Definition

The freefall portion of a signup cohort's drop-off curve — the period of steep early losses during which users haven't yet established whether the offering will satisfy their intent.

Two regimes in the drop-off curve

A signup cohort's survival curve has two structurally different regimes. The first is starting churn: a steep drop-off in the first days or weeks after signup, where losses are concentrated. The second is sustaining churn: the flatter portion that follows, where the cohort has filtered down to a more stable population and losses run at a much lower rate.

The two regimes reflect a structural difference in what's happening to users at each point in the cohort's life. During starting churn, users are still assembling evidence about whether the offering will deliver on their intent, and small failures weigh heavily because the evidence base is small. By the time the cohort has crossed into sustaining churn, the users who remain have accumulated enough confirming evidence — and enough switching cost — that day-to-day issues are absorbed as noise.

Blended metrics like net churn collapse the two regimes into a single number, which hides that most actionable loss is concentrated in starting churn.

Why early losses compound

Four mechanisms compound to make starting churn structurally distinct from sustaining churn:

• Evidence-weight asymmetry: a single failed interaction in the first session represents a significant fraction of the user's total evidence about whether the offering will satisfy their intent. The same failure in the hundredth session runs against hundreds of confirming sessions and is absorbed as noise.
• Volume: most users are still early; most events happen there. The cohort drop-off concentrates loss in this period by definition.
• Lowest cost of switching: early users haven't accumulated investment — no learned habits, no stored data, no integrations. The cost to leave is at its lowest, so trust failures that tenured users absorb actually cause early users to leave.
• Unsupported intents concentrate here: first-encounter, orientation, fit-evaluation, and getting-to-first-value intents cluster in this window. These are often the intents object-shaped teams are most blind to, because they're not the "main use case" the offering is built around.

Why value milestones work hardest here

Starting churn is where value milestones do the most work. Users in this window are still updating their prediction about whether continued engagement will pay off, so interventions that produce intent-satisfaction events directly affect the evidence base they're weighing. Once a cohort has crossed into sustaining churn, the same interventions produce smaller marginal updates because users have already accumulated their evidence base.

The apparatus generalizes in principle — multiplier analysis, milestone work, and intent-anchored path design can be applied to sustaining-churn regimes — but the operational focus sits in the freefall window because the leverage is highest and the mechanisms are most active there.

Example

A freemium SaaS signup cohort plotted as a survival curve shows the canonical shape: a steep early descent — the majority of the cohort lost within the first few weeks — followed by a much shallower slope that continues for months or years. The early descent is starting churn; the long tail is sustaining churn. A team reporting "monthly churn" as a single averaged percentage across the cohort's lifetime is publishing a blended number that obscures which regime is driving loss. The same data shown as a cohort-anchored survival curve makes the regime split visible immediately, and makes clear that interventions targeting the long tail are operating on a much smaller population than interventions targeting the early descent.

See also

• First paid conversion
• Missed conversion
• Profitable customer
• Behavioral proxy
• CAC payback
• Concentrated leverage