What does affiliate management cover?
How do you choose between CPA and revenue share?
A cost-per-acquisition (CPA) deal fits markets where lifetime value (LTV) is short and fraud is high. Revenue share fits products with a real month-three tail. Hybrid is a compromise that still needs a negative-carryover rule written into the contract. We set out the worked maths in our article on valuing affiliate revenue-share deals.
What does quality control look like in practice?
Checks for discount abuse, duplicate accounts, incentivised junk traffic and partners bidding on your brand terms. Partners who cannot pass a cohort review lose the deal, however large last month's clicks were.
How do you recruit without damaging the existing programme?
New partners are added to fill a gap in market or product, not to hit a volume target. Recruiting a copy of your best partner at a higher CPA is how programmes leak money.
What reporting do affiliates actually need?
New customers, net revenue, refunds and chargebacks, and a status they can trust. A 40-column portal that disagrees with finance is how good partners leave.
Where do affiliates meet CRM?
A revenue-share customer who churns on day four is a customer relationship management (CRM) failure billed to the partner. Retention and affiliate terms have to share a definition of value, so this service works closely with retention and CRM.
How is affiliate management measured?
| KPI | What it tells you | When we act on it |
|---|---|---|
| Net revenue after quality | Revenue that survives fraud, refund and discount filters. | When headline revenue and finance's revenue diverge. |
| Customer quality mix | Share of new customers who make a second purchase. | When a partner's repeat-purchase rate collapses. |
| Effective CPA | True cost per quality customer across deal types. | When the CPA implied by revenue share beats the CPA deals you are signing. |
| Partner concentration | Dependence on the top three sources. | When one partner effectively is the business. |
What happens in the first 90 days?
- Days 1–20: deal inventory, cohort quality by partner, a view of refunds, chargebacks and discount abuse, and reconciliation of the portal against finance.
- Days 21–45: freeze or renegotiate deals that fail on quality. Negative-carryover and brand-bidding rules go into the contract, not into Slack.
- Days 46–70: recruit only to fill a named gap. New partners are onboarded with the same event definitions media uses.
- Days 71–90: lock a quality scorecard and report net revenue after filters as the number the programme exists to grow.
Where do affiliate programmes usually go wrong?
Paying CPA on customers who never come back
A cheap customer who leaves after the welcome discount is not cheap. If the deal cannot see day-7 (D7) repeat purchases, you are paying for a screenshot.
Revenue share with no negative carryover
Partners keep the good months and you keep the refunds and discount losses. That is a one-way option, not a partnership.
Letting the loudest affiliate set terms for everyone
Most-favoured terms spread quickly. Each deal should be priced on the value it brings, which is the opposite of matching terms in a panic.
Related services: attribution so partner conversions match your own records, retention economics that show whether a revenue-share customer was worth it and paid acquisition, whose CPA is the benchmark affiliate deals compete with. For the numbers behind this approach, read a worked example of valuing CPA, revenue-share and hybrid deals.