Attribution · Published 2026-06-10 · 1 min read
Adapty has updated its prediction model
Adapty's note: «We recently updated our prediction model, so you may notice slightly different revenue and LTV figures. The new model is more accurate and less likely to overestimate revenue.»
What this is
Predicted LTV estimates the lifetime value — roughly, how much a user will pay — for a cohort of new users, say the last week's.
How it used to be
I never used the model before, because it overstated my predictions by around 30%. With this update it lands where mine do. I can cautiously recommend using it.
How I estimate it myself
My model is very simple: I look at the LTV curve. If, with made-up numbers, the LTV increment was 10% in January, 8% in February and 4% in March, do not expect 10% again in April. More likely you get 2% in April and 1% in May.
That is enough to judge when the curve reaches a plateau, and to estimate LTV from it.