Binaries and borders
A couple of weeks ago, I was sitting in on a deep dive analysis presentation that one of our senior analysts was giving to a high-value client. I’m not really a hardcore numbers guy, so when the analyst starts parsing satisfaction data for very specific customer segments, I’m sometimes prone to taking a peak at the Blackberry and letting my attention move away from the presentation (slightly and momentarily).
But then I heard the word “Iraq” and my head snapped up. The analyst was reviewing some open-ended comments that pertained to why certain visitors didn’t complete a purchase during their online session. The verbatim comment in question was from a parent whose son was shipping off to the Middle East, and it ran something like: “I’m looking for an affordable laptop for my son, who’s leaving for Iraq next month, but I was unable to find anything within my budget range.”
Powerful words, huh? Think about how we usually analyze conversion data. Yes and no. Simple binaries. We have that 2% that converts online and that 98% that doesn’t. And what does that 98% constitute? Unrealized opportunity? Virgin territory ripe for harvesting?
Actually, it constitutes real people, operating within the strictures of real-world circumstances. Sometimes, cart abandonment means that the user can afford college tuition for their kids or a new computer, but not both.
Therein lies the burden of survey research. It’s not always clean and pristine numbers; sometimes…it gets messy. You actually get your hands dirty, because you’re drawing on the raw, and sometimes emotional, verbatim output of why visitors didn’t buy.
Like I said, I’m more right brain than I am left, but I do know that clickstream analysts have got it easier. Pull the data for cart abandonment, slice by frequency of visit, referrer, product line—then, power down your machine and head home. The survey research analyst is the one who has to wade through the tricky morass of visitor frustration, of dashed hopes, and of foregone needs, and then put distance between themselves and the stream of commentary in order to provide objective insight.
But I guess it’s an added motivator. When it comes to making recommendations to clients, you’re advocating on behalf of people, rather than numbers.

Good post Jon. Lots of things can explain abandonment: a credit card being maxed out would be one of them. Or any other of the wide variety of reasons why someone doens't go through with an expected action here and now.
Behavioral analysis cannot dig efficiently into that dimension of the web experience; only attitudinal analysis can.
It escapes me why so many organizations still do not devote time and effort to doing it.
Jacques WarrenDecember 03, 2007