How the average sales cycle calculator works
Average sales cycle measures elapsed time between a chosen start event and a won deal. A start event might be opportunity creation or qualification, but the definition must stay consistent. This calculator accepts the elapsed day count for each completed win and returns the mean, median and 90th percentile.
The mean includes every day in the sample and is sensitive to unusually slow deals. The median describes the middle observation and can better represent a typical deal. The 90th percentile uses nearest rank: sort the sample and take the observation at the rounded-up 90% position.
Average cycle = sum of closed-won cycle days ÷ won deals; median = middle sorted duration; P90 = nearest-rank 90th percentile
Worked example
Start with the example inputs below. The calculator opens with these values; reset the example at any time to compare your own scenario.
| Measure | Value |
|---|---|
| Closed-won cycle lengths (days) | 20, 25, 30, 35, 90 |
| Measure | Value |
|---|---|
| Average | 40 days |
| Median | 30 days |
| 90th percentile (nearest rank) | 90 days |
| Measurements | 5 |
Use elapsed calendar days from one consistent start event to closed-won. Open and lost deals are excluded from this sample.
How to use the result
Export elapsed durations for one segment and a consistent closed-won reporting period. Paste only numbers, separated by commas, spaces or new lines. Compute elapsed calendar days from your timestamps before entry; this tool does not infer working-day calendars or time-zone rules.
Compare the mean with the median before choosing a planning assumption. A much higher mean suggests a long tail of slow deals. Review those deals for procurement delays, stalled decisions or inconsistent start dates rather than assuming every opportunity needs the same time.
Assumptions and common mistakes
Completed wins exclude opportunities still open and deals that were lost. That can make a recent report look faster when slow opportunities have not yet finished. A small sample makes percentiles unstable. The maximum is 500 measurements and each duration must be non-negative.