The four steps
| Step | What you do |
|---|---|
| 1 | Pick the cohort key, usually created month or created quarter |
| 2 | Freeze the membership list, since a deal belongs to the cohort it was created in even if it is reopened later |
| 3 | Wait out the maturity window before reading the result |
| 4 | Report won, lost, and still open as three shares of the same denominator |
Deciding when a cohort is readable
Every cohort is provisional until enough of it has resolved. ORM models this directly, predicting a close curve for each group of opportunities, with curves running from 1 to 80 weeks and most of the expectation before week 12. Very few groups carry expectation past 52 weeks.
That shape gives you a practical rule. Read a cohort at 12 weeks as an early indicator, treat it as reliable once the open share falls into the low teens, and consider it closed out at 12 months. Reading a three week old cohort tells you which deals move fast rather than which deals win.
What the open column is worth
The unresolved share is a leading indicator on its own. A cohort that is aging with a large open share and no stage movement is not undecided, it is stalled, and those deals mostly resolve as no decision losses. Compare the open share across consecutive cohorts and the trend appears before it reaches bookings.
Cohort win rates also give a forecast a defensible conversion assumption, since they are measured on the same kind of grouping a forecast has to make. See sales forecasting for where that assumption belongs, and win rate for the parent metric and its other cuts.
Frequently Asked Questions
What is the difference between cohort win rate and period win rate?
Period win rate counts deals that closed inside a window no matter when they started, so a quarter's number blends fast transactional deals with enterprise deals created a year earlier. Cohort win rate counts deals created inside a window and follows them to resolution. Only the cohort version connects an outcome back to the sourcing and qualification decisions that produced it.
How long do you wait before reading a cohort?
Long enough for the majority of that group to resolve. ORM groups opportunities with a machine learning model and predicts a close curve for each group, and those curves run from 1 to 80 weeks with most of the expectation landing before week 12. Very few groups carry expectation past 52 weeks, so a cohort older than a year is close to final for most B2B SaaS motions.
What do you do with deals from the cohort that are still open?
Report them as a third column rather than dropping them or counting them as losses. A cohort at 22% won, 55% lost, and 23% still open is telling you something different than a cohort at 22% won and 78% lost. Dropping the open deals inflates both the win rate and the loss rate on young cohorts.
What does a cohort view show that a period view cannot?
When a change actually happened. If you tightened qualification in March, the effect appears in the March creation cohort and stays invisible in period reporting for months, because the deals created under the old rules keep closing. Cohorts date the change to the moment it was made rather than to the moment its results arrived.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like how do you calculate win rate by cohort? into prescriptive action for your team.
Schedule a Demo