Other Mental Models · OM-11
You can arrive at a surprisingly accurate estimate of an apparently unknowable quantity — how many piano tuners are in Chicago, how many gas stations are in a country — by breaking it into a chain of smaller, individually estimable quantities and multiplying them together.
A technique for producing a rough, order-of-magnitude estimate of a quantity that seems impossible to know directly, by decomposing it into a chain of smaller component quantities that can each be estimated with reasonable confidence, then combining those component estimates — the individual estimation errors tend to partially cancel out, producing an overall estimate that is far more accurate than the apparent difficulty of the original question would suggest.
Named for physicist Enrico Fermi, who was known for posing and solving such estimation problems, including his famous in-class example of estimating the number of piano tuners in Chicago, and for his own real-time estimate of the yield of the first atomic bomb test using scraps of paper dropped during the blast.
The Mechanism
Break the unknowable into a chain of knowables, then multiply
Fermi's own most famous demonstration came at the Trinity nuclear test in 1945, when he dropped small scraps of paper as the blast wave passed and used the distance they were carried to estimate the bomb's explosive yield in real time — an estimate that later proved remarkably close to the value obtained from full instrumental measurement, illustrating how a simple, well-chosen chain of physical reasoning can substitute for direct measurement.
01 · THE TECHNIQUE WORKS BECAUSE INDEPENDENT ESTIMATION ERRORS TEND TO PARTIALLY CANCEL, NOT COMPOUND
This is the statistical reason the method is more accurate than it seems it should be
If each component estimate in the chain has roughly independent errors (some too high, some too low), the errors across the whole multiplicative chain tend to partially offset one another rather than compounding in the same direction — this is why a chain of five rough guesses, each individually uncertain by a factor of two, often still produces a combined estimate within a much smaller factor of the true value.
02 · THE GOAL IS THE RIGHT ORDER OF MAGNITUDE, NOT PRECISION
Fermi estimation deliberately trades precision for speed and tractability
A Fermi estimate is not attempting to determine an exact figure — it's attempting to establish whether the true answer is closer to 10, 100, 1,000, or 10,000, a level of precision that is often sufficient for a real decision (should we enter this market, is this claim plausible) and can be produced in minutes rather than requiring extensive research or data collection.
03 · IT'S WIDELY USED AS A SCREENING TOOL BEFORE COMMITTING TO A MORE RIGOROUS ANALYSIS
Business, engineering, and scientific practice all use it as a fast first-pass sanity check
Before commissioning detailed market research or a full engineering analysis, a quick Fermi estimate is frequently used to sanity-check whether a proposed number is even plausible — if a detailed later analysis produces a result many orders of magnitude away from the Fermi estimate, that mismatch is itself a valuable signal that something in the more detailed analysis needs re-examination.
Where It Fails / Inversion
Where it fails / inversion
Fermi estimation performs poorly when the underlying quantity is genuinely governed by a small number of highly correlated, non-independent factors, rather than a chain of roughly independent ones — in those cases, errors in the component estimates can compound in the same direction rather than canceling out, producing a final estimate that's badly wrong despite each individual step seeming reasonable.
How To Use It
Worked example · sanity-checking a business's total addressable market
Before committing significant resources based on a market-research firm's claimed total addressable market figure, a quick Fermi estimate — population size, times the percentage plausibly needing the product, times a reasonable price point — provides an independent, fast sanity check; if the Fermi estimate and the market-research figure differ by orders of magnitude, that discrepancy deserves investigation before either number is trusted for a major resource-allocation decision.
How to use it
When facing a quantity that seems impossible to know directly, try breaking it into a chain of smaller quantities you can estimate with reasonable confidence, then multiply them together — the goal is the right order of magnitude, achieved quickly, not false precision achieved slowly.
See Also