Data & Analytics

We Don't Scale Traffic. We Scale Predictable Economics.

Product events, subscription and payment records, attribution data and advertising spend are joined in one model. Every acquisition decision is taken against that model.

Attribution & Cohorts

Acquisition Data and Product Data in One Record

A user is not a click. Each cohort keeps its acquisition context for its whole life, so revenue can be attributed back to the exact creative, campaign, country and price it came from.

Acquisition date

Cohorts are cut by install or signup week.

Channel

Platform and network the user came from.

Campaign

Campaign and ad set structure.

Creative

The exact concept and variant that converted.

Geography

Country and language of the user.

Product

Application and platform.

Funnel

Onboarding and paywall variant seen.

Pricing

Price point, trial length and billing period.

Cohort Prediction

Early Signals, Long-Term Revenue

Behaviour in the first days of a cohort — activation, trial starts, engagement depth, first charges — is used to model revenue months ahead, and the forecast is corrected as the cohort matures.

Observed revenue vs modelled forecast SCHEMATIC

              D0D1D3D7 — MODEL INPUT CLOSESD30D90D180 PREDICTED
            

Observed

Realised revenue from the cohort, reconciled with processor and store records.

Forecast

Modelled continuation of the curve, with a confidence band that narrows as data arrives.

Decision point

Spend is adjusted once the forecast crosses or misses the payback threshold.

Unit Economics

The Metrics Every Decision Is Measured Against

CAC Customer acquisition cost Total spend divided by acquired subscribers, per channel and country.

CPA Cost per action Cost of the intermediate steps: install, signup, trial start.

TRIAL CVR Trial conversion Share of users who start a trial after reaching the paywall.

PAID CVR Paid conversion Share of trials that reach a first successful charge.

ARPU Average revenue per user Net revenue per user over a fixed observation window.

LTV Lifetime value Modelled revenue per user across the subscription lifetime.

ROAS Return on ad spend Realised and forecast revenue against spend, per cohort.

PAYBACK Payback period Days until cumulative cohort revenue covers acquisition cost.

RETENTION Retention Product retention and subscription renewal by period.

Decision Engine

Collect, Model, Predict, Allocate, Scale

Collect

Product events, payments, attribution and spend, in one pipeline.

Model

Cohorts built per channel, creative, country, funnel and price.

Predict

Forecast revenue and payback from early cohort signals.

Allocate

Budget moved to the segments that clear the threshold.

Scale

Volume increased while the model holds, and reviewed continuously.