Software, technology, sysadmin war stories, and more. Feed
Thursday, December 20, 2012

97% happy customers, sure, but what about former customers?

I try to zap past as many TV ads as possible, but now and then I'll catch a chunk of one while watching something else. It's usually the last 5 or 10 seconds of the last ad in a package. I think the introduction of Tivo and the "jump back after pushing play" correction factor has probably turned that time into the most valuable time now. I'd love to hear from someone in the TV ad business if that's actually true.

Anyway, one of these blipverts said something about how 97% of their customers were happy with the service they were getting with this big insurance company. Just from seeing that, my head filled up with all sorts of ways for that number to be gamed and/or weasel-worded so as to become useful.

If you take it at face value, then okay, perhaps 97% of their customers are in fact happy. That means 3% of their customers aren't exactly happy. That looks like a pretty good situation, but it's missing another group of people: former customers. You'd think that the ones who were really unhappy with a company would bail out for some other place. It's not like they have a monopoly on insurance, right?

It seems like a lot of these numbers are set up to make you think there are only two groups (happy people and unhappy people) when in fact there are at least three: those two groups, plus the group of unhappy former customers who won't show up on the numbers. Those are the ones you'd probably really want to know about if you were trying to make an informed decision, right? We'll assume that the (much larger) set of people who have never been customers isn't interesting and so can be rightfully omitted.

The same sort of weaselly behavior might apply to these so-called "best places to work" surveys. If you only ask the people who are currently employed at a particular place, how can you possibly see the whole picture? I submit that this omission is used to confuse people into seeing things which may not actually exist.

Consider that both employees and customers can be "managed out" just by making their experiences suboptimal. You don't need to explicitly fire them if you can just sour the milk so much that they want to quit, right?

I have another thought about surveys in general. If you're really fed up with something, are you even going to bother to answer one about how happy you are? Maybe you're stuck in a situation until something vests, or matures, or otherwise pays out. During that time, are you really going to answer a "happy check" questionnaire? I bet a bunch of people will find creative reasons to "forget" to fill them out.

This is why I always wonder about what sort of signals can be inferred from the responses, or lack thereof. It would be interesting to see if anything pops out of the data. For instance, maybe formerly reliable people stop filling them out 6 months before they quit.

If you're really "over it", it might be an interesting chance to find out how many "required" things really aren't, especially if you always obediently filled them out before. You might just discover that a great many things aren't worth the paper they're printed on.