There is a story about a company that set a target for ending customer service calls quickly, after which staff began hanging up before they had finished solving customers' problems. Call times went down, but the number of the same customers calling back actually went up. The number improved, but the work itself got worse.
The phenomenon in which a number that measures something loses its original meaning the moment it is made a target is called ‘Goodhart's law.’ It is commonly known through a single sentence: “When a measure becomes a target, it ceases to be a good measure.”
In this article, we will look in turn at where Goodhart's law came from, why it happens, what examples can be seen around us, and how to handle numbers wisely.
Understanding Goodhart's Law
Where Did Goodhart's Law Come From?
The law is named after the British economist Charles Goodhart. He worked on monetary policy at the Bank of England, the UK's central bank, and first set out the idea in a paper presented in 1975 at a conference hosted by the Reserve Bank of Australia, Australia's central bank.
At the time, the British government was trying to steer the economy using the amount of money circulating as an indicator. Goodhart pointed out that “any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.” He meant that once the government latches onto a particular indicator and starts controlling it, people and financial institutions change their behavior accordingly, so the indicator no longer shows the real state of the economy.
The short sentence widely used today was formulated by the British anthropologist Marilyn Strathern in a 1997 article on the assessment system of British universities. Thanks to this, Goodhart's law spread beyond economics to almost every field that measures performance with numbers, including education, public administration, and management.
Why Do Numbers Break Down When They Become Targets?
An indicator is originally a window that shows us, indirectly, something that is hard to see directly. Test scores give us a sense of ability, sales figures of customer satisfaction, and call times of the efficiency of customer service. But each pair is not exactly the same thing; they only tend to move together.
When rewards and penalties are attached to a number, people look for the easiest way to push up the number itself rather than serve the original purpose. They memorize only the questions likely to be on the test instead of building their ability, or hurry to end calls instead of helping customers. Then the number and reality diverge, and the number increasingly tells lies.
The American social psychologist Donald Campbell pointed out the same problem around the same time. He held that the more a quantitative indicator is used in social decision-making, the more it is subject to pressures of corruption and distortion, and this idea is called ‘Campbell's law.’
There is one more trap. Those who manage the numbers easily come to believe that things are going well when they see the indicator improving. So even as the number drifts away from reality, they notice the problem late and make their next decisions based on an indicator that is already broken.
Goodhart's Law Around Us
A widely told anecdote is the ‘cobra effect.’ The story goes that when Britain ruled India, a bounty was paid for every cobra caught in Delhi to reduce their number, and people started breeding cobras outright. When the government stopped the reward, the now-useless snakes are said to have been released, making the problem worse. Whether it really happened is unclear, but it is often used as an example of the side effects that arise when rewards are attached to numbers.
In schools, when test scores become the yardstick for evaluating schools, teaching can drift toward test preparation. In hospitals, delaying the moment patients are registered in order to meet waiting-time targets has also been criticized. On the internet, it is common to put sensational headlines that don't match the content in order to boost view counts. In every case, the number improves while the original value it pointed to shrinks.
How to Use Numbers Wisely
Goodhart's law does not mean we should stop using numbers. Numbers are still an essential tool for understanding a situation. But we must be careful to treat them as a mirror reflecting the original purpose and not make the numbers themselves the goal.
One way to do this is to look at several complementary indicators together rather than relying on just one. If you look at call time along with the repeat-contact rate and customer satisfaction, it becomes hard to force up just one of them. It is also important to look directly at the real situation behind the numbers and to boldly change an indicator once it starts to stray from its original meaning.
Above all, if people share an understanding of why a number is being measured, the temptation to game it also decreases. Sharing the reasons behind a goal is, in effect, the best prevention. In fact, many organizations try to reduce such distortions by reviewing their indicators at set intervals and creating opportunities to hear voices from the front line that numbers do not capture.
Goodhart's law reminds everyone who deals with numbers to be humble. Numbers are only a window onto reality, not reality itself, and polishing the window does not change the view outside.
Never forgetting the original purpose beyond the numbers: that is the first step toward keeping numbers a trustworthy tool.
