Profit on Ad Spend
ROAS tells you what you sold. POAS gets closer to what you made.
Google Ads reports revenue extremely well. But ecommerce businesses don't keep revenue. From every order come product costs, discounts, payment fees, fulfilment, shipping, returns and eventually advertising spend.
What's left matters considerably more than the number Google puts in the conversion value column. That's why we use Profit on Ad Spend alongside conventional advertising metrics.
Because £1 of revenue isn't always worth £1 of revenue.
What is POAS?
POAS measures the profit generated relative to the advertising investment required to generate it.
The formula
POAS = Profit ÷ Ad Spend
£20,000 of profit from £10,000 of advertising spend is a POAS of 2.0: £2 of profit for every £1 invested, before whatever costs have deliberately been excluded from your chosen profit definition. That last part matters a lot.
What do we mean by "profit"?
You need to define it before calculating POAS. There isn't one universal calculation suitable for every business. You could calculate against gross profit, contribution before advertising, contribution after certain variable costs, order-level profit or another commercially useful definition.
For ecommerce advertising we usually want to get considerably closer to contribution than revenue alone: product cost, discounting, payment processing, pick and pack, shipping contribution, royalties, returns and other transaction-level costs.
Consistency matters more than inventing the world's most philosophically perfect definition of profit. Agree what goes into the calculation, apply it consistently, then use it to make decisions.
POAS vs ROAS
ROAS asks how much revenue advertising generated. POAS asks how much profit it generated. Neither question is stupid. They can produce very different answers.
Product A
£500 revenue. £200 variable costs. £300 contribution before advertising. £100 ad spend.
5.0 ROAS · 3.0 POAS
Product B
£500 revenue. £375 variable costs. £125 contribution before advertising. £100 ad spend.
5.0 ROAS · 1.25 POAS
Same ROAS, completely different economics. Managed towards a single 5.0 ROAS target, Google sees both as equally successful. Your P&L doesn't.
Blended ROAS makes the problem worse.
| Product | Ad spend | Revenue | ROAS | Contribution | POAS |
|---|---|---|---|---|---|
| A | £10,000 | £50,000 | 5.0 | £30,000 | 3.0 |
| B | £10,000 | £50,000 | 5.0 | £15,000 | 1.5 |
| C | £10,000 | £50,000 | 5.0 | £8,000 | 0.8 |
At account level: £30,000 spend, £150,000 revenue, 5.0 ROAS. Lovely. Commercially, A is creating significant value, B is much weaker and C isn't covering its advertising cost under this definition. The blended number has hidden the entire problem, and scale makes averages more dangerous: with 10,000 SKUs you will not spot it manually.
Work out your own POAS
POAS Calculator
Enter your numbers. See ROAS, POAS and net profit after ads.
ROAS
6.67x
Gross profit (before ads)
£6,000
POAS
4.00:1
Contribution after ads
£4,500
POAS = gross profit before ads ÷ ad spend. A POAS of 1.5:1 means £1.50 of gross profit per £1 of ad spend, leaving £0.50 of contribution after the ad cost. These figures exclude warehousing, staff, platform fees and any other overhead, so they are not net profit. There is no universal "good" POAS: the right target depends on your margin structure, fixed costs and what each product is there to do. POAS is a registered trademark of ProfitMetrics.
What is a good POAS?
There isn't one universal answer. A POAS of 2.0 might be excellent for one business, poor for another and intentionally aggressive for a third. The right target depends on your cost structure, fixed costs, commercial objective, customer value, growth expectations, cash position, product lifecycle and the profit definition you're using in the first place.
Anyone giving every ecommerce business the same "good POAS" target is missing the point. The target should come from your economics.
What is breakeven POAS?
If your numerator is contribution before advertising, a POAS of 1.0 means advertising has consumed all of that contribution: £10,000 of contribution against £10,000 of spend. Under another definition of profit, the interpretation changes. Which is why we don't throw POAS numbers around without defining the calculation first.
POAS isn't a magic metric.
Changing the dashboard from ROAS to POAS does not suddenly make the strategy commercially intelligent. You can optimise POAS badly too: underinvesting in profitable growth, stopping acquisition of valuable customers, ignoring strategically important products, failing to clear ageing inventory or overprotecting short-term margin.
POAS improves the measurement. You still need to decide what you're trying to achieve.
POAS measures. BOI® decides.
POAS asks what economic return the advertising generated. BOI® asks what commercial job the product should be doing. The hierarchy is commercial objective → BOI® → appropriate measurement → Google execution. Not "POAS good, ROAS bad".
A Scale SKU
Strong margin, healthy stock, growing demand. We care about POAS, but maximising it isn't the objective. If we can accept a lower POAS while generating substantially more total contribution, that may be the better business decision.
A Profit SKU
Established demand and strong economics. Here POAS matters much more: we're deliberately trying to maximise contribution from the available demand.
A Gateway SKU
Modest first-order economics, but it disproportionately acquires customers with strong repeat behaviour. Cutting investment because first-order POAS looks weak could be exactly the wrong decision.
A Recovery SKU
Ageing inventory, a new range arriving, capital trapped in stock. We may deliberately tolerate lower POAS to increase sell-through before the product needs deeper markdown.
Higher POAS is not automatically better.
Scenario A: £10,000 spend, £40,000 contribution, 4.0 POAS. Scenario B: £50,000 spend, £125,000 contribution, 2.5 POAS. If the objective is maximising total contribution and the additional investment is commercially acceptable, Scenario B creates £85,000 more contribution despite the lower POAS. Efficiency and total economic output are different things.
Average POAS vs marginal POAS
Your historic average tells you what existing investment produced. It doesn't tell you what the next £1 will produce. Spending £100,000 a month at 3.0 POAS says nothing about whether the next £50,000 returns 2.8, 1.8 or 0.7. Scaling should depend on the economics of incremental investment, not the average efficiency of existing spend.
POAS helps expose cross-subsidisation.
One category produces enormous contribution. Another barely breaks even. Managed together, the profitable category subsidises the weak one and the account-level result still looks healthy. That can continue for years. Product-level POAS makes the subsidy visible, so you can decide whether it's intentional, strategically justified, or simply waste.
Where POAS changes the decision
Google Shopping
Shopping operates at product level, and products have different economics. Connecting product identifiers with commercial data gives spend, revenue, margin, contribution, POAS, stock and commercial role at SKU level. The feed tells Google what the product is. Commercial data tells us what it's worth. Explore Google Shopping.
Performance Max
Google has substantial freedom over where budget goes, and high-revenue products naturally look attractive when judged on conversion value. We want to know which products PMax spends against, what they contribute, and whether brand demand or existing customers are involved. Explore Performance Max.
New customers
A customer generating £30 first-order contribution who never returns can look better on first-order POAS than one generating £10 now and £150 over the following year. For acquisition-led businesses we combine POAS with nCAC, first-order contribution, lifetime value, repeat rate and payback. Explore customer acquisition.
Returns and discounts
£100,000 of attributed revenue with £30,000 returned no longer represents the final economics of those orders. And two £100 orders, one full price and one heavily discounted, look identical in revenue and nothing alike in contribution.
Inventory
With £300,000 of ageing stock approaching markdown, strict POAS optimisation might tell you to protect efficiency, when the better commercial decision is increasing investment at a lower POAS to release cash earlier. This is why we don't describe ourselves as "a POAS agency". We're a commercially-led ecommerce PPC agency. POAS is one input into the decision. An important one, but still an input.
How we calculate POAS
1
Agree the commercial definition.
What costs matter? What does finance consider contribution?
2
Establish product economics.
COGS, margin, variable costs and returns where available.
3
Connect transactions to products.
Usually through ecommerce and order-level data.
4
Reconcile identifiers.
Less glamorous than strategy and considerably more important than people expect. If product IDs don't match across the platform, Merchant Center, Google Ads and commercial data, the beautiful profitability model becomes a beautiful collection of unrelated numbers.
5
Calculate contribution.
Using the agreed cost definition, applied consistently.
6
Connect contribution with advertising investment.
At the most useful level available: account, campaign, category, product, market or customer type.
7
Validate it.
Does the result reconcile with the business? If Google Ads says the company has discovered perpetual motion while finance says profit is falling, we investigate.
The formula isn't the hard part.
Anyone can put profit ÷ ad spend into a spreadsheet. The difficult parts are reliable cost data, matching product IDs, accounting for returns, handling bundles, understanding discounts, dealing with shipping, defining contribution consistently, separating new and returning customers, and deciding what to do with the result. POAS is a data problem before it becomes a bidding strategy.
You don't need perfect data to start asking better questions. But you do need to understand what's missing. False precision is worse than an honest estimate.
Does POAS replace finance reporting?
No. Please don't tell your CFO that a Google Ads metric has replaced the P&L. POAS is a decision metric for advertising. Finance remains responsible for the complete economics of the business. The objective is alignment, not creating a second accounting department inside Google Ads.
Does POAS prove incrementality?
No. POAS can tell us whether attributed activity generated sufficient profit under the chosen model. It doesn't prove those sales would not have happened anyway. That needs holdout testing, geo experiments, brand suppression tests or other causal analysis. Profitable attribution isn't the same thing as incremental profit.
POAS isn't perfect.
Neither is ROAS, attribution, lifetime value or MER. There isn't one metric that perfectly describes an ecommerce business. Sometimes POAS is central. Sometimes nCAC matters more. Sometimes inventory matters more. Sometimes incrementality is the unanswered question. The metric follows the commercial problem.
When POAS is particularly useful
- Product margins vary significantly.
- Your catalogue is large.
- Discounting materially affects profitability.
- Returns differ between products.
- Google Shopping or PMax represents meaningful spend.
- Revenue is growing faster than profit.
- Finance doesn't recognise the performance marketing is reporting.
- You need better product-level allocation decisions.
What profit measurement changed for our clients
Thermos
After establishing more accurate measurement and rebuilding advertising decisions around commercial performance, contribution increased 94%.
Read the Thermos case study →Flavour Blaster
International growth managed against regional economics rather than one global revenue target, including +114% POAS alongside substantial international revenue growth.
Read the Flavour Blaster case study →The important question isn't "what's our POAS?"
It's what decision knowing our POAS changes. Should this product receive more investment, or less? Should its target change? Should it move into another commercial job? Are we acquiring customers profitably? Is one product subsidising another? Has scaling reached diminishing returns?
Measurement earns its keep when it changes allocation.
Managing to profit already? See how we implement it in client accounts on our POAS agency page, or read the deeper comparison of POAS vs MER vs ROAS.