Recently, there was a post over at DiscoBlog on the Discover Magazine site that mentioned a study that claims that women who wax have better sex. It reminded me how quickly we can be led astray by settling on the first assumption. At first blush, one would assume that if you are a woman and you want to have better sex, wax. Reading the comments on the post showed me that the other readers (at least the ones who commented) didn't get much beyond this first level assumption. They seemed to be stuck on the physicality of waxing and how it either increases stimulation for the female, or makes a female more attractive to the male (especially for certain sex acts).
It strikes me that the appropriate inference to make is that women who wax are more tuned in to their sexuality. Waxing is just a symptom of a deeper level of engagement with their libido. It seems like a pretty safe bet that a female who thinks about sex enough to undertake a fairly involved and painful regimen like waxing would be more likely to get a desirable outcome from sex.
We fall into the same trap when looking at system usage. We tend to make the easiest assumption. For instance, I've seen "number of pages viewed" thrown around as a desirable statistic. The more someone clicks around your site, the more they like it... Right? Well, the truth is probably a bit more complex than that. It could be that they couldn't find what they wanted and began desperately clicking around. What about "time on site" as a metric? Does it hold that the longer someone's on your site, the more of their attention you've received? Could it also be that people pull your site up, and after not finding what they want, the open a new tab and start looking elsewhere? The only way to answer these questions is to ask the users.
I was working with a prospective client a while ago. They were building a new system for their agents. They were very focused on whether or not our persona development met the standards set by Forrester. I told them that, while our personas were in line with Forrester's standards, that they had a finite user population. They knew who there users were. They knew where they worked. Personas are a useful way to create a contraction for a large body of user data. But, given the choice between talking to data and talking to people, choose people every time. There really is no substitute for ethnographic research to get a feel for the way users interact with you system. You also get a feel for how your system fits into the larger context of the user's world. There is no data that will show you what else is going on while the users interact with your application.
If the folks who conducted the waxing study had spoken to the respondents instead of conducting an internet study, it's very likely that their findings would have been, "woman who are focused on their libido have better sex." However, this wouldn't have made as good a headline.