Marketplace Unit Economics: Model Contribution and Payback

Founders often assume that scaling volume will automatically absorb acquisition costs. If your unit economics and fee structures remain constant while your transaction contribution is negative, more volume simply multiplies your losses and drains your capital faster.

A reliable unit economics model traces what the platform actually earns from a single transaction, then measures that yield against the cost of acquiring a defined group of buyers or suppliers. When Directorism designs and builds custom marketplaces and directories for clients, we structure the underlying database around the precise transaction, refund, and payout records required to answer these questions factually.

Defining the Unit and Observation Window

Before you calculate anything, establish boundaries. Choose one currency, one reporting period, and a strict definition of a completed transaction. A successfully delivered service, a paid booking, and a refunded order carry entirely different financial consequences. Document how fee credits are assigned to specific periods, and hold those definitions constant across scenarios. If you are unsure what event drives your revenue, start with the business-model selection guide.

Gather your input records first:

Start with a raw transaction export mapped to customer IDs and dates. Match fees, credits, and direct costs back to those individual records. Pull acquisition spending from your marketing and sales logs, clearly documenting what is included. If a record is missing, leave the input blank rather than assuming it is zero.

GMV is not revenue. As a16z explains, the total sales volume passing through a marketplace is distinct from the fees the platform actually keeps. Neither metric equals profit. In this model, contribution represents your platform revenue after fee credits and direct costs, but before acquisition spend and fixed overhead. Stripe's contribution explanation defines contribution as the selling price less variable cost; for a commission-driven marketplace, the platform's effective selling price is its fee, not the gross buyer payment.

Transaction Economics Example

Calculate contribution from your platform revenue and direct costs first. Leave acquisition and fixed overhead out of this step.

Consider a hypothetical September 2026 example in USD, assuming zero subscription or external revenue:

This hypothetical platform nets a contribution of $31.78 per transaction. The $822 represents combined direct costs for the period; it is not a raw percentage quote from a payment gateway.

Notice how the denominator dictates the narrative. That $3,178 represents 79.45% of platform revenue ($3,178 / $4,000), but only 15.89% of GMV ($3,178 / $20,000). Always label the base when discussing margins. Neither percentage proves the business is profitable overall.

Defining the Acquisition Cohort

With per-transaction contribution established, evaluate the cost to acquire the users driving those transactions. This requires isolating a cohort: a defined group of users who made their first payment within a specific time window.

Assume a separate hypothetical cohort of 50 buyers who first paid in June 2026. If the acquisition spend definitively assigned to that group is $1,500, the buyer CAC is $30 ($1,500 spend / 50 buyers).

Acquisition costs must be fully accounted for, but counted only once. State clearly whether you are measuring the buyer side or the supplier side, and document your allocation rules.

Observing Contribution Without Survivor Bias

Observed cohort performance measures the actual contribution a specific group generates over a fixed historical window.

To prevent survivor bias from inflating your numbers, you must divide the total cumulative contribution by the original number of acquired customers—including those who bought once and never returned.

Suppose the June cohort generated $750 in total contribution across each buyer's first 90 days following their initial payment. Because every buyer has passed the 90-day mark, the observation is complete. Divide the $750 by the 50 original buyers: $15 of observed 90-day contribution per acquired buyer.

This hypothetical $750 is tied exclusively to the June cohort. It is entirely separate from the $3,178 total platform contribution recorded in September. Furthermore, a 90-day observation is not a lifetime value. If a newer cohort has only existed for 30 days, their performance is incomplete and cannot be directly compared to this fully observed 90-day baseline.

Separating Observations from Assumptions

Founders frequently blur historical facts with future hopes. A contribution-based lifetime value estimates the total amount remaining after direct costs over an assumed customer lifespan. A revenue-based lifetime value ignores those direct costs entirely. Specify which calculation you are using, and state the evidence supporting your assumed horizon.

Moving to assumed future performance, imagine a scenario projecting a $5 monthly contribution per original acquired buyer over a 12-month horizon:

• Assumed Contribution Lifetime Value: $60 ($5 × 12 months)
• Simple Payback Period: 6 months ($30 CAC / $5 monthly contribution)

If you want to stress-test that assumption, run a more conservative model projecting only $3 in monthly contribution:

• Assumed Contribution Lifetime Value: $36 ($3 × 12 months)
• Simple Payback Period: 10 months ($30 CAC / $3 monthly contribution)

The hypothetical 90-day observation above does not guarantee this assumed 12-month retention. Simple payback assumes a constant stream of positive monthly contribution; when contribution varies, track cumulative cohort contribution over time to see whether and when it recovers acquisition cost. With no positive contribution, this simple payback calculation has no meaningful positive recovery period, and the workbook leaves it blank.

Reconciling Shared Transactions

Marketplaces feature two-sided acquisition. Calculating buyer CAC and supplier CAC as independent metrics is highly useful for diagnosing channel performance. However, because buyers and suppliers share the same transaction, you must reconcile the platform's overall economics before setting budgets.

If a single transaction generates $31.78 in contribution, you cannot spend that same $31.78 twice to justify both the buyer's and the supplier's CAC. Count each transaction's contribution once, and evaluate the combined acquisition cost against it.

Modeled payback is not a cash-flow forecast. Simple payback arithmetic ignores cash timing. Delayed settlement, refunds, and the timing of service costs change when cash is actually available. Buyer funds still owed to suppliers are not spendable platform contribution. Fixed overhead also has to be covered. If a modeled contribution change comes from fewer completed transactions, use the liquidity metrics guide to investigate the underlying transaction outcomes.

The Decision and Review Record

Models are only useful if they can be understood and updated by someone else. Keep a plain-text record alongside your spreadsheet:

If direct support costs are undocumented, assign a specific team member to measure support hours against completed transactions before you make any claims about margin. The next step follows the weakest input, rather than a universal CAC or margin target.

Moving to the Excel Workbook

Download the Excel unit economics workbook to model your own figures. The file requires one reporting period to calculate platform revenue and transaction contribution; you can leave the acquisition fields blank if you only need those figures. To calculate cohort CAC, assumed lifetime contribution, and simple constant-contribution payback, you must also input a defined acquisition cohort.

Keep in mind that the workbook relies on the evidence and assumptions you provide; it does not automatically measure retention or forecast future behavior. Reference the calculator walkthrough for a detailed cell-by-cell guide. Save your documented assumptions alongside the file, and secure your baseline measurements before allocating capital to acquisition.