Interview With Avinash Kaushik, Author of Occam's Razor

Sonicko President Jeff Lawrence recently sat down with Avinash Kaushik, author of the popular web analytics blog Occam's Razor about his views on web analytics, what he hopes to see from Microsoft's upcoming web analytics application, and Web 2.0 technologies. Avinash has also authored an upcoming book Web Analytics: An Hour A Day which you can preorder on Amazon.

1. You've been in the field of web analytics for quite some time, did you just wake up one day and think to yourself that this is something that you wanted to do, or were you thrown into the role and simply adapted to it?

At my last job with DirecTV, Sr. Manager for Enterprise Analytics, I had small amounts of exposure to Web Analytics (someone supplied log file parsed numbers into the dashboard). When I interviewed for the job at Intuit (Manager for Web Analytics) I was quite excited about the possibility of taking all my experience in Decision Support and apply it to a 100% exclusive web environment.

There is something so beautiful and scary and challenging and fun about data on the web. It was too hard to pass up. But it would be fair to say that when I took the job at Intuit I had no idea what "web analytics" was, I had not yet had the fortune to have used any web analytics application. Blaire Hansen, my hiring manager, certainly made a huuuge leap of faith in hiring me.

It has been a amazing ride and yes to answer your question I have simply adapted to it, but since my post MBA experience has been almost solely focused on Decision Support Systems I think I have brought all the learnings from traditional data warehousing and business intelligence and applied it to my current role.

2. What problems if any do you foresee with the implementation of Web 2.0 technologies such as AJAX and the explosion of tab based browsing? Are you concerned about problems of people keeping tabs open when they are not actively browsing the site?

I have blogged about the fact that slowly but surely the page paradigm is dying. That is not saying that the big problem is that the page view metric is going to be crap. It is more that currently almost all web analytics applications are constructed, from an architecture perspective, on the fact that a page view has to happen and all things go from there.

The challenge with AJAX or Flash or RIA's is that the page view model does not work (remember this is more from a data capture and data analysis perspective). I have blogged about (http://snipurl.com/1al4p) moving to an "event" based data capture and analysis model. The challenge for the web analytics vendors is to change their underlying architectures to accommodate for a event based model and not just try to stuff events into page views because that won't work in the long term.

There is a bigger challenge from Ajax / RIA's for business users. We are used to slapping a tag on the page and expecting most data we need to show up. With web 2.0 we are going to have to think way up front exactly what we want to measure, what is success and then instrument these new experience in Ajax or RIA's to give us the data we need. Most companies and practitioners are not yet prepared for this mental shift.

In terms of tabbed browsing, this simply exposes one more limitation of clickstream data to ultimate actionable insight. Even with tabbed browsing it is hard to make sense of clickstream data. My personal point of view with tabbed browsing the order in which a visitor sees something and follows a "path" is fairly messed up. Depending on how your sessions are initiated in terms of your analytics apps this could also screw up other reports such as referring urls etc.

But it is important to note that if you have a tab open and as long as you are not auto-reloading it, the web analytics application will terminate your session after 30 mins of inactivity and that is not a biggie. Tabbed browsing, and its related impact on clickstream data, is yet another reminder that each website owner should have a mechanism to collect qualitative data (http://snipurl.com/1al54) and to have a Trinity mindset (http://snipurl.com/1al53). Without that they will not truly be able to get actionable insights from their data.

3. What tools, features, and reports would you like to see in the upcoming Microsoft Gatineau product?

Hmm…. I don't think my friend Ian Thomas has quite the luxury to build whatever I want, but let's assume he does. J

I hope that with Gatineau Microsoft figures out exactly who their target audience is and then delivers a tool exactly and specifically just for that audience. Being all things to all people means being nothing to anyone. I guess I am saying I hope their tool does not have a billion standard reports out of the box, just the six that their target audience needs. Atleast initially.

Efficient segmentation. In four clicks (see I am generous!) anyone should be able to segment out traffic from the search engines or from a top referring url /'s or visitors who see x number of pages or come on a particular campaign (whose id is in the url or cookie). It is very hard to dumb down the ability to do intelligent segmentation, yet that is the key to finding actionable insights.

Some useful reporting for Search Engine Optimization. I love free traffic and with all the changes (especially at Google, such as increased "personalization") the PPC gravy train is going to pause. SEO will become more effective at getting the right kinds of traffic yet today most tools pay lip service to the measurement of the results of SEO efforts, all you can do is measure organic traffic and if it goes up (that is hardly a measure of SEO). I hope Gatineau can at least tap into the MSN data and providing efficient reporting for atleast MSN SEO efforts.

Ok maybe I will ask for a reporting feature. I hope that all the reports will show one extra time period by default. For example show eight days in a "weekly" trend and thirteen months in a "yearly" trend. Seems like a small thing but most web analytics tools are not great at giving context, and context is king. If you look at a eight day trend you could compare this Monday to last Monday and get a feeling for if you are doing better or worse this Monday, with most tools you don't see last Monday. Ditto for this month vs. same month last year. It gives context to your past performance and is a "internal benchmark" that can frame current performance. Might not scream answers at you but will get you to ask the right "why" and "what" questions.

There is nothing uniquely Microsoft Gatineau about the above three requests, though if they are really starting with a open mind it might be easier for them to consider requests from random bloggers such as myself.

4. Do you foresee a decline in the major players in the web analytics field such as Omniture and WebSideStory based upon free web analytics packages, or do you believe that they fulfill a niche and will remain?

It is important to realize that I am a practitioner. I am not a vendor, I not a consultant, I am not a analyst from Forrester or Jupiter or any other esteemed organizations. In as much I probably have no idea what I am talking about when I answer this question.

One overall fact to consider is that the web analytics space is growing by leaps and bounds, driven by the fact that the web in general becoming a medium that is increasing been monetized (to huge amounts). At the moment anyone in the field can do great because web analytics is a baby and the there are way too many people who are falling in love with this cute baby. Near term there is hardly a worry on the horizon.

Longer term both Google and Microsoft will prove to be excellent disruptors. If they provide solutions that are value add (rather than being YATR - Yet Another Tsunami of Reports) and keep improving, as their deep pockets would enable them to, then they should own the small to mid sized clients. There is really no need for you to pay for clickstream reporting and some of the always required analysis.

That leaves some of the mid-market and the "high end". These will continue to be with the paid-vendors for some time for a whole host of reasons, for the next couple years at the minimum. After that the paid-vendors that and provide more than clickstream analysis (or indeed web analytics) will thrive (those that enable what I call the Trinity Strategy - http://snipurl.com/1al53). Others will feel perhaps more than bearable pressure from the for free vendors and get squeezed.

In a few years it will be hard to find vendors who will just do traditional web analytics and will be paid for those services. There I have gone out on a limb!

5. We know that you've become quite a football fan, but what else do you do to unwind after a long day besides blogging and answering questions?

By the time I am done with the blogging, and answering atleast twenty fairly detailed emails from the blog readers, it is usually around 0100 hrs and that is rather late! So I unwind by going to sleep. J

With a full time job, the blog, emails from blog readers, the book, speaking engagements and seminars, two small kids and travel required by business it is really tough to find time. No tv for the last year, the super bowl game was the only game I saw all year long (and boy was it nice).

I suppose for me unwinding is writing my blog. It is way more work than I ever imagined (approximately twenty hours a week at least). But when I write I am fully absorbed in writing, I get an absolute thrill when I get comments, it makes me happy beyond reason when I get emails from readers who appreciate the small amount of wisdom that is one the blog. There is no monetization tied to the blog for me (just lots of work!) but it is great feeling that in my own small tiny way I can help someone in Sweden or Iran or Australia or Brazil or Russia or South Africa or Canada or many other places.

I suppose few things are this much work just to "unwind" and few have such delightful rewards.

Avinash is the author of the upcoming book Web Analytics: An Hour A Day http://www.webanalyticshour.com and also the highly rated Web Analytics blog Occam's Razor http://www.kaushik.net/avinash

By day he is the Director of Web Research & Analytics at Intuit Inc where he is responsible for the business, technical and strategic elements of the web decision making platform supporting 60 plus Intuit websites.

His professional career has been focused on Decision Support Systems at Fortune 500 companies such as Silicon Graphics, DirecTV Broadband, Intuit and DHL in Asia, Middle East and the US.