A £15,000 campaign can look like a 4.0x success because the ad account shows £60,000 of conversion value. Evaluating a Google Ads programme properly means removing £6,000 of refunds and credits, then adding £3,250 of management, creative and tracking costs; the true ROAS calculation falls to 2.96x.
Apply delivery costs, transaction fees and sales commission, and only £7,940 of contribution profit remains.
That gap is not an attribution debate. It is arithmetic. The useful calculation starts with net realised revenue, divides it by the cost required to acquire it, and then tests the result against contribution margin.
The true ROAS calculation: two numbers, not one
Standard ROAS divides attributed revenue by media spend:
Platform ROAS = attributed conversion value ÷ media spendThat number is useful for managing campaigns, but it is incomplete for commercial decisions. A defensible calculation uses net revenue and fully loaded acquisition cost:
True revenue ROAS = net realised attributable revenue ÷ fully loaded acquisition costCalculate the two inputs as follows:
Net realised attributable revenue
= attributed gross revenue
− refunds and returns
− cancellations and chargebacks
− credits and retrospective discounts
− VAT or other taxes collected on behalf of governmentFully loaded acquisition cost
= media spend
+ campaign management fees
+ allocated creative and landing-page costs
+ platform or tracking fees directly required by the campaignA 3.0x result means the campaign generated £3 of net revenue for each £1 of acquisition cost. It does not mean it generated £3 of profit.
Margin belongs in a second calculation:
Contribution before acquisition
= net realised attributable revenue
− cost of delivery or goods
− payment fees
− variable sales commission
− other variable fulfilment costsContribution ROAS
= contribution before acquisition ÷ fully loaded acquisition costContribution profit after acquisition
= contribution before acquisition − fully loaded acquisition costReporting both true revenue ROAS and contribution ROAS preserves the distinction between revenue efficiency and profit. Our practical method for finding the real return from Google Ads explains how this commercial view fits into wider performance analysis.
The starting conversion value still needs a defensible basis. If that is uncertain, use the separate examination of what Google Ads ROAS does and does not accurately represent before applying the formulas above.
What belongs in a true ROAS calculation
Use figures from the same customer cohort. Dividing revenue from old customers by this month’s acquisition cost can produce an impressive ratio with no economic meaning.
For revenue, start with sales attributed to the acquired customers and exclude VAT. Deduct refunds, returns, cancellations, service credits and discounts that were not already reflected in the original value.
For B2B contracts, choose a revenue basis and label it clearly. Recognised revenue is conservative. Collected revenue is useful where bad debt matters. First-year contract revenue can support acquisition decisions when delivery costs and early churn are also included. Avoid presenting unweighted pipeline as revenue.
For acquisition cost, include spending that was necessary to run the campaign:
- Media spend, including non-recoverable platform charges
- Agency or freelance management fees
- Campaign-specific creative production
- Landing-page work allocated across its expected useful period
- Incremental call tracking, feed or measurement costs
Shared retainers need a documented allocation rule. General company overhead does not belong in campaign ROAS. Internal marketing salaries can appear in a separate fully loaded view, but an arbitrary salary allocation should not be treated as spend that disappears when a campaign is paused.
Classify fees once. A management fee belongs in acquisition cost. A card-processing fee belongs among variable costs when calculating contribution. A refund reduces revenue. Putting the same cost in two places understates the result.
Use one consistent source for each input rather than averaging incompatible figures. Where Ads, analytics and CRM reports conflict, apply a defined framework for reconciling ROAS discrepancies across tools before doing the maths.
Worked true ROAS calculation
Consider a B2B supplier selling through an online quotation and checkout journey. All values below exclude VAT.
| Input | Amount |
|---|---|
| Ad-account conversion value | £60,000 |
| Refunds and cancellations | £5,000 |
| Credits and discounts | £1,000 |
| Media spend | £15,000 |
| Management fee | £2,000 |
| Creative and landing-page allocation | £1,000 |
| Incremental tracking cost | £250 |
| Delivery cost | £24,300 |
| Payment fees | £810 |
| Variable sales commission | £2,700 |
Step 1: Calculate net realised revenue
£60,000 − £5,000 − £1,000 = £54,000The campaign produced £54,000 of usable attributable revenue, not £60,000.
Step 2: Calculate fully loaded acquisition cost
£15,000 + £2,000 + £1,000 + £250 = £18,250The platform calculation used only the £15,000 media spend:
£60,000 ÷ £15,000 = 4.00xThe true revenue ROAS is:
£54,000 ÷ £18,250 = 2.96xThat is 26% lower than the headline 4.00x figure.
Step 3: Apply margin and variable costs
Delivery consumes 45% of net revenue, leaving a 55% gross margin. Payment fees and sales commission reduce the usable contribution further:
£54,000 − £24,300 − £810 − £2,700 = £26,190The contribution margin before acquisition is:
£26,190 ÷ £54,000 = 48.5%Contribution ROAS is:
£26,190 ÷ £18,250 = 1.44xContribution profit after acquisition is:
£26,190 − £18,250 = £7,940The campaign remains profitable, but the commercial position is much tighter than a 4.00x platform ROAS suggests.
True ROAS for B2B lead generation
Lead-generation campaigns need an interim forecast because revenue can arrive months after the advertising cost. Keep that forecast separate from realised ROAS.
Suppose a campaign incurs £12,000 of fully loaded acquisition cost and generates ten sales-qualified opportunities. Historical cohorts show a 30% win rate, an average first-year contract value of £25,000 and an 8% reduction for cancellations and credits.
Expected gross revenue = 10 × 30% × £25,000 = £75,000
Expected net revenue = £75,000 × 92% = £69,000
Expected true revenue ROAS = £69,000 ÷ £12,000 = 5.75xAt a 40% contribution margin before acquisition:
Expected contribution = £69,000 × 40% = £27,600
Expected contribution ROAS = £27,600 ÷ £12,000 = 2.30x
Expected contribution profit = £27,600 − £12,000 = £15,600Call this expected ROAS. Once the cohort matures, replace probabilities with revenue from won customers. The distinction matters in Google Ads lead generation for B2B companies, where lead volume alone says little about commercial return.
If the path from ad click to signed revenue still has gaps, Actualyse will trace each hand-off with you — book a call
Set a break-even ROAS before spending
Break-even depends on contribution margin, not an industry benchmark:
Break-even true revenue ROAS = 1 ÷ contribution marginFor the worked example:
1 ÷ 48.5% = 2.06xEvery £1 of acquisition cost therefore needs £2.06 of net revenue to cover variable costs and acquisition. The achieved 2.96x is profitable.
A break-even target leaves no allowance for fixed overhead or desired profit. If the company requires contribution profit after acquisition equal to 15% of revenue, the target becomes:
Required ROAS = 1 ÷ (48.5% − 15%) = 2.99xThe campaign’s 2.96x result now sits just below target. That leads to a different budget decision than the reported 4.00x.
Whether campaigns are run internally or through Google Ads management tied to commercial outcomes, the margin and target should be agreed before budgets change. When reviewing case studies showing campaign results, check whether the quoted return uses gross conversion value, net revenue or contribution. Ratios with different definitions are not comparable.
Common calculation mistakes
Do not mix revenue including VAT with costs excluding recoverable VAT. Do not use gross margin from the company accounts if campaign customers have different delivery costs or discount levels.
Avoid dividing lifetime revenue by the cost of acquiring a recent cohort unless lifetime value has been adjusted for churn, collection risk and future delivery cost. Do not compare a forecast ROAS for open opportunities with realised ROAS from closed customers.
Finally, keep a short calculation sheet showing the period, cohort, revenue basis, deductions, cost allocation and margin assumption. A repeatable definition is more useful than a supposedly perfect number that changes every month.
FAQ
What is a good true ROAS?
There is no universal figure. At a 25% contribution margin, break-even revenue ROAS is 4.00x. At a 60% margin, it is 1.67x. Add the company’s required profit above those thresholds.
Should agency fees be included in ROAS?
Include them when deciding whether the acquisition programme is commercially worthwhile. You can retain a media-only ROAS for campaign management, but label it separately and do not use it as the profitability measure.
Is true ROAS the same as profit?
No. True revenue ROAS measures net revenue relative to acquisition cost. Contribution ROAS applies margin and variable costs. Contribution profit then subtracts acquisition cost in pounds.
How should late refunds or cancellations be handled?
Restate the original customer cohort when material refunds arrive. For quicker forecasting, apply a historical refund or cancellation rate, then replace the estimate with actual deductions as the cohort matures.
Summary
- True revenue ROAS divides net realised attributable revenue by fully loaded acquisition cost.
- Deduct returns, cancellations, credits and VAT before calculating revenue ROAS.
- Include management, creative, landing-page and directly required technology costs.
- Apply delivery costs, transaction fees and commission through contribution ROAS.
- Set the target from contribution margin: break-even ROAS equals 1 ÷ contribution margin.
Actualyse builds measurement and attribution setups that tie B2B ad spend to real revenue. Book a call to talk through where yours stands.
