Affiliate program KPIs: EPC, conversion and payout ratio

The seven affiliate program KPIs worth tracking monthly are EPC (commission earned ÷ clicks), click-to-customer conversion (new paying customers ÷ clicks), payout ratio (total partner cost ÷ partner-attributed revenue), active partner rate (partners with a new customer ÷ approved partners), revenue per earning partner, median time to first sale, and concentration (the share of commission earned by the top 20% of partners). In the worked example below, a SaaS program with 60 approved partners, 4,800 clicks and 48 new customers in a month has a conversion rate of 1.0%, a first-month EPC of about $0.10 but an expected 12-month EPC of $0.90, a payout ratio of 24.2% and a top-20% share of 81.6%. Each number answers a different question, and each has a specific fix when it moves the wrong way.

Which affiliate program KPIs actually matter?

Most dashboards show clicks, signups and commission, and stop there. Those are counts, not KPIs. A KPI is a ratio that tells you whether the program is healthy independent of its size, so you can compare this month with last month even if you doubled the partner count in between. The seven below cover the three things a partner program must do: send traffic that converts, cost a sensible share of what it brings in, and not depend on one or two people.

How to calculate EPC for a recurring program

EPC is commission earned divided by clicks. It is the number affiliates use to compare programs, because it combines your conversion rate, your price and your commission rate into one figure they can multiply by their own traffic. A publisher who sends 1,000 clicks a month can read an EPC of $0.90 as roughly $900 a month in earnings.

Recurring commission complicates it. A customer acquired this month pays a commission this month and, if they stay, every month after. If you divide this month's commission from new customers by this month's clicks, you get a first-month EPC that understates the value of the click. If you divide all commission this month by this month's clicks, you mix in renewals from customers who clicked a year ago. Report both of the following, and label them.

• First-month EPC = commission on first invoices of new customers this month ÷ clicks this month. It is what a partner sees quickly.
• Expected EPC = new customers this month × expected commission per customer ÷ clicks this month. It is what a click is worth over the commission period, using your churn.

With a $49 plan at 20% for 12 months, the first invoice pays $9.80. With 5% monthly churn, expected paid invoices in 12 months are (1 − 0.95¹²) ÷ 0.05 ≈ 9.2, so expected commission per customer is about 9.2 × $9.80 ≈ $90. At 48 new customers from 4,800 clicks, first-month EPC = 48 × $9.80 ÷ 4,800 = $470.40 ÷ 4,800 ≈ $0.098, and expected EPC = 48 × $90 ÷ 4,800 = $4,320 ÷ 4,800 = $0.90. The same month, nine times apart. Publish the expected figure with its assumptions, because that is the honest comparison with a one-time bounty program.

What is a good affiliate conversion rate?

Click-to-customer conversion is new paying customers divided by clicks. Count paying customers, not trial signups, because a SaaS trial that never converts costs you support time and pays the partner nothing. If you have a free trial, it helps to track the two halves separately: clicks to trials, and trials to paid.

There is no universal good number, because conversion depends on price, trial design and how well the partner's audience matches your buyers. A $9 tool bought on impulse and a $499 tool bought after a demo will never share a benchmark. What matters is the comparison inside your own program: the same partner over time, and partners of the same type against each other. A comparison site converting at 3% and a YouTube channel at 0.4% can both be fine; the same comparison site dropping from 3% to 1% in a month is a signal.

Conversion is also the KPI most damaged by bad attribution. If customers type a partner's coupon code instead of clicking, or buy on a different device weeks later, the sale is still the partner's but the click and the customer no longer line up. Personal coupon codes and a referral window long enough for your sales cycle, such as 60 days, keep conversion honest.

Payout ratio: what share of revenue goes to partners?

Payout ratio is total partner cost divided by partner-attributed revenue in the same period. Total cost means everything you pay partners: recurring commission, first-payment bounties, tier bonuses and fixed fees for creator or newsletter placements. Revenue means invoices paid by customers attributed to partners, net of refunds.

If you only pay a flat recurring percentage, the payout ratio equals that percentage and the KPI tells you nothing new. It becomes useful once you add bounties, bonuses and fixed fees, because those can push the real cost well above the headline rate without anyone noticing. Refunds matter too. Stripe does not return its processing fees when you refund a payment [2], and disputes pull the payment plus dispute fees from your balance [3]. A program that reverses commission on refunds and disputes keeps the ratio honest; a program that does not pays commission on revenue it never kept.

Active partner rate and revenue per partner

Active partner rate is the share of approved partners who brought at least one new paying customer in the period. It is usually low, and that is normal: many people join programs, few promote them. The KPI is useful as a trend and as a check on your recruiting. If you approve 30 new partners a month and the active rate falls, you are approving people who will never promote.

Revenue per earning partner divides partner-attributed revenue by the number of partners who had any commission row in the period, including renewals. It answers how big a working partnership is. Combined with the active rate, it tells you which lever to pull: if revenue per earning partner is healthy but the active rate is low, help more partners start; if the active rate is fine but revenue per partner is small, help existing partners do more, for example with a better comparison page or a coupon.

Time to first sale and partner concentration

Time to first sale is the median number of days from a partner's approval to their first attributed paying customer, measured over partners approved in the last six months. The median is the right average here because a few partners who take a year would distort a mean. If it grows, onboarding is the first suspect: unclear links, no ready-made assets, or a program page that does not explain what to say.

Concentration is the share of commission earned by the top 20% of earning partners. In most partner programs a small group produces most of the result, so a high number is expected. The KPI exists to tell you how exposed you are. If one partner produces half of all partner revenue and changes their recommendation, the program halves overnight. Track the share, know who is in the top group, and treat a rising number as a reason to recruit, not as a success.

A worked monthly affiliate report

Here is one month for an illustrative SaaS program. The plan is $49 a month, commission is 20% for 12 months, and the brand also paid one creator a $600 fixed fee for a video. All counts are invented for the example; every KPI is computed from them.

Now the arithmetic, step by step. First-month EPC = 48 × $9.80 ÷ 4,800 = $470.40 ÷ 4,800 ≈ $0.098. Expected EPC = 48 × $90 ÷ 4,800 = $0.90. Click-to-customer conversion = 48 ÷ 4,800 = 1.0%. Payout ratio = ($2,842 + $600) ÷ $14,210 = $3,442 ÷ $14,210 ≈ 24.2%. Active partner rate = 12 ÷ 60 = 20%. Revenue per earning partner = $14,210 ÷ 20 = $710.50. Time to first sale: the nine values are already sorted, so the median is the fifth, 31 days. Concentration = $2,320 ÷ $2,842 ≈ 81.6%.

Reading the report: the payout ratio of 24.2% is above the 20% headline rate because of the creator fee, which is fine if that video brought customers and worth checking if it did not. The active rate of 20% is typical of a roster that has never been pruned. Concentration of 81.6% means four partners carry the program; the first recruiting goal is a fifth and sixth partner of the same type as the top two.

What to do when a KPI is off

There are no reliable public benchmarks for most of these numbers across SaaS programs, so the useful question is direction: compared with your own last three months, is the number moving the wrong way? Each KPI has a short list of likely causes.

One number deserves a hard threshold: disputes on partner-referred customers. Stripe notes that the card processing industry treats dispute activity above 0.75% as excessive, and that card networks run monitoring programs that can lead to fines [1]. Stripe also notes that cardholders can dispute a charge up to 120 days after payment, so recent months will keep moving [1]. If disputes cluster on customers from one partner, stop paying that partner's commission until you know why.

Building the report every month

1. Close the month on the same day each time, after renewals and refunds for the month have posted.
2. Export clicks, new customers, revenue and commission rows by partner. Use net revenue (after refunds) and net commission (after reversals).
3. Compute the seven KPIs with the formulas above. Keep the formulas in the spreadsheet so anyone can check them.
4. Compare each KPI with the last three months, not with an industry number.
5. Pick one KPI that moved the wrong way and one action from the diagnostic table. Do that one thing this month.
6. Share a short version with partners: their own clicks, customers, EPC and next payout date. Partners who can see their numbers tend to fix their own links.

RelayWonder keeps commission in an append-only ledger with one row per Stripe invoice, refunds and disputes as negative rows, so net commission and net revenue for a month come from the same source. That is the part of the report that is hardest to get right by hand.