Data Inertia or the Persistence of Uncertainty

I don’t often claim to have coined phrases, but one that came to mind a long time ago is “[tag]data inertia[/tag]”, since (in my own head anyway) it seemed to really carry weight as a real thing and keeps confirming its existence as a real thing and an accurate description.

Today I came across the story in Wired that was a perfect example.   The story, sharply and humorously described by Wired, notes the article in Slashdot that sloppily reported that cybercrime worldwide had reached $105 billion annually, a figure greater than the drug trade.  I won’t go into the details here, since Wired did a splendid job lampooning the number.

My point is that, once a piece of data has been given any credibility — and I mean ANY credibility — by virtue of being written down and reported by almost anybody (and cited again by anybody else), that credibility sticks, has inertia, that will make it persist almost in defiance of other facts.  This is an issue I constantly face in medical technology market analysis.  If that number — the size of a particular market, the number of patients with a specific diagnosis, an incidence rate, really is inert, static, invariable, you name it, then it’s a nice easy factoid to lean on.  Like 650,000 hysterectomies in the U.S. annually, like 90% undiagnosed type 2 diabetes, like the U.S. medical market (any product!) being 50% of the worldwide market, like a “flat” market must only be growing between 4% and 5%.  There are so many of these examples, I could not count them (lest I be quoted and challenged later).  Arguably, some of them are not a bad starting point for estimation, but none of them should be assumed true without annual (or more frequent) reassessment.  But that doesn’t stop my customers or my competitors (but not my authors, who I browbeat) from hanging on to those datoids like life preservers as if they were their only hope of not drowning in the sea of data that has waves and ripples (and sharks) and simply does not remain static. 

So, if you can’t hang on to static chunks of data, what can you?  Well, it’s that the chance of the “true” number being different than your assumption is directly proportional to the significance of the number in your analysis. Or that the more stable the number is, the less likely anyone really cares about it.  Or that the more insightful you may be, drawing upon your awareness of previously unnoticed trends, new developments or other hidden shifts in the market, the more flak you will get when you announce it with a flourish.  Now that’s something you can hang on to.

Leave a Reply