Affiliate Program KPIs: The 12 Metrics That Actually Matter
How to measure affiliate program performance from the brand side: 12 KPIs with formulas, how Awin, CJ, Rakuten Advertising and impact.com define EPC, sourced benchmarks and a simple dashboard layout.
Most affiliate dashboards show dozens of numbers. Only a handful tell you whether the program is healthy, whether it is growing for the right reasons and what to fix next. This guide covers the 12 affiliate program KPIs worth tracking from the brand side, how to calculate each one, where networks define them differently, and what to do when one moves the wrong way.
The short version
- Track outcomes (revenue, orders, AOV, conversion rate), partner health (EPC, active partner rate, concentration, new partners activated) and quality (new-customer share, reversal rate, cost of sale, incrementality).
- Check how your network defines EPC before comparing it. CJ reports it per 100 clicks; Awin, Rakuten Advertising and impact.com describe it per click.
- Your own trend is the best benchmark. Published figures are useful context, not targets.
The 12 KPIs and their formulas
Use the same date range and the same attribution rules for every metric, and decide up front whether you report gross figures (all tracked sales) or net figures (after returns and declined sales). Mixing the two is the most common reason affiliate reports don’t match finance.
| KPI | Formula | What it tells you |
|---|---|---|
| 1. Affiliate revenue | Sum of approved affiliate-attributed sales in the period | Size of the channel |
| 2. Orders | Count of approved affiliate-attributed orders | Volume, without AOV swings |
| 3. Average order value (AOV) | Affiliate revenue ÷ orders | Basket size from partner traffic |
| 4. Conversion rate | Orders ÷ clicks × 100 | How well partner traffic converts |
| 5. EPC | Commission paid ÷ clicks (× 100 for CJ-style EPC) | How attractive you are to partners |
| 6. Active partner rate | Partners with 1+ sale in the period ÷ all approved partners × 100 | How much of the partner base is working |
| 7. New-customer share | Affiliate orders from new customers ÷ all affiliate orders × 100 | Whether partners bring new buyers |
| 8. Reversal rate | Declined or reversed orders ÷ all tracked orders × 100 | Returns, fraud and tracking quality |
| 9. Cost of sale | (Commission + network fees + bonuses) ÷ affiliate revenue × 100 | What the channel really costs |
| 10. Incremental revenue | Revenue with partners − revenue you would have had without them | Whether partners create sales or just take credit |
| 11. Partner concentration | Revenue from top 10 partners ÷ affiliate revenue × 100 | How fragile the program is |
| 12. Partners activated | Partners whose first sale fell in the period | Whether recruitment turns into revenue |
Revenue and orders
Revenue and order count are the headline numbers, but read them together. Revenue can rise on a few large orders while order volume falls, and order volume can rise on cheap baskets that barely cover the commission.
How to measure: use approved (validated) sales, not pending ones, and report them net of returns once your validation window closes. Also track affiliate revenue as a share of total online revenue so growth isn’t confused with seasonality.
When it moves the wrong way: split the change by partner. If one or two partners explain the whole drop, it’s a relationship issue. If it’s spread across everyone, check tracking and your offer first.
AOV and conversion rate
Awin defines average order value as sale amount divided by the number of sales, and conversion rate as the percentage of referred visitors who made a purchase (Awin glossary). Together they tell you whether partner traffic is the right traffic.
How to measure: compare the affiliate conversion rate with your site-wide rate from your own analytics, and break both down by partner type. Coupon and loyalty traffic often converts higher than content or creator traffic because it arrives later in the purchase, so a blended rate hides a lot.
When it moves the wrong way: a falling conversion rate with steady clicks usually points to tracking gaps, out-of-stock or mismatched landing pages, or a competitor with a better offer. Our guide to why affiliate programs stop growing covers how to check each one.
EPC (earnings per click)
EPC is the commission partners earn on average for each click they send. It is the number publishers use to compare programs, so it is effectively your program’s price tag. Definitions differ by network:
- CJ reports 7-day and 3-month EPC as average earnings per one hundred clicks (CJ developer docs).
- Awin calculates EPC as commission in a timeframe divided by the number of clicks (Awin glossary).
- Rakuten Advertising divides commissions earned by the click-throughs that generated them (Rakuten Advertising glossary).
- impact.com describes EPC as the average commission partners make for each click they generate (impact.com glossary).
So, as a simple example, a CJ EPC of 50 and an Awin EPC of 0.50 (in the same currency) describe the same earnings. Always check the unit before comparing programs or networks.
When it moves the wrong way: EPC falls when conversion rate, AOV or your commission rate falls, or when more sales are declined. Work out which of the three moved before you change commissions.
Active partner rate
This is the share of approved partners who drove at least one sale in the period (90 days is a sensible window for most brands). It is often the most revealing number in the program, because a long list of approved partners can hide the fact that very few are promoting you.
How to measure: export partners with their sales for the last 90 days, count those with one or more approved sales, and divide by the total approved. Track the number of partners activated each month alongside it (KPI 12), so you can see whether recruitment is turning into revenue.
When it moves the wrong way: clear the application backlog, refresh creatives and contact your best-fit inactive partners personally before you recruit more.
New vs returning customers
Affiliate sales from existing customers aren’t worthless, but they are the sales most likely to have happened anyway. New-customer share shows whether partners are growing your customer base.
How to measure: pass a new/existing customer flag with each conversion. Most platforms support it, and some let you pay different commissions on it; Awin, for example, lists new or existing customer as a commission rule on its advertiser plans (Awin pricing).
When it moves the wrong way: raise the new-customer rate for content and creator partners, and check whether a coupon partner is capturing your existing customers at checkout.
Reversal rate
The reversal (or decline) rate is the share of tracked orders you later reject for returns, cancellations, fraud or rule breaches. Too high and partners stop trusting your program; too low may mean nobody is checking.
How to measure: declined orders divided by all tracked orders, by partner and by reason, each month.
When it moves the wrong way: a spike at one partner is a compliance question. A spike across the program usually points to returns, a tracking change or a checkout bug.
Incrementality
impact.com’s glossary defines incrementality as the lift in a metric (such as sales) brought about by the marketing investment, comparing people who had a touchpoint with people who didn’t (impact.com glossary). It is the hardest affiliate KPI to measure and the most important one for deciding where commission goes.
How to measure: start simple. Track new-customer share and conversion paths by partner type. Then run holdout tests: pause a partner, a code or a placement in some regions or weeks and compare against a control. Treat brand-search and coupon-at-checkout sales as low incrementality until a test shows otherwise.
When it moves the wrong way: shift commission toward partners who introduce new customers, and tighten brand bidding and code rules.
Partner concentration
This is the share of revenue from your top partners. Track the top 1, top 5 and top 10 each quarter. A high share isn’t automatically bad, but it means one partner’s change of terms can move your whole channel.
When it moves the wrong way: set a target for revenue from partners ranked 11 to 50 and point recruitment and activation there. Our guide to recruiting quality publishers covers where to find them.
Benchmarks (and how to use them)
Published affiliate benchmarks are rare, and most come from one platform’s own clients, so treat them as context, not targets. Two useful references:
- impact.com’s 2025 benchmark, covering 2,368 North American retail brands, found clicks up 2% year on year, conversion rates down 6% and transactions down 5%, while AOV rose 4% from $118 to $123. Loyalty and rewards partners drove 50% of transactions (impact.com 2025 benchmark).
- Partnerize says it often sees conversion rates of 4–8% for retailers focused on bottom-of-funnel tactics, and stresses that the right rate varies with your objectives, partner mix and product price (Partnerize).
Your most reliable benchmarks are your own: the same month last year, your trailing 12-month average and your site-wide conversion rate. If your network shows peer data (Awin’s Index, for example, scores programs on click-to-sale conversion, percentage of sales approved, EPC and validation turnaround), compare against programs in your own category.
Dashboard example
A one-page monthly view is enough for most programs. This is an example layout, not a template with target figures:
| Section | KPIs | Compare against | Review |
|---|---|---|---|
| Outcomes | Revenue, orders, AOV, conversion rate | Last month and same month last year | Monthly |
| Partner health | EPC, active partner rate, partners activated | Trailing 3 months | Monthly |
| Quality | New-customer share, reversal rate, cost of sale | Your targets | Monthly |
| Risk | Top 1/5/10 concentration | Last quarter | Quarterly |
| Value | Incrementality tests and results | Control group | Per test |
Add one line of commentary per section: what moved, why, and what you’ll do about it. That turns a report into decisions. For the wider operating rhythm, see our guide to affiliate program management, and for a full health check use the affiliate program audit checklist.
If you want help setting up the reporting or acting on what it shows, our affiliate program management service covers both. Get in touch.
Frequently asked questions
What are the most important affiliate program KPIs?
For most brands: revenue, orders, AOV and conversion rate for outcomes; EPC, active partner rate and partner concentration for program health; and new-customer share, reversal rate, cost of sale and incrementality for quality.
How is EPC calculated in affiliate marketing?
EPC is commission divided by clicks. Networks differ on the unit: CJ reports EPC per 100 clicks, while Awin, Rakuten Advertising and impact.com describe it per click. Check the unit before comparing programs.
How do you measure affiliate program performance?
Use one date range and one attribution rule for every metric, report approved sales net of returns, compare each KPI with last month and the same month last year, and add a line of commentary on what moved and what you will do about it.
What is incrementality in affiliate marketing?
Incrementality is the extra sales a partner creates beyond what you would have had anyway. You estimate it with new-customer share, conversion paths and holdout tests that pause a partner or code for part of your audience and compare the results.
Get in touch
Let’s talk about your program.
Tell us about your brand and what you need: launching a program, fixing one, or finding better partners. We’ll come back within one business day with how we’d approach it.
