Wednesday, September 11, 2013
Noticing their similarly low success rates, blogger Sean McBeth draws some
striking similarities between online dating and job boards. Among them:
People talk about their "type": the ideal man or woman that they see themselves dating, presumably even to eventually marry. You know you think red-heads are pretty, but if you are including it in your "type", then you don't know that it's inconsequential to being happy in a relationship, and you miss out on what a wonderful individual who just so happens to be blonde or a brunette could offer.I've both tried online dating and, when networking has failed me, resorted to applying for jobs online. I think McBeth is on to something -- at least regarding using ads to find a serious relationship viz-à-viz hiring in the professions.
So it is with job listings that say they need someone with "15 years of C#" or "MSCE Certified blah blah blah". That stuff doesn't mean anything. That could be 15 years of hiding behind far more competent teammates at a company that has a pathological problem with not firing under-performers. That cute Asian girl might like taking rides with you in your car everywhere because she's a drunk who lost her license from a DUI conviction.
McBeth concludes that the problem is that "both people and [the] problems cannot be quantified". I wouldn't put it quite this way, although this is true for all practical purposes. I think that the problem is that people and matching them to some purpose have a huge number of variables to consider. Discovering the variables required, what value range is acceptable, and what values someone brings to the table are all very difficult problems on their own. (Fewer variables or wider ranges or both make, say, finding someone for a card game or hiring a ditch digger, infinitely easier.) But, yes, attempting to solve all these problems at once with severe, self-imposed data limitations, will usually get "abysmal results".