Choosing Percentage, Fixed, or Qualified-Lead Commissions

The best commission model is the one that aligns Affiliate effort with verified Brand value. A familiar percentage is not automatically fair, and a high fixed payment is not automatically sustainable.

Percentage of eligible sale

Use a percentage when order values vary, transaction value is reliably tracked, and the Brand can define which components count. State whether taxes, shipping, discounts, gift cards, tips, financing, returns, and recurring renewals are included. Percentage commissions share upside, but they also expose both parties to refund and margin variation.

Fixed amount per conversion

Use a fixed amount when the qualifying event has predictable value or when simplicity matters more than order-size variation. Fixed commissions are easier to communicate and forecast. They become dangerous when low-value orders can trigger the same payment as high-value ones without sufficient margin.

Qualified lead

Use a lead commission only when the Brand can define quality before launch. Specify required fields, geography, consent, uniqueness, intent, contactability, exclusions, validation window, and what happens to duplicates. Paying for form submissions without a quality definition invites fraud and disputes.

Hybrid or tiered models

A Brand can combine a small qualified-lead amount with a larger completed-sale commission, or raise rates after verified volume or quality thresholds. Keep tiers prospective and measurable. Avoid opaque discretionary bonuses that cannot be reconciled from the program record.

Run the unit-economics test

Model customer value, contribution margin, expected refund rate, payment costs, support cost, Affiliate commission, platform fee, and expected conversion rate. Then test weak, expected, and strong scenarios. The program should not depend on an unrealistic zero-refund or perfect-retention case.

Publish examples

Show two or three commission examples using realistic order values and exclusions. An Affiliate should be able to reproduce the result from the written rule. If the Brand cannot explain the math in a short example, the system is not ready to scale.