Interactive demo

Cohort retention & LTV

The instrument panel a consumer or games business actually runs on. Six levers decide whether a player is worth what they cost to acquire — and the uncomfortable arithmetic is that a product can look healthy on every headline metric while every install quietly loses money. Move the sliders and watch where it breaks.

This is a model, not a simulation. Nothing is sampled and nothing is random: every number follows deterministically from the levers, through a power-law retention curve fitted to hit the D1 and D30 you set. That is the honest way to build a planning tool — it answers "what would follow if retention looked like this", which is the question the levers are actually asking.

Levers

Start from

D1 retention 45% D30 retention 14% ARPDAU $0.12 Cost to acquire $2.20 Installs / day 1,500 Cohort quality drift -1% / cohort

D7 retention

—

Avg lifetime

—

LTV @ 180d

—

Payback

—

LTV : CAC

—

Steady-state DAU

—

Retention curve

The area under this curve is the average number of days a player stays active — which is the only thing that turns ARPDAU into LTV.

LTV vs acquisition cost

Where the curve crosses the acquisition line is payback day. Past 180 days, most of the curve is flat — late revenue rarely rescues a bad cohort.

DAU at steady state

Hold installs constant and DAU converges to installs × average lifetime. Growth stalls at that ceiling no matter how much you spend.

Cohort triangle

Each row is a weekly cohort, each column its age. Cohort quality drift is the thing a single headline retention number hides: scale acquisition hard enough and later cohorts come in worse, so the business degrades while the average looks stable. Set the drift negative and read down a column.

Percentage of each cohort still active. Read across for one cohort's decay; read down for whether the cohorts you are buying now are worse than the ones you bought before.