`Cumulative Rate (Stage N) = Step Rate (N to N+1) x Step Rate (N+1 to N+2) x ... x Step Rate (final to won)`
The compounding is the finding
A worked example with three gates:
| Stage | Step rate to next | Cumulative rate to won |
|---|---|---|
| Discovery | 60% | 21% |
| Demo | 50% | 35% |
| Proposal | 70% | 70% |
Use it in place of the static probability field
A weighted pipeline is only as good as the percentages applied to it. Most CRMs ship with probability values assigned by stage at implementation, and those values survive for years without anyone testing them against outcomes. Swapping in a measured cumulative ladder changes the weighted number immediately, usually downward in early stages.
That single change is where a lot of forecast accuracy improvement comes from, before any model is involved. See weighted pipeline for how the weights are applied once you have them.
Two conditions the math depends on
Stages have to be entered in order, and entry has to be recorded. Stage skipping starves the skipped step of entrants and inflates the steps that follow, so the ladder describes a path no deal took. The fix is opportunity field history rather than current stage, since history counts entries even when a deal moved twice in a day.
The cohort also has to be mature. A step measured on deals created last month is measuring speed, not conversion, because slow deals have not had their outcome yet. Hold the window open long enough that the majority of that cohort has resolved, then recompute. The compounded rate feeds directly into win rate analysis, and the two numbers should reconcile at the final stage.
Frequently Asked Questions
How do you calculate cumulative stage conversion rate?
Multiply the step rates from the stage in question through to closed won. If discovery to demo runs at 60%, demo to proposal at 50%, and proposal to won at 70%, then a deal sitting in discovery converts at 0.60 x 0.50 x 0.70, which is 21%. The same multiplication run from each stage gives you a full ladder of stage level odds.
Is this the same as the probability field in the CRM?
No. Default probability values are set once during implementation and almost never revisited, so they encode an assumption rather than a measurement. Cumulative conversion is computed from resolved history and moves when the pipeline moves. Replacing the static field with the measured ladder is one of the cheapest forecast corrections available.
Should open deals be in the calculation?
No. Measure each step on deals that entered the stage far enough back to have resolved, then treat anything still open as excluded rather than lost. Including open deals in the denominator drags every step rate down and makes the compounded number look worse than performance was.
What breaks the multiplication?
Stage skipping and reopened deals. If reps jump a deal from discovery straight to proposal, the skipped step shows an artificially small entry count and the compounded number stops describing any real path. Reopened opportunities double count, because the same deal enters a stage twice and can produce more exits than entries.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like cumulative stage conversion rate into prescriptive action for your team.
Schedule a Demo