Signup cohort
A signup cohort is a group of users who signed up during a defined window, typically a calendar week or month, tracked together as a fixed group from that point forward. The cohort exists as a unit so the team can ask, of everyone who came in the door together, what happened to them downstream.
Definition
The set of users who signed up within a defined window, before conversion or retention outcomes have resolved — the population over which multiplier analysis operates.
Inclusion is fixed at signup
Cohort inclusion is determined by who signed up in a defined window — say, all users who created accounts between March 1 and March 31. That population doesn't change as outcomes resolve. Some cohort members will convert; some will churn; some will remain free indefinitely. All of them remain in the March signup cohort regardless of what happens afterward.
This is structurally different from a conversion cohort (those who became paying customers in a window) or a customer cohort (those who were paying customers in a window). Both are downstream populations: the conversion cohort filters out users who never paid; the customer cohort additionally filters out churned customers. The signup cohort sits upstream of both, before any downstream filter has run.
Why signup is the right boundary
Multiplier analysis compares the conversion rate of a candidate-behavior subset against the full cohort baseline. The comparison only carries signal if the cohort contains both converters and missed conversions at meaningful representation. The signup cohort does; downstream cohorts don't.
A second reason concerns regime. Starting churn and sustaining churn are structurally different drop-off regimes, and the patterns distinguishing profitable customers from missed conversions concentrate in the starting-churn window. The signup cohort lives entirely in that window. A baseline computed on a longer-duration population averages across both regimes; a baseline computed on customers averages over survivors of starting churn whose profitability-relevant behaviors have already been selected for. Either move dissolves the signal multiplier analysis is trying to surface.
Self-serve scope
The signup cohort assumes signup is the user-driven entry into value-pursuit — the user encountered the offering, decided to engage, and created an account as part of that decision. In sales-driven contexts this isn't true: accounts get created through partner cross-promotion, post-contract provisioning, or sales-rep handoffs, with no consistent relationship to user-driven conversion progression. A signup cohort assembled from such events isn't a behaviorally coherent population, and multiplier analysis run on it produces noise rather than signal.
Example
A self-serve SaaS company defines its March signup cohort as the 10,000 users who created accounts between March 1 and March 31. By July, 1,200 of those have converted to a paid plan, and 800 of those are still paying.
To run multiplier analysis on a candidate behavior — say, "completed a project import in week 1" — the baseline conversion rate is computed against all 10,000 (e.g., "12% of March signups converted by month 3"). The subset rate is computed against the slice of those 10,000 who did the behavior. The multiplier compares the two.
The 1,200 converters and 800 sustained customers can't substitute for the 10,000 here. Both populations exclude the missed conversions that the contrast depends on, and both have been pre-filtered by the very behavior the analysis is trying to investigate.