Precisely Wrong

One of the most closely watched metrics at the annual NFL Scouting Combine this time of the year is each athlete’s 40 yard dash timing.  At the MIT Sloan Sports Analytics conference last year a panel decomposed that metric. How few times in real life play does a player get to run 40 yards uninterrupted? The first 10 yard burst is way more important. Speed/strength after contact at 10 yard point is way more important. It was fascinating to listen to this because slight variations in the 40 yard metric can affect which round a player gets drafted in and affect millions in his contract value.

Frank Markus has been writing for years about technology for Motor Trend magazine which constantly calibrates a whole set of metrics on many cars. He recently wrote an article titled “Surrender” and says “Engine control computers know pretty well how much fuel they're metering through the injectors, but that information is not publicly shared on the CAN data bus. Even when we've invested in unlocking the rich OBD data stream, we've found it difficult to capture accurate data except during steady-state cruising, and obtaining that rich data for every (sometimes preproduction) car is impossible.” Now abstract that many layers higher and see how imprecise the MPG, emissions and other data auto salespeople use and we as consumers take for granted from the car sticker.

I have been enjoying the fact that my FitBit tells me I have been working away 2,500 to 3,000 calories a day. My weight has dropped a bit, but I got a rude surprise in the men’s store recently. The salesman said I should be asking  for “executive cut “ (ahem – you know what that means) not just 42 or 44 Long in my blazers. He laughed when I asked for the classic Brooks Brothers cut. Mea culpa – I should have read this 26 slide GQ guide to suits which shows a wide range of lapels, tapers, vents and other jacket features the tailoring world has been improvising

My point with all this is over and over in our personal and business lives even as we are into “quantified self” metrics and Big Data projects we are anchored around a handful of old-faithful units of measure. We benchmark by Yahoo finance metrics and Hackett process metrics and the underlying world has changed so much. Part of good analytics is to challenge sacred cow measures like that panel at the MIT conference was doing.