not even exaggerating. And everything you used to do in three months, you need to do in the Web in every three days, at least three weeks, at least not three months. What this calls for is this fundamental shift from the few to a true data democracy….

Article Tags