,
20 tweets,
3 min read
Read on Twitter
This is one of the best books I’ve read this year. Go read it! It articulated the key problems that can happen when society over relies on measurement. 
1. First, the good things about measurement. If you measure a problem, you can solve it better, because you can figure out where you're failing. If you have 5 employees making widgets, without measuring worker production you won't know who's slacking and making too few widgets.
2. So you measure how many widgets each worker makes. A few problems can arise.
3. Workers know they are being measured, so they start to cheat the system. Maybe they only work on days the inspector comes around. They could end up being less productive overall, because you are measuring productivity.
4. Workers are incentivized to cheat by reducing widget quality. If you measure the total number of widgets produced, they'll make lots of crappy widgets. (In academia, this is done through p-hacking, small sample sizes, salami publishing, etc.)
5. But if you measure widget quality, the workers will produce very small numbers of high quality widgets, but fewer get produced overall, and maybe average widgets are good enough. People will give you more of whatever quality you measure.
6. Measuring can sap morale. Workers who once came in loving their jobs may writhe at the idea that management sees them only for the number of things they produce, not as the human beings they are.
7. Measurement becomes a goal in itself that distracts from the primary goal of the organization. Schools become places where the primary goal is increasing test scores, whereas the goal should instead be learning. Subtle difference, but huge.
8. Administration builds. E.g., in universities, the primary functions are teaching and research--both done by profs. Admin develops to make sure profs are doing those jobs effectively. They hand out surveys, find some useful tidbit and pass it to profs.
9. But there are only so many useful tidbits. And there is a cost to handing out surveys: every survey a prof takes is time not spent doing teaching or research. Admin doesn't count these costs, so they keep increasing surveys and forms, to keep profs accountable.
10. The proliferation of admin eventually decreases university performance because it takes up profs time, while also increasing costs. Paperwork time is huge, and counter productive.
11. Externalities. Things that aren't measured don't influence decision-making. So companies measure profits, but not environmental damage? Expect a lot of profits and also a lot of environmental damage.
12. Short term thinking. If you measure a company's performance quarterly, the executives will target programs that help right now, but pay minimal attention to programs with long term benefit. The executives will be at another company by then, so why bother thinking long term?
13. Rule cascades. Admin knows workers try to game the system. So they make a rule to fix it. And workers find another way to game it. So there's another rule. And another. Now you need a team of lawyers to ensure compliance, at huge cost.
14. Discouraging risk-taking. Workers know they could try a new technique, which might benefit the organization. But it might not work, and then they'd be fired. So they stick with the status quo, and the organization loses out on a million dollar idea.
15. Rewarding luck. If luck plays a role in outputs, you might be rewarding lucky people rather than hard workers. In academic publishing, there's a fair amount of luck in when papers get published, and where. In the long run, skill prevails, probably, but for students? Some luck
16. Discouraging cooperation. If I'm rewarded for being the most productive worker, it's in my self-interest to sabotage others in the organization. That's not good for creating a cooperative culture.
17. Overall, Muller makes the case that reliance on metrics cause a lot of problems, and metrics tend to proliferate, because the solution provided is often another metric.
18. You can measure the wrong thing, or measure the right thing in the wrong way. And measuring itself has costs that quickly balloon.
19. The book's author is on Twitter. Great book. @jerryzmuller
