My son’s participating in a (virtual) summer reading club. He’s supposed to set a goal to read a certain number of books each week, and track them on a calendar. If he’s successful, he’s eligible to win a prize.

My son’s already a big reader. We decided that maybe his goal for the first week should be 14 books. On the first day, he read two books. So far, so good.

But my son’s very clever.

At the end of the next day, I asked him how many books he’d read. He proudly told me that he’d gotten through fourteen.

He’d changed his strategy.

He put it together that there were no requirements around the length of the book. So, instead of reaching for the chapter books he has, which are more appropriate to for his reading level, he pulled out a box of books from when he was younger. These are books he can probably get through in a couple of minutes each.

Goal achieved. Technically, anyway.

If nothing else, it’s a good example of how when we use quantitative measures to track behaviour, it provokes change behaviour.

It’s something akin to the observer effect. When we are aware that some of our activities are being counted, we start to prioritize behaviour that we know will be “seen.”

This happens at the individual level, like with my son. As Goodhart’s law tells us, once a measure becomes a target, it’s no longer any use as a measure. As I like to put it, what gets measured gets manipulated. My son changed his reading habits to “game” the system. In other cases, we might become more risk averse, out of fear that we’ll expend resources on things that don’t count.

This can fundamentally change our relationship to our work. As Jerry Z. Muller points out in The Tyranny of Metrics, not everything we do is easily measurable. But once our performance is tracked, rewarded, or punished based on quantitative measurements, we start to prioritize work that can be measured. We made a trade-off, sacrificing other aspects of our job that perhaps not so easily quantified.

Think about that. Of all the things that we’re responsible for, which are the most easily counted? Are they really the most important things to your success?

When we start trying to reduce performance to numerical outputs, we lose a lot. It’s a kind of lossy compression: there are aspects of human behaviour that simply do not translate well to scorecards or dashboards. When we try to force our activities into those kinds of charts and graphs, we sacrifice a lot of rich data. There’s always going to be a gap.

Can you measure, quantitatively, the quality of something you’ve written? Of a presentation you’ve delivered? Of how well you collaborated with others, or the camaraderie you’ve built with your team?

What about your last great idea? Was it something that “counted” on your personal performance indicators?

When we start to reduce performance to quantitative data, we risk alienating ourselves from the very things that make many of us good at our jobs. Our aversion to risk intensifies, and we start to avoid the big problems where we might actually provide the greatest value out of fear that it will hurt us on our scorecard. Our expertise shifts from intrinsic dedication to our craft, and toward the more extrinsic motivation of manipulating the metrics. We feel more like cogs processing inputs into outputs, and less like, well, people.

Organizations should be supporting those employees who are willing to go out on a limb in support of a good idea—even if it isn’t something that will “count” against a standardized benchmark.

But that means setting aside the idea that everything that counts can be counted, and that everything that can be counted counts.