18 · Deciding
Pareto Principle
A small share of causes produces most of the results. The usual shorthand is 80/20.
Evidence: Rule of thumbWhat we do with it
- 01
Check the numbers before you pick. The pages that carry most of the traffic are rarely the ones a team argues about.
- 02
Most complaints trace back to a small number of problems. Fix those before the long tail.
- 03
Don't use it as an excuse to ignore everyone else. The 20% you skip might be someone's entire job.
Examples
Getting it wrong, getting it right.
Don't
A redesign that gives every page the same budget.
Do
Analytics first. The three pages that carry most of the traffic get redesigned and tested before anything else.
The evidence
What the research says.
An observation about uneven distributions, and a useful one, but 80/20 is a shorthand. The real split might be 60/20 or 100/10, and you only know by measuring. Norman Fenton and Niclas Ohlsson measured it in 2000 across two releases of a large telecom system: 20% of the modules held nearly 60% of the faults found in testing, and 10% of the modules caused 100% of the failures reported from the field in one release and 80% in the next. Those modules were small, too. In the first release they made up 12% of the code.
Sources
- Pareto, V. (1896). Cours d'économie politique. Lausanne: F. Rouge.
- Persky, J. (1992). Retrospectives: Pareto's law. Journal of Economic Perspectives, 6(2).
- Fenton, N. E., & Ohlsson, N. (2000). Quantitative analysis of faults and failures in a complex software system. IEEE Transactions on Software Engineering, 26(8).
Last reviewed November 7, 2026. How we keep this current