Benchmarking answers a question no internal report can: is this number normal. This guide is for anyone who has looked at their cost per acquisition and had no idea whether to be pleased. It covers what the comparison set is, which metrics benchmark usefully, and how to avoid the trap benchmarking sets.
What you are being compared against
The comparison is against anonymised brands of similar size on Tatheon, not against published industry averages.
That distinction matters. Public benchmarks are compiled from whoever chose to report, which skews towards brands with something flattering to say, and they usually span such a wide range of business types that the average describes nobody. A comparison against businesses of your scale, measured the same way from the same underlying sources, is a genuinely different thing.
No individual brand is identifiable, and yours is not identifiable to anyone else either.
Which metrics benchmark well
Efficiency ratios benchmark well because they are scale independent. Conversion rate, repeat purchase rate, average order value relative to category, blended efficiency, gross margin. Two businesses of different sizes can be compared meaningfully on these.
Absolute figures benchmark poorly. Knowing that a peer earns more revenue tells you nothing you can act on.
Cost per acquisition sits in between and needs care. It is only comparable against businesses with similar order values and margins, because the acceptable cost of acquiring a customer follows directly from what that customer is worth. A high cost per acquisition against a high lifetime value is a healthy business.
The trap
Benchmarking invites you to treat the median as a target. It is not one.
Being below the median on a metric is only a problem if that metric is a constraint on your business. Plenty of successful brands sit well below average on conversion rate because they sell considered, expensive products where a long decision is normal. Optimising towards a benchmark set by impulse purchase brands would be actively wrong.
The useful reading is not the position, it is the outlier. If most of your metrics sit near the middle and one is far below, that one is worth investigating. A uniform position slightly below the median across everything is usually a business model difference rather than a set of problems.
Reading it in combination
The strongest signal comes from two metrics read together.
A high cost per acquisition alongside a high repeat purchase rate is a business that buys customers expensively and keeps them, which works. A high cost per acquisition alongside a low repeat rate is a business paying premium prices for customers who leave, which does not.
Similarly, a low conversion rate alongside a high average order value is often a considered purchase behaving exactly as it should. A low conversion rate alongside a low order value usually is a problem.
What to do with it
Use benchmarking to decide where to spend your attention, not what to aim for. Find the metric where you are furthest from your peers, then check whether it is genuinely a constraint on your business or a consequence of what you sell. If it is a constraint, that is the highest value thing on your list. If it is a consequence, note it and move on.