Bikeshedding
C. Northcote Parkinson observed that a committee reviewing plans for a nuclear reactor will spend little time on the reactor itself — too complex for most members to evaluate — and enormous time on the design of the bike shed: simple enough for everyone to have an opinion. This is bikeshedding, also called Parkinson’s law of triviality: the time spent on any item is inversely proportional to its importance.
The mechanism is asymmetric competence. The reactor requires specialized knowledge; most committee members must defer. The bike shed requires no expertise; everyone can contribute. Contributing feels like participation, so people gravitate toward decisions where their contribution is possible. Trivial items attract opinions; important items repel them.
The perverse result: the most consequential decisions get the least scrutiny, while the least consequential decisions attract endless debate. The nuclear reactor is approved in minutes; the bike shed color is debated for hours. Resources are allocated by who can participate, not by what matters.
This extends beyond committees. Online discussions spiral over formatting, not substance. Product teams debate logos while shipping broken code. Voters focus on personality while ignoring policy. Engagement concentrates where competence is easy, not where impact is high.
Partial remedies: time-box trivial discussions, require expertise for major decisions, explicitly weight agenda items by impact. But the underlying problem — that contribution is easier on easy matters — doesn’t disappear. People want to participate; participation on hard topics requires preparation that easy topics don’t.
The awareness itself helps. Noticing when you’re bikeshedding lets you ask: is this discussion worth the time? Would my energy be better spent understanding the hard decision? Sometimes the honest answer is no — the bike shed color is still undecided. But at least you’re choosing the trivial consciously.
Related: sayre law, mcnamara fallacy, streetlight effect, coordination, systems