Conversion cohort
A conversion cohort is the group of users who became paying customers during the same time window: everyone who first paid in February, for example. What unites them is when they converted, not when they signed up. A February conversion cohort would include people who signed up in December, January, and February, provided that they all happened to pay for the first time during that month.
Definition
A cohort grouped by users' first paid conversion month, distinct from the signup cohort which groups by signup month.
Anchoring at the conversion event
A user belongs to a conversion cohort based on when they first paid, regardless of when they signed up. The January 2024 conversion cohort contains users who signed up in January, December, November, or earlier — what unifies them is that they all became paying customers in January 2024.
Conventional SaaS finance tooling adopts this reading direction. ChartMogul, ProfitWell, MRR cohort views, retention curves past conversion: the customer is the unit of accounting, and the customer comes into existence at the moment of payment. Tracking starts there and runs forward through retention, expansion, and churn.
The cohort contains only converters
Because the anchor event is conversion, the cohort by construction excludes anyone who never converted. Signups who entered the funnel but didn't reach payment are never members. The signup→conversion loss — typically the largest drop-off in the user funnel — is invisible to conversion-cohort views.
For activation questions, this is the wrong unit. Activation work asks what determines whether a signup becomes a sustainably-paying customer; the population that didn't convert is half the question. Multiplier analysis and related apparatus operate on the signup cohort for exactly this reason — the denominator has to include the missed conversions the work is trying to recover.
For finance and revenue analysis past conversion, the restriction fits the question. Retention, LTV, and payback questions are about the population of paying customers; the conversion cohort contains exactly that population.
Where this framing fits
Conversion cohorts are the right unit for:
- Retention and churn analysis past the conversion event (sustaining churn regime)
- LTV and payback computation at the customer level
- MRR cohort views and revenue accounting
They are not the right unit for activation work — the signup cohort is. Calendar-duration substitution bridges between the two when full signup-level attribution isn't available, treating conversion-cohort data as approximately equivalent to signup-cohort data under specified assumptions: short signup-to-conversion lag, stable conversion mix, uniform acquisition pacing.
Example
A team reviews their February 2024 cohort in ChartMogul. The view shows 240 customers, their MRR trajectory, a 12-month retention curve, and projected LTV. The cohort answers questions like "are February customers retaining better than January customers?" and "what's our payback period for this cohort?"
What the view doesn't show: the ~4,800 signups during January and February who never paid. For activation questions — what drives signups to convert, where missed conversions concentrate by behavior — those 4,800 belong in the denominator, and the conversion cohort excludes them by construction. Activation work needs a parallel signup-cohort view alongside the finance view.