Making Measurement Count

There’s a saying I’ve heard many times that goes something like this: what gets measured gets done.

And though I completely agree with that saying, I think it misses a crucial point: that before measuring anything, we have to make sure that what’s getting done is what you want to get done.

After the army started testing for push-ups in their standard physical fitness tests, I started training for it. After they stopped testing for pull-ups, I stopped training for it.

I now do 50 push-ups without too much difficulty. I used to struggle at 30.

I now do 8 pull-ups with difficulty. I used to be able to do 12 with both hands tied behind my back.

Both were intended to act as proxies to my fitness. So, can someone tell me: am I less fit or more?

To answer that question, I propose we ask two further questions, the answers of which will then determine the answer:

• What’s the definition of fitness?
• Which exercise, pulls-ups or push-ups, is more closely correlated to that definition?

It would be great if fitness could be measured directly, but t can’t. So we use proxies or indirect measures such as pull-ups, push-ups, or running times to help approximate it. Sometimes these proxies make sense, but sometimes (many times) they don’t (e.g. how high or far you jump is often used as a proxy for fitness, but it’s actually probably more technique than anything else).

Proxy measures are used all the time in business as well: the performance of a business is often measured via financial proxies (revenue, costs, profit) and these often make sense – when there’s no profit for too long, companies go bankrupt and fold.

But many companies don’t stop there (nor should they). Things like staff turnover, lead times, number of customers and more are also used as proxies as well to determine if a business is “doing well”.

I do worry sometimes though, that in our quest to measure everything in order to understand where we’re at and seek “improvement” in some process or activity, we forget to ask what our true end-goal is (the formal definitions of what we’re trying to achieve), and whether or not the measurements we have in place are truly correlated to that goal.

What does it really mean for a business (or business unit or employee) to be “doing well” (in the context of this organisation)? And are the measures we have in place to measure if the above are “doing well” really correlated with what we have define as “doing well”?

Because if we don’t, we could well be working hard at improving the proxies without improving anything of real value.