The Tyranny of Metrics, by Jerry Z. Muller
I finished reading The Tyranny of Metrics by Jerry Z. Muller about two years ago, and its arguments have stayed with me ever since. If you have ever spent three hours filling out a ‘productivity tracker’ or some soul-crushing ‘self-evaluation software’ instead of actually doing the work you are being paid for, this book is for you. Muller gives a name to that nagging feeling we all have: that our modern obsession with counting things is actually making us less efficient and, frankly, a bit more miserable.
The issue isn’t that numbers are inherently evil. In fact, people love them because they provide a sense of orientation; they help us feel like we know where we are in a complex world. The problem is that we often forget that numbers are usually just bad proxies for the things we actually care about. We mistake the metric for reality, when in truth, it is just a ‘croquis’—a rough sketch that leaves out all the vital details.
This obsession with proxies tends to kill off professional judgement and what Muller calls ‘tacit knowledge’—that deep, messy expertise you gain from years of actually doing a job. Consider an official working in international development who has to decide which NGO to fund. A spreadsheet might show one organisation has ‘higher output metrics’, but the official’s intuition—developed over decades—might suggest the other group has a much better rapport with the local community and is a safer investment. That intuition is a sophisticated form of expertise, yet the system often forces them to ignore their gut and tick the box for the one with the better numbers. We treat humans like vending machines where you input a KPI and expect a perfect result, ignoring the fact that humans are far more complicated.
This leads us straight into the trap of Goodhart’s Law, which states: ‘When a measure becomes a target, it ceases to be a good measure.’ Once a number is tied to a reward or a threat, people stop focusing on the purpose of their work and start focusing on the number itself. To see this in the wild, you only have to look at the buses in the city where I live.
The system sets punctuality targets where drivers are judged on reaching certain checkpoints at a specific time. It sounds logical on paper, but in reality, once that arrival time becomes a metric the driver is punished for missing, everything goes sideways. You will see buses flying past crowds of waiting passengers, leaving people stranded on the pavement, because the driver is terrified that stopping to actually let someone on will make them ‘late’ for the checkpoint. The metric says the bus is ‘on time’, but for the passengers who were never picked up, the service is a total failure. You get exactly what you measured—a bus hitting a mark—but you lost the point of the service entirely.
Muller isn’t suggesting we should bin the spreadsheets and go back to the Middle Ages. He is just saying we need to stop treating metrics as if they are infallible. They are like salt: a little bit makes the stew better, but if you dump the whole pot in, it becomes inedible. We need to leave room for trust, common sense, and the realisation that not everything that matters can be measured, and not everything that can be measured matters.