Published
Methodology
Quantitative analysis of 1,400 performance indicators across 40 organisations, with 22 follow-up interviews
Co-Authors
Dr. Tomas Vidal
Nadia Karam
Paper Length
36 pages
Cite as
Ellery, J., Vidal, T., & Karam, N. (2025). Measuring the wrong thing: Indicator design and behavioural drift. Public Performance Review, 27(2), 165–200.
Copy citation
Measures do not simply describe performance; they redirect attention. This paper analyses 1,400 indicators in use across 40 organisations and shows how a small number of poorly specified measures shaped behaviour far beyond their intended scope.
What 1,400 indicators taught
Across forty organisations, most indicators were introduced to describe performance and ended up defining it. Within two years of adoption, teams reliably optimised the measure rather than the outcome it was chosen to represent.
“
A measure is never neutral. Once published, it becomes an instruction, and it will be obeyed more faithfully than the strategy it was meant to serve.
Accumulation without retirement
The median organisation added eleven indicators a year and retired one. Reporting effort grew steadily while the number of measures anyone could act on stayed close to five.
A test before adoption
Before a measure is published, ask what behaviour it would reward if pursued dishonestly. In twenty-two follow-up interviews, that single question predicted the failures we later documented more accurately than any technical review.
