Most brands asking about ecommerce cpa benchmarks are asking the wrong question. Not because benchmarks are useless, but because a headline number on its own tells you almost nothing about whether your paid media is healthy, scalable, or quietly burning margin.
A £25 CPA can be excellent for one retailer and disastrous for another. If you sell repeat-purchase skincare with strong lifetime value, you can afford to buy customers more aggressively. If you sell low-margin homeware with little repeat rate, that same CPA can wreck contribution profit. That is why serious operators do not treat CPA as a trophy metric. They treat it as one piece of a commercial model.
Why ecommerce CPA benchmarks get misused
The main problem with ecommerce cpa benchmarks is context. Agencies, platforms and industry reports often bundle together wildly different account types – branded search, prospecting social, Shopping, Performance Max, remarketing, high AOV brands, low AOV brands, seasonal peaks and clearance periods. Then they publish an average as if it is useful.
It usually is not.
A benchmark can tell you whether you are broadly in range. It cannot tell you whether your account is efficient, whether your offer converts, or whether you should scale. A brand with a £60 CPA might still outperform a competitor on profit if average order value, gross margin and repeat purchase behaviour are stronger. A brand with a £15 CPA might still be in trouble if returns are high and fulfilment costs are eating the order.
That is the first discipline. Benchmark CPA against business economics, not against internet averages.
What a “good” CPA actually depends on
If your team is reviewing paid media properly, CPA should never be looked at in isolation. The real question is simple: what is the maximum cost you can pay to acquire a customer while still hitting your margin target?
That number changes based on average order value, gross margin, new versus returning customer mix, discounting, shipping costs and channel intent. It also changes by growth stage. Brands trying to hold profit steady will tolerate a very different CPA from brands deliberately acquiring customers for repeat purchase.
Margin comes first
This is where weaker agencies usually lose the plot. They optimise for lower CPAs because lower looks better in a report. A specialist eCommerce team works backwards from breakeven cost per acquisition and target contribution.
If your average first order is £80 and your gross margin is 65%, you have far more room to acquire a customer than a brand selling a £35 basket at 45% margin. Add subscription or repeat purchase and the acceptable CPA can move again. Remove that repeat behaviour and the benchmark tightens fast.
Channel intent changes the benchmark
Search and Shopping traffic usually carry stronger purchase intent than paid social prospecting. That means CPA benchmarks should be tighter on bottom-funnel search terms and more flexible on cold audience acquisition, provided downstream conversion is proven.
If Google Shopping is producing a higher CPA than broad prospecting Meta campaigns, something is probably off. That could be feed quality, bid structure, pricing pressure or simply weak product-market fit. But if prospecting social is more expensive than branded search, that is not automatically a problem. It is expected.
New customer acquisition is meant to cost more
This point gets glossed over far too often. If you separate new and returning customer performance properly, new customer CPA will often look worse. That is normal. Returning users already know your brand. They are cheaper to convert.
The mistake is comparing blended account CPA with new customer acquisition goals. If your blended CPA looks good because remarketing and brand search are carrying the account, you may be underinvesting in actual growth.
Useful ecommerce CPA benchmarks by channel
There is no universal number, but there are useful commercial ranges if you understand what they are for.
For branded search, CPA should usually be your cheapest traffic source because intent is strongest. If it is not, there may be a conversion issue on site or poor query control. For non-branded search and Shopping, CPA often sits in the middle. These campaigns are where structure, feed quality and product segmentation matter most. For paid social prospecting, CPA is typically higher because you are creating demand rather than simply capturing it.
Performance Max muddies the picture because it blends inventory and intent types. That can make headline CPA look efficient while hiding waste. If you cannot see where spend is going, benchmarking becomes less useful. In those cases, profitability and incrementality matter more than the platform-reported acquisition cost.
As a broad commercial rule, many established eCommerce brands can tolerate the following pattern: cheapest CPA on brand and remarketing, moderate CPA on Shopping and non-brand search, and highest CPA on paid social prospecting. If your account breaks that pattern, there may be a strategic reason, but it deserves scrutiny.
When benchmark CPA is less important than MER or ROAS
Some businesses obsess over CPA when they should be managing a wider model. If your product range has varied order values, bundles, subscriptions or strong repeat purchase, CPA can distort decision-making.
Take two campaigns. One delivers a £20 CPA on a £50 first order. The other delivers a £35 CPA on a £140 first order. If you only optimise to lowest CPA, you cut the campaign creating more revenue and potentially more gross profit. That is bad account management dressed up as efficiency.
This is why many mature retailers balance CPA with ROAS, MER and contribution margin. CPA tells you what it cost to win the order. It does not tell you enough about what that order was worth. Good operators know when to prioritise each metric.
How to set your own benchmark properly
Start with breakeven. Work out what you can spend to acquire a first-time customer without losing money after cost of goods, fees, shipping and typical returns. Then decide whether you are willing to acquire above breakeven because of lifetime value.
Next, split targets by channel. Expecting Meta prospecting to hit the same CPA as branded Google Ads is fantasy. Set realistic benchmarks based on intent, not hope.
Then segment by product category. Your bestseller may support a stronger CPA than low-margin accessories. If everything is blended together, you will scale the wrong products and pause the right ones.
Finally, separate reporting by customer type where possible. New customer CPA, returning customer CPA and blended CPA all tell different stories. If you do not split them, your benchmark is weak before the analysis even starts.
Red flags hidden behind “good” CPA numbers
A low CPA can still mean poor account performance. That is uncomfortable for some teams because it removes the easiest talking point in a monthly report.
One common issue is over-reliance on remarketing. CPA looks strong, but the account is harvesting existing demand instead of generating new demand. Another is excessive brand bias. You can make numbers look tidy by pumping spend into branded terms, but that does not mean you are growing market share. A third is discount dependency. If CPA improves only when promotions get deeper, your media efficiency may be fine while your commercial model is getting weaker.
There is also the issue of scale. The best CPA in the account is often attached to the least scalable audience. That does not make it the best opportunity. A serious growth strategy weighs efficiency against room to expand.
What established brands should do with ecommerce CPA benchmarks
Use benchmarks as guardrails, not targets to worship. If you are materially outside expected range for your channel and business model, investigate. If you are inside range but profit is weak, investigate anyway.
The strongest accounts are not built by chasing average benchmark figures. They are built by tightening feed quality, improving landing page conversion, structuring campaigns around intent and margin, and cutting wasted spend without starving scale. That is where specialist management earns its keep.
For established retailers, the right question is not “what is the average CPA in my sector?” It is “what CPA can this business support by channel, by product set and by customer type while still growing profitably?” That question leads to better decisions.
Oxedent’s view is simple. If your CPA benchmark does not connect back to margin and scale potential, it is not a benchmark worth managing against.
The useful number is not the one that makes the report look cleaner. It is the one that tells you whether the next pound of ad spend is likely to produce profitable growth.