Fintech Metrics: Unit Economics, Take Rate, and Risk Analysis

Fintech unit economics are unforgiving — thin margins, real default risk. Here are the metrics that matter and how to compute them from transaction data.

Thin margins, real risk

Fintech businesses run on basis points. A few bps of take rate and a few bps of losses decide whether the model works. You have to track revenue and risk on the same data.

The revenue metrics

• Take rate: net revenue ÷ total payment volume (TPV). The core fintech margin.
• TPV (Total Payment Volume): the top-of-funnel volume metric.
• Contribution margin per user: revenue minus processing, fraud, and servicing costs per active user.
• ARPU: average revenue per user, tracked by cohort.

The risk metrics

• Default / charge-off rate: defaulted balance ÷ total outstanding, by vintage.
• Delinquency buckets: % of balance 30/60/90 days past due.
• Loss-adjusted yield: gross yield minus expected losses. The real return.
• Cohort loss curves: how each origination vintage performs over time.

Why cohorts are everything

Blended default rates lie when you're growing fast — fresh loans haven't had time to go bad, so the average looks great while the underlying vintages deteriorate. You must analyze by origination cohort to see the truth.

How Talon does it

Upload your transaction or loan-tape CSV and ask the Fintech Analyst: "Compute take rate, contribution margin per user, and default rate by origination cohort, then chart the loss curves."

Talon runs the deterministic Metric Pack, builds cohort loss curves, and flags any vintage trending above target loss rates.

Bottom line

Track take rate and loss-adjusted yield by cohort, never blended. Talon computes both from a raw transaction export in seconds.