Week 2: Rewarding A, hoping for B

Kerr, Goodhart, Campbell, and the model underneath all three. What happens to a measure once the people it measures know it is a target, and to the work the measure cannot count.

2 March 2027

Why it worksPay for what you can count and people do the countable thing; the uncountable work is what the measure was for.

In 1975 Steven Kerr published a paper whose title is the whole argument: “On the folly of rewarding A, while hoping for B”. Organisations reward what they can see and count, hope for what they actually want, and are then surprised to get the first. Universities want teaching and count publications, so they get publications. The same year Charles Goodhart wrote that any observed statistical regularity will tend to collapse once pressure is placed on it for control purposes. Donald Campbell wrote in 1979 that the more a quantitative indicator is used for decision-making, the more it will be corrupted and the more it will corrupt the process it was meant to monitor. The version you have heard, “when a measure becomes a target, it ceases to be a good measure”, is Marilyn Strathern’s 1997 paraphrase, and it is the one that fits on a slide.

Software teams rediscover this every quarter. Pay by lines of code and the code gets longer. Count closed tickets and tickets get smaller. Set a velocity target and story points inflate. None of this requires anyone to be dishonest. It requires only that the number be watched and that moving it be cheaper than doing the work.

Holmstrom and Milgrom gave the folly a model in 1991. Give someone two tasks, one you can measure and one you cannot, pay for the measured one, and effort moves to it from the other. That is not a flaw in the person; it is the contract working. The model’s least welcome result is that when the task you care about cannot be measured, the right contract often pays for nothing at all. No company running a curve wants to hear that, and week 12 will ask you to put a price on it.

Outline#

• Kerr’s folly, and the three phrasings of Goodhart that followed it
• the multitask model: why the uncounted work is the first to go
• lines, tickets, points: three metrics and how each one died
• why the dashboard cannot tell the difference
• what a measure can still do once everybody knows it is a target

Sources

• Kerr, On the folly of rewarding A, while hoping for B (Academy of Management Journal, 1975)
• Holmstrom and Milgrom, Multitask principal-agent analyses (Journal of Law, Economics, and Organization, 1991)
• Goodhart's law
• Campbell's law
• Goodhart's law in engineering metrics