We Know What You Did Last Summer But We Don't Know Why

Measuring clickthroughs, pageviews and revenues is revealing, but it’s a bit like asking an in-store shopper how well they like your store based only on the time of day they came in, which departments they went into and how much they bought. Do they *like* the store? Different story. Today’s web analytics tools are miraculous. They can track an individual from click to click. They can collaboratively filter that behavior to create profiles. They can, on the fl y, determine which dynamic content to serve to improve conversion. In other words, they watch what you do and deliver an offer that you are statistically more likely to accept. Brilliant. Spectacular. But the entire right brain is missing from the equation. The following scenario illustrates the point.
Michelle and Alex are expecting a baby. Alex fires up the computer to look at strollers and finds a website he likes and trusts. He looks at a variety of models and shows Michelle the features he thinks are important. Michelle wants to see them in person so the two of them head off to the store to take a look.
In the store, they see which brands are sturdy, which models are too heavy to lift or too big for their car. They find a couple of models they like and head back home to price them out online. Michelle logs on from her laptop and decides which features she cannot live without. She emails this to Alex (six feet away) and Alex finds just the brand they want with just the features they need at the best price he can get. When he tells Michelle that it’ll take 7 to 10 days for delivery and delivery is neither free nor economical, they pile back into the car, head down to the store to buy one.
Web analytics tools are great for their purpose. However, while the clickpath record of how this couple went about shopping is interesting, it is far from complete and far from informative. Two individuals clicking for several different reasons from two different computers makes for an indecipherable web visit experience.

Project Prioritization and Confidence
Those who actively make use of attitudinal data are vociferous about how much they have come to depend on it. Hearing opinions from the horse’s mouth provides process improvement, project prioritization and the confidence that the right decisions are being made.
The web development industry also makes use of focus groups and usability studies to gauge customer experience. Put a volunteer in front of a browser, give him a task (find the return policy on left-handed scissors) and watch him fl ounder. Then record his actions and ask for his feedback. The Nielsen/ Norman Group (www.nngroup.com) and User Interface Engineering (www.uie.com) have been doing this for years. They both have excellent newsletters on the subject as does Good Experience (www.goodexperience.com).
But usability testing is a time consuming and somewhat artificial process. Yes, after testing only five or six people, you can come up with a list of website mistakes that will give you a year’s worth of development tasks to fix. But how do you know you found the most important problems? The most universal problems? The ones you should fix first? For that, you need more subjects and more realistic conditions.
The web gives us the ability to interview a large number of actual customers while they are actually interacting with the site.