What's A Good ROAS Or CPA? Benchmarks And How To Set Your Own

“We’re doing a 3.2 ROAS. Is that good?” I get asked this most weeks, and the honest answer is that I have no idea until I know the margin. A 3.2 on a 60% gross margin business is money in the bank. The same 3.2 on a 25% margin is a slow, well-attributed loss. There is no industry number that settles it. Your break-even ROAS is 1 divided by your gross margin, and everything above that line is the only benchmark that matters.

Same story for CPA. A £120 cost per lead is either brilliant or ruinous depending on what a lead is worth to you once your close rate has had its say. Below: why the benchmark tables you find on Google are close to useless for a decision, the arithmetic that gives you your own numbers, and the four figures I would rather look at than any benchmark.

Why industry benchmarks mislead

Benchmarks are not fabricated. They are just answering a different question to the one you are asking.

Take the most widely cited set, WordStream’s 2026 Google Ads search benchmarks, drawn from thousands of campaigns across more than 20 industries:

Metric	All-industry average
Click-through rate	6.64%
Cost per click	$5.42
Conversion rate	8.18%
Cost per lead	$66.69

Useful context. Terrible target. Three reasons.

The spread inside a category is wider than the gap between categories. In the same dataset, cost per lead runs from $26.84 in arts and entertainment to $131.63 for attorneys and legal services. Sit inside legal and the range between a personal injury firm and a conveyancing practice is just as brutal. “Legal” is not a market. It is a filing cabinet.

The figures are US dollars and US-weighted. UK auctions price differently, VAT changes the revenue you report, and a US-heavy average tells a Manchester plumber very little about their local auction.

Averages hide the mix. An account running heavy brand traffic posts a flattering blended ROAS and a low CPL because it is buying back demand it already had. Strip the brand campaigns out and the number often halves. Compare your blended figure to someone else’s blended figure and you are comparing two traffic mixes, not two performances.

Use benchmarks for one thing only: a sanity check that you are in the right order of magnitude. If your CPL is 10x the category average, something is broken. If it is 20% off, that tells you nothing at all.

Work out your break-even ROAS

This is one line of arithmetic and it beats every benchmark table.

Break-even ROAS = 1 ÷ gross margin.

Gross margin here means revenue minus variable costs, divided by revenue. Cost of goods, payment fees, shipping, packaging, returns. Not your net margin, and definitely not a number you half-remember from a board deck.

Gross margin	Break-even ROAS
20%	5.00
30%	3.33
40%	2.50
50%	2.00
60%	1.67
70%	1.43

Read that table across and the “is 3.2 good?” question answers itself. At 50% margin, a 3.2 ROAS is 60% clear of break-even and you should be pushing spend. At 25% margin, break-even is 4.0 and a 3.2 means every extra pound you spend loses money faster.

Two things people get wrong here. First, they use revenue including VAT, which inflates ROAS by 20% on paper and hides a loss. Report ex-VAT conversion values. Second, they forget returns. If 15% of orders come back, your effective margin is lower than your product margin and your real break-even sits higher than the table suggests.

Turn break-even into a target you can bid to

Break-even is where you stop losing money. It is not where you want to run. You need a target that leaves profit on the table after the ad spend has been paid.

Target ROAS = 1 ÷ (gross margin − desired profit margin).

At a 45% gross margin, wanting 15% profit after media: 1 ÷ (0.45 − 0.15) = 3.33. That is the number you bid to.

A simpler version I use when a client wants a quick answer: take break-even and multiply by 1.3 to 1.5. It builds in a buffer for the gap between platform-attributed revenue and what actually lands in the bank, which is real and usually runs in Google’s favour.

One practical note if you are handing that number to Smart Bidding. Google’s Target ROAS is entered as a ratio against your reported conversion value, and Google’s own guidance for Shopping is that a strategy needs a floor of conversion volume to work with, around 15 conversions in the last 30 days per Merchant Center ID. Below that, you are asking an algorithm to optimise on noise. Get volume first, then get clever. Our breakdown of Google Ads bidding strategies covers which strategy suits which data volume.

Lead gen: the same maths, run through your close rate

Ecommerce gets to work in ROAS because the sale happens on the site. Lead gen does not, so you run the same logic in reverse and land on a target CPA.

Four numbers:

1. Gross profit per closed customer. Average deal value multiplied by gross margin.
2. Lead-to-customer close rate. From your CRM, not from a guess.
3. Maximum acceptable CAC. The share of that gross profit you are willing to spend to win a customer.
4. Target CPA = maximum CAC × close rate.

Worked through: a £4,000 average deal at 40% margin gives £1,600 of gross profit. You will spend a third of it to acquire, so £533 maximum CAC. Close 20% of leads and your target cost per lead is £533 × 0.20 = £107.

Now that $66.69 all-industry average looks exactly as relevant as it is. Your number is £107 because of your deal size, your margin and your sales team. Nobody else’s average knows any of that.

The catch in lead gen is lead quality. Optimise to raw form fills and Smart Bidding will happily find you cheap, worthless ones. Feed qualified leads and closed revenue back into the platform through offline conversion imports, and the target CPA you set starts meaning something.

Four numbers I would rather look at than a benchmark

Once you have your own break-even, these tell you more than any category average. They also sidestep the trap we wrote about in why ROAS is misleading: a single blended ratio flattens very different traffic into one number.

• Contribution after ad spend. Gross profit minus media cost, in pounds. A ROAS that drops from 4.0 to 3.4 while contribution climbs is a good month. Ratios go down as you scale. Profit is what you bank.
• Marginal ROAS. Not the account average, but the return on the last chunk of budget you added. Scaling decisions live here. The average is history.
• New customer CAC. Strip out returning buyers and brand searches. What does it cost to buy a person who did not know you? That number decides whether the channel is growth or just harvesting.
• Payback period. How many months until a customer’s gross profit covers their acquisition cost. For subscription and repeat-purchase businesses this beats first-order ROAS outright. Our post on media buying unit economics goes deeper on the four metrics worth tracking.

How often to revisit the numbers

Recalculate break-even ROAS whenever your margin moves. A supplier price rise, a shipping increase or a discount campaign all shift the line, and most accounts are still bidding to a target set against last year’s cost base.

Practically: check margin quarterly, review your target ROAS or target CPA against actual profit monthly, and never change a bid target on less than two or three weeks of data. Changing targets weekly keeps campaigns in permanent learning and teaches you nothing.

Want a second opinion on your numbers?

Most accounts I look at are bidding to a target nobody can explain. If you want to know whether yours clears the margin, our free Google Ads audit reviews tracking, structure and the targets you are actually bidding to, and tells you where the profit is leaking. If it turns out the maths is fine and the execution is not, that is what our paid search management is for.

FAQ

What is a good ROAS? Any ROAS above your break-even, which is 1 divided by your gross margin. At 50% margin that is 2.0, at 25% margin it is 4.0. A 4x ROAS is often quoted as a general benchmark, but it is only good if your margin clears it.

What is a good ROAS for ecommerce? Most ecommerce brands run at 30% to 50% gross margin, putting break-even between 2.0 and 3.3 and a sensible target somewhere between 3.0 and 5.0. Work out your own figure from margin rather than adopting the range, because product mix and returns move it significantly.

Is a 2.5 ROAS good? It is profitable if your gross margin is above 40%, and a loss if it is below. On a 40% margin, 2.5 is precisely break-even, so you are working for nothing. Check the margin before you judge the number.

What is a good cost per lead? The one that sits below your target CPA, calculated as your maximum acceptable acquisition cost multiplied by your lead-to-customer close rate. All-industry averages sit near $67 in US search data, but a high-value B2B business can profitably pay several times that.

Should I use ROAS or CPA as my target? Use ROAS when order values vary and you can pass reliable conversion values back to the platform. Use CPA when every conversion is worth roughly the same, which is most lead generation. Passing a fake or flat conversion value into a ROAS target is worse than bidding to CPA honestly.