Monday, February 25, 2013

Under the Lamppost

The drunk had been wandering around the lamppost for a few minutes. Finally, another person wanders over to him. "Can I help you?" he asked the drunk. "I lost my house keys" was the reply. "Here?" the man asked. "No, over there in the alley," the drunk answered. "Why aren't you looking in the alley?" "The lights better here."

We treat data the same way. We will analyze the data we have, because the 'lights better' rather than do the difficult job of getting the data we don't have. Customer surveys are sliced and diced till we understand every demographic profile, product feature and service desire we can. We treat the conclusions as if they're gospel truths.

We forget that the data is biased. When was the last time you filled out a survey? Do we really have a random sample of our customers, or only ones that take the time to fill out surveys? A lot of times people self-select themselves out of participation and they rationalize why they shouldn't fill out the survey: don't have anything to complain about; don't have anything to rave about either; too busy; not a loyal customer; one-time purchase; responded before and nothing's changed; responded before and I keep coming back; don't believe anything will change; don't want anything to change...and so on.

The same thing happens when we look at computer data. It just happens to be the data we've collected, not necessarily the data we should have to answer the important question. Without data, we are allowed to rely on our intuition, hunches and hopes....and the few anecdotes of the customer experience or process that we remember. And memory is a plastic thing relying on what we want to see and remember because of an emotional connection. We discard other bits that don't seem so important.

Are you looking at information because it's the easiest or because it's the most meaningful?

No comments:

Post a Comment