A four-step framework for CFOs to control AI token costs.

Read Now

Cohort modeling

View how to build a customer cohort model from a new-cohort line and a retention curve, extend it to usage-based revenue and platform fees, and use the results in reports.

How cohort modeling works in Drivetrain

Each part builds on the rows created before it, so set them up in the order shown. Usage-based revenue and platform fees are optional extensions.

Enter the new cohort line

  • Add a row for the number of new customers you expect each month.
  • Enter one value to fill every month.
  • Overwrite any month where you expect a different number.
Drivetrain Cohort Model with a New Cohorts (M0s) row and the forecast formula set to 300

Set up the retention curve

  • Keep your customer retention curve in its own model.
  • Express it as the share of customers still active in each month after they join.
  • Adjust any month's value directly.
Cohort Retention Curve model showing customer retention of 100%, 97%, 95%, 91%, 89%, 85% and 81% for months 0 to 6

Build the customer cohort table

  • Add a customer cohort row, with cohort as a dimension.
  • Multiply new customers by the retention curve to get a cohort for every start month.
  • Use time selectors to line each cohort up with its start month and restart the curve for every new cohort.
  • See each cohort's first month and how many customers remain from earlier cohorts, month by month.
Customer Cohort table expanded by cohort month, with the diagonal of new cohorts and monthly totals of 300, 591, 876 and 1,179

Extend the model to usage-based revenue

  • Add a price row.
  • Add a usage retention curve, which can rise above 100% as customers expand.
  • Multiply new customers by usage retention and price to get a revenue cohort.
  • See revenue grow even as the customer count falls.
Revenue Cohort table expanded by cohort month, showing January's revenue rising from 6,000 to 12,000 in its second month

Add a platform fee cohort

  • Add a platform fee as a percentage of revenue.
  • Multiply the revenue cohort by the fee to get a platform fee cohort.
  • Change the rate for any month to see the effect on fees.
Cohort Model with platform fees changed from 3% to 2% from July 2025 and the Platform Fees Cohort expanded by cohort month

Use the cohort model in reports

  • Show customers, revenue and fees from the cohort model as charts, tables or KPIs.
  • Explore trends and drill down by any dimension.
Cohort Analysis report with Customers and Customer Retention KPIs and Revenue Cohort and Customer Cohort charts

How Drivetrain simplifies cohort modeling

  • Forecast how revenue and usage grow after signup, not just customer counts.

  • Forecast churn with a retention curve that changes as cohorts age, not one flat rate.

  • Drill into cohort results by any dimension in reports.

FAQs

How do we build a revenue model from scratch based on cohort behavior?

Start with two inputs: the number of new customers each month and a retention curve. Multiply them to get customer cohorts, then multiply by usage and price to get revenue cohorts.

What about churn? How does Drivetrain forecast it?

Churn comes from the retention curve, which shows the share of each cohort still active in each month after it starts. You enter or edit the percentages yourself.

Our customers expand usage over time. Can the model show that?

Yes. A separate usage retention curve can go above 100%, so revenue from a cohort can grow even as its customer count falls.

Can we use the cohort results in reports?

Yes. Customers, revenue and fees from the cohort model can be shown in reports as charts, tables or KPIs.

See Drivetrain in action

Fewer spreadsheets, faster planning, better decisions.

Book a Demo