Nothing happens just because you measure. You don’t lose weight just by weighing yourself. Or maybe I did not stand there long enough? (Bjarte Bogsnes on Twitter, 16/02/2014)

Count something (Atul Gawande, Suggestions for becoming a Positive Deviant, in Better

We’ve all learnt from lean that measuring is good. But the passion for measurement is sometime, not matched by actually doing anything meaningful as a result of having that data. Ten out of ten for data, minus several million points out of ten for not actually using it.

What’s more, I’ve seen folk insist that data (x) is the most important metric, then the next week it’s ‘ah, we need to be tracking (y)’, after a few weeks you end up with a mountain of data and no sense of either what any of it means, nor what you would do in response to that data, or the experiments you’d run to move the needle. Not just standing on the scales, but skipping from scale to scale, buying a new one every other day, then deciding that what you’re really interested in is your shoe size, not your weight.

I do see two uses for measuring, though. The first is the scientific, experimental use. We hypothesise X, decide what to measure that will validate or disprove the hypothesis,

There’s also measuring for monitoring. Not just strong signals (hey, the internet’s down), but also weak signals (mmm… the engine sounds slightly different today, I wonder what’s going on). This kind of operational measurement is just as important.

We need both kinds of measurement: the first to guide us to where value lives, the second to keep us close to how that value’s being delivered.