BNPL unit economics: how the margin is really made
Buy-Now-Pay-Later looks simple — the merchant pays a fee, the customer pays in instalments. But the margin per order is thin and four variables decide whether you make money or quietly bleed it. Here is how they combine.
LAST REVIEWED: JUNE 2026 · ILLUSTRATIVE BENCHMARKS, NOT A FORECAST
A BNPL provider earns most of its revenue from the merchant discount rate (MDR) — the fee the merchant pays for higher conversion and basket size. Against that sit three costs: defaults/losses, the cost of funds to finance the receivable, and operating cost per order. What's left is contribution margin. The trap is that each looks small on its own; together they can erase the spread.
The four levers
Lever
What it is
Typical range
MDR
Merchant fee as % of order value — your gross revenue
~3–6%
Loss rate
Share of financed value never recovered (default, fraud)
~2–8% of GMV
Cost of funds
Annual financing cost on the outstanding receivable
Rate × the weeks you're out of pocket
Opex / order
Payment processing, servicing, collections, support
Fixed + variable per order
Cost of funds is the one founders underweight. If you finance a 6-week instalment at an annual rate, you only carry the money for ~6 weeks — but at scale across a large book, even a small annualised rate is a real line item.
A worked example (per AED 1,000 order)
MDR revenue (4.0%)+40.00
Expected loss (3.0% of GMV)−30.00
Cost of funds (6 wks @ ~12%/yr)−1.40
Opex / order−6.00
Contribution / order+2.60
In this illustration a 4% MDR order nets ~AED 2.60 — a ~0.26% contribution margin on GMV. Nudge the loss rate from 3% to 5% and the order turns negative. That razor-thin sensitivity is the entire BNPL game: underwriting quality and MDR negotiation are the business, not a detail.
What actually moves the needle
Loss rate is king. A one-point change in losses can swing you from profit to loss. Underwriting, limits, and repeat-customer data matter more than headline volume.
MDR has a floor. Merchants resist above ~5–6%; competition pushes it down. You rarely solve thin margins by raising MDR.
Repeat customers carry the model. Acquisition cost is one-time; a customer's second and third orders are where contribution compounds. Cohort retention is the real lever.
Late fees are fragile revenue. Regulators across markets are tightening on consumer-credit fees; do not build the model on penalties.
Model your own BNPL business
Adjust MDR, loss rate, cost of funds and opex — see the contribution waterfall update live, benchmarked to MENA presets.
Contribution per order is typically a fraction of a percent of GMV once losses, funding and opex are netted off MDR. Profitable players win on low loss rates and high repeat usage, not on a fat per-order spread.
Why do BNPL companies lose money even with growth?
Because each order's margin is thin and front-loaded costs (acquisition, losses on new, unseasoned customers) hit before repeat-customer contribution compounds. Growth without underwriting discipline scales the losses, not the profit.
What loss rate kills a BNPL model?
It depends on MDR, but in a 4% MDR model a loss rate drifting from ~3% toward ~5% of GMV can flip the per-order economics negative. Loss rate is the most sensitive single variable.
Thinking about building a BNPL or lending product?
Pressure-test the whole idea — market, regulation, economics, competition — before you build.