Here's the most basic thing I learned from studying mass media over the past few years: Advertisers love metrics.
For instance, there are only two particularly interesting numbers involved in buying print advertising (circulation and the price of the ad), whereas the metrics for a Web ad (traffic, rotation, cost per click, click-through rates, impressions, uniques, etc), give staffs of ambitious white-collar employees endless things to do with spreadsheets and Powerpoint. This has nothing to do with the relative value of print-v.-Web advertising, but it's a definite selling point for the Web guys.
Which is actually a problem, because as our stockpile of numbers grows, so grows our illusion of control. And the stronger our belief in illusions, the greater the likelihood of catastrophic errors.
Not that I fault people for measuring. The digital world gives us ever-expanding stockpiles of data, and analyzing it for trends and indicators makes perfect sense.
The question is, how many of us really understand what we're counting? Because if you've made one or more fundamental yet erroneous assumptions at the start, your staff of business-school-educated apparatchiks is going to generate reams of statistics reinforcing those assumptions.You'll make hundreds, thousands of decisions based on those fundamentally flawed but exceedingly precise projections, removing layer after layer of redundancy and inefficiency, re-shaping your organization into the perfect delivery system for your product or service.
And on the day that those projections stop working and everything collapses because you followed the metrics right down the rabbit hole, everyone will look at each other, shrug and say "Who knew?"
I loves me some data, but I remember that what I don't know about a topic usually exceeds my mastery. I understand what the apparatchiks can't afford to admit: We aren't in control. For all our planning and power, we are all ultimately just winging it.
It's why hiring the smart, experienced veteran is better than following the metric and hiring only cheap, specialized appartachiks. Unless you're doing discovery informatics, spotting and interpreting patterns is an art, not a science. Seriously: It says so right here, with 0.01 statistical significance.
Comments