Measurement
Part of How to build affiliate marketing measurement that stands up
Affiliate marketing key metrics without the vanity numbers
A practical guide to the affiliate marketing key metrics that matter, with a scoring rubric for English retailers, plus reporting and compliance pointers.
What to take away
- A Manchester retailer running a small affiliate programme often sees clicks rise while sales stay flat, because those are not the affiliate marketing key metrics that decide budget.
- Track sales, new customer share, commission cost and incrementality together; a single ratio such as EPC or ROAS tells you little on its own.
- Score each publisher against a rubric before you renew a contract or shift spend, so the decision rests on evidence rather than anecdote.
- Reporting should separate observed sales from modelled or incremental sales, and name the source and date for every figure.
- Data protection and advertising rules apply to how you collect and publish these numbers, not only to the campaign itself.
Start from the numbers a finance director will ask about
A finance director rarely asks about click-through rate. They ask what the programme cost and what it returned. Sales, commission paid and net revenue after commission are the three figures that survive a budget meeting. Everything else supports or explains those three.
For an English retailer, the practical set is small: transactions, average order value, commission rate, new customer share and the cost per acquisition. Add a retention measure if your programme targets repeat buyers. If you want the full reporting structure around these figures, the affiliate marketing measurement and reporting guide 2027 sets out how they fit together.
Write a one-line definition for each metric and keep it fixed. If average order value excludes delivery one month and includes it the next, the trend tells you nothing. Note the currency, the tax treatment and the date range on the report itself, so any reader can rebuild the figure.
Separate observed sales from incremental sales
An affiliate sale is observed when a tracked click precedes a purchase. It is incremental only when the sale would not have happened otherwise. Those are different claims, and mixing them flatters performance.
Use a holdout group, geo test or audience split where you can. Where you cannot, label the figure as modelled and say so on the report. The IAB Measurement Centre publishes measurement guidance that helps teams describe attribution method consistently across channels.
Record the test window, the size of the holdout and the confidence you place in the result. A single fortnight rarely settles the question. If the uplift sits inside the noise, say so rather than rounding it up into a headline.
Build a scoring rubric before you judge a publisher
A rubric turns a subjective review into a repeatable decision. Score each publisher or partner on the metrics below, then set a threshold for renewal, review or removal.
| Metric | What it measures | Suggested weight | Score 1 (weak) | Score 3 (acceptable) | Score 5 (strong) |
|---|---|---|---|---|---|
| Incremental sales share | Sales that would not have occurred otherwise | 25% | Below 10% | 10% to 30% | Above 30% |
| Net revenue after commission | Revenue left once commission is paid | 20% | Negative | 0% to 10% margin | Above 10% margin |
| New customer share | Buyers new to the brand | 15% | Below 5% | 5% to 20% | Above 20% |
| Cost per acquisition | Commission spend per acquired sale | 15% | Above target CPA | At target CPA | Below target CPA |
| Return rate | Refunds and cancellations after commission | 10% | Above 15% | 5% to 15% | Below 5% |
| Reporting timeliness | Days from month end to reliable data | 10% | Over 10 days | 5 to 10 days | Under 5 days |
| Compliance evidence | Disclosures and data handling checks | 5% | None recorded | Partial | Documented and current |
Weights are illustrative. For example, a team paying £400 a month in commission may weight new customer share higher than a mature programme would. Fix the weights before the review, not after it, and keep the same rubric across publishers so comparisons stay fair.
Keep the data lawful and the report honest
The Information Commissioner's Office publishes audits and overview reports that show how organisations handle personal data in practice. Read those before you expand any pixel or server-side tracking, because consent choices decide which sales you can measure at all.
Advertising effectiveness research from the IPA research on advertising effectiveness can help you place affiliate results alongside other channels rather than judging them in isolation. That context matters when a publisher claims credit for sales that brand search or email also influenced.
If your dashboard needs rebuilding around these metrics, the affiliate marketing reporting dashboard in England covers layout, refresh cadence and the fields to expose.
Common questions
Which affiliate marketing key metrics should a small English retailer track first?
Start with transactions, commission paid, net revenue after commission and new customer share. Add cost per acquisition once you have enough volume to compare publishers fairly.
How do I know if affiliate sales are incremental?
Run a holdout or geo test where feasible. If that is not possible, present the figure as modelled and state the method and assumptions in the report.
How often should the rubric be reviewed?
Quarterly is workable for most programmes. Review weights when strategy changes, such as a shift from volume to new customer acquisition, and record the change.
Do reporting rules differ across the UK?
Data protection law applies UK-wide, while advertising codes are administered nationally. Check the relevant guidance for where your customers sit before publishing claims.
