Measurement
Part of How to build affiliate marketing measurement that stands up
Affiliate marketing benchmark research rarely settles a performance argument
How to commission and read affiliate marketing benchmark research without misleading comparisons, covering sample design, disclosure and the evidence you need.
What to take away
- Decide first whether you need a sector benchmark or an internal baseline: they answer different questions and carry different risks.
- Check the sample before the headline numberwho was surveyed, how they were recruited, and whether the data covers England, the wider UK or another market entirely.
- Treat published benchmarks as context, not targets, because commission structures and tracking methods differ between programmes.
- Keep your own definition of a conversion stable across periods so your trend line means something.
Benchmark or baseline first
A benchmark compares your programme with others. A baseline compares your programme with its own past. Most teams ask for the first when the second would settle the argument faster.
Benchmark vs baseline
Benchmark
- Compares against
- Other programmes
- Data source
- Industry average
- Control
- External
- Settles argument
- Rarely
Baseline
- Compares against
- Own past
- Data source
- Internal data
- Control
- You control
- Settles argument
- Faster
If you cannot state your own conversion rate for the last six months, an industry average will not help you. Internal data is also the only evidence you control.
Fix your conversion definitions before you go shopping for comparison data. Your measurement and reporting standards should be agreed with finance and legal before any external figure enters the budget conversation.
What benchmark research covers
Published research from networks such as Awin, CJ Affiliate, Impact.com and Tradedoubler usually reports click-through rates, conversion rates, average order values and commission rates. Some studies add publisher mix and device splits.
Each figure depends on definitions the researcher chose. A conversion might mean a first purchase, any purchase, or a validated sale after the returns window closes.
A benchmark built on validated sales will look worse than one built on raw orders. That gap is definitional, not a performance difference.
Sample and geography checks
Ask where respondents came from. A survey of English retailers is not the same as a UK-wide study, and neither describes behaviour in the United States.
For example, a team reading a benchmark drawn from 200 English retailers should not apply it to a Scottish or Welsh programme without checking the sample notes.
Sample size matters less than how the sample was recruited. Self-selected panels attract particular kinds of respondent.
Reading a number honestly
Ranges beat single figures. If a study reports a median commission rate of 8 per cent, ask about the spread around it.
Suppose your programme pays a 6 per cent commission and converts at 3 per cent, against a cited median of 8 per cent and 2.5 per cent. The comparison tells you nothing until you match product margin and return rates.
For a worked example, a team paying £400 a month in platform fees and £2,000 in commissions against £18,000 of tracked revenue is working from its own numbers, not a benchmark. Those pound figures are illustrative.
Compliance evidence in research
Benchmark reports often describe disclosure practice. Verify claims against the regulator rather than the report.
The UK GDPR and the Privacy and Electronic Communications Regulations (PECR) set out lawful basis, consent and electronic-marketing rules that any affiliate campaign touching personal data must meet. The ICO's direct marketing guidance explains how these apply in practice.
In the United States, the FTC warning letters on deceptive marketing show how disclosure failures have been pursued. Read them as evidence of enforcement direction, not as rules binding UK programmes.
If a platform's conduct distorts the market, the route for reporting a competition or market problem to the CMA is the formal channel.
Reporting what you find
Write down the source, the date and the definition behind every figure you reuse. A benchmark without that context will be quoted back at you in a budget meeting.
Build the internal view first. Your own reporting dashboard should track the fields that make the trend line comparable month to month: validated conversions, commission paid, returns and net revenue.
Then place external research beside it, clearly labelled as context.
Common questions
Can I use a benchmark as a performance target?
Only with caution. Benchmarks describe other programmes with different products, margins and tracking. Use them to sanity-check a figure, not to set a target someone is held to.
How old can benchmark data be before it is useless?
Commission structures and tracking rules change, so treat anything more than two years old as historical context. Always check the publication date before quoting a figure.
Does UK-wide data apply to my English programme?
Usually as a broad guide, but check the sample. If the study only covers England, do not present it as a UK figure in reporting.
What should I record alongside a benchmark figure?
Record the source, publication date, sample description, geography and the definition used for each metric. Without those five items the number cannot be defended.
