Your real sales cycle — and why the average lies
Post 28 Aug 2026 MeasurementForecasting
Ask any sales manager about their sales cycle and you get one number: "about two months". Then look at the data and find deals that closed in a week and others in seven months. The single number was not just wrong, it was misleading.
Use the median, not the average
One deal that took a year lifts the average by a full month. The median — the value half your deals fall below and half above, is unmoved by it.
| Measure | What it tells you |
|---|---|
| Median | Your normal case |
| Upper quartile | When a deal becomes "stuck" |
| Gap between them | How random your process is |
The practical rule: a deal past the upper quartile is not slow, it needs a decision, a push or a close.
Segment before you measure
One cycle for everything is a meaningless number. Split at least by:
- Deal size, small and large are not the same process.
- Customer type, private sector versus public tender (the second runs on a calendar you do not own).
- Source, a referral always closes faster than an ad.
Forecasting: weighted probability, not hope
Multiply each deal's value by the historical close rate of its stage, not the rep's estimate. If deals in "Negotiation" historically close at 45%, a 200,000 deal is worth 90,000 in the forecast.
Compare forecast to actual every month. A persistent gap in one direction is not bad luck, it is a wrong stage rate.
What to watch weekly
Three numbers are enough: deals past the upper quartile · new value entered · last quarter's close rate. Anything beyond that is a report read once and forgotten.