Profit-Based ROAS: Optimising for Margin, Not Revenue

A 6.0x revenue ROAS can make less gross profit than a 3.0x return.

That is not a rounding error. It happens whenever the 6.0x product carries thin enough margin, and scaling it through a Google Ads programme can reduce post-media contribution while the dashboard improves.

Optimising to revenue in a mixed-margin business systematically buys your worst customers, and the fix is arithmetic you can do in a spreadsheet before touching the ad account.

Profit-based ROAS replaces revenue with verified profit as the conversion value. Google Ads then sees the commercial outcome finance cares about, rather than turnover with the costs hidden.

Why revenue ROAS buys low-margin demand

Google Ads ranks conversion opportunities using the values it receives. When those values represent revenue, every £1 of turnover looks equally useful regardless of what fulfilling the sale costs.

The algorithm is not making a financial mistake. The input never gave it the financial problem.

For this article, profit on ad spend means gross profit before media cost divided by ad spend:

POAS = gross profit before media cost ÷ ad spend

That definition must be agreed with finance before implementation. Calling revenue “profit” in a dashboard does not make it so.

Take an illustrative UK industrial supplier spending equally across two product groups during the same period. The two input sets are:

1. Product A inputs: £10,000 ad spend, £60,000 attributed ex-VAT revenue and 15% gross margin.
2. Product B inputs: £10,000 ad spend, £30,000 attributed ex-VAT revenue and 50% gross margin.

The complete Product A calculation is:

Revenue ROAS = £60,000 ÷ £10,000 = 6.0x

Gross profit before media = £60,000 × 15% = £9,000

Profit ROAS = £9,000 ÷ £10,000 = 0.9x

Post-media contribution = £9,000 − £10,000 = −£1,000

The complete Product B calculation is:

Revenue ROAS = £30,000 ÷ £10,000 = 3.0x

Gross profit before media = £30,000 × 50% = £15,000

Profit ROAS = £15,000 ÷ £10,000 = 1.5x

Post-media contribution = £15,000 − £10,000 = £5,000

Product A wins on revenue ROAS, 6.0x versus 3.0x. Product B wins on gross profit, £15,000 versus £9,000, and post-media contribution, £5,000 versus −£1,000.

The same illustrative account produces:

Account revenue ROAS = (£60,000 + £30,000) ÷ (£10,000 + £10,000) = 4.5x

Account profit ROAS = (£9,000 + £15,000) ÷ £20,000 = 1.2x

The account therefore reports revenue ROAS of 4.5x versus profit ROAS of 1.2x. That is the gap between an attractive marketing report and a modest commercial return.

This calculation assumes the attributed revenue is already trustworthy. If that number is disputed, resolve it using our approach to finding the real Google Ads return before introducing margin.

Revenue rewards volume; profit rewards the right volume.

Calculating profit-based ROAS before changing bids

Finance already holds the decisive variable: how much revenue remains after supplying each offer.

Build six spreadsheet columns: offer, net revenue, attributable delivery cost, gross profit, gross-margin rate and ad spend. Avoid a single account-wide margin unless every offer genuinely has similar economics.

The break-even revenue ROAS follows directly from gross margin:

Gross profit = Revenue × gross-margin rate

At media break-even:

Revenue × gross-margin rate = Ad spend

Divide both sides by ad spend:

Revenue ÷ Ad spend = 1 ÷ gross-margin rate

Therefore:

Break-even revenue ROAS = 1 ÷ gross-margin rate

For an illustrative offer comparison, the margin inputs produce two very different floors:

40% gross margin: 1 ÷ 0.40 = 2.50x break-even revenue ROAS

70% gross margin: 1 ÷ 0.70 = 1.43x break-even revenue ROAS

The first offer needs 2.50x versus 1.43x for the second merely to cover media spend. One universal revenue target cannot represent both accurately.

Once gross profit becomes the uploaded conversion value, break-even POAS is 1.0x. The target should sit above 1.0x by whatever post-media contribution finance requires.

Apply these three spreadsheet rules:

1. Gross-margin spread of at least 10 percentage points between offers → use offer-level or order-level profit values before enabling value-based bidding.
2. Target revenue ROAS below break-even for any offer receiving at least 20% of spend → freeze budget increases and raise the target to the calculated floor.
3. Margin data older than 30 days, or a verified margin movement of at least 5 percentage points → refresh costs before changing bids.

The honest limit here is that gross margin may exclude refunds, variable fulfilment, payment fees or sales commission. Where those costs move with each sale, contribution profit is the better input. Fixed overhead remains a finance planning decision.

Future customer value is deliberately excluded. That separate decision belongs in our guide to customer lifetime value and ROAS, not in an unverified multiplier added to current margin.

Margin turns an impressive return into a financeable decision.

Feeding margin data into profit-based ROAS

A submitted B2B form rarely carries a dependable profit value. The lead must keep its identity while the CRM and finance systems determine what was sold, whether it closed and what delivery will cost.

The margin handoff needs four controls:

1. Identity: Google Tag Manager or the site captures a stable transaction or lead identifier, subject to the visitor’s consent, and the CRM preserves it. Identifier loss needs a separate measurement plan for reduced-cookie environments.
2. Commercial key: The form, order or CRM record identifies the offer, customer segment and currency required to select the correct cost basis.
3. Economic state: Finance attaches verified gross profit at purchase or closed-won stage. If an earlier qualified-lead value is necessary, use verified segment close rate × expected deal gross profit, not one invented value for every enquiry.
4. Delivery and reconciliation: The value is imported into Google Ads against the original conversion, then aggregate profit is checked against CRM and finance totals for the same records, currency and period.

Website architecture matters because an ambiguous enquiry cannot be mapped reliably to an economic category. Our case studies of complex website journeys document that segmentation work: Lanteria routes a broad capability set for multiple stakeholders, Savgen structures a technical multi-industry offer, and Lake Erie Shores separates stay and ownership audiences. Those are architecture decisions, not performance claims.

Qualification also needs a stable commercial definition. A sales-accepted opportunity and an uncontactable form submission cannot carry the same expected profit, which is why lead-quality measurement must precede bid automation.

Clean profit signals give bidding systems an economically valid objective.

If the path from ad click to signed revenue still has gaps, Actualyse will trace each hand-off with you — book a call

What POAS does not fix

Four popular fixes leave the underlying signal broken.

1. Recalculating POAS after a bad collection event. POAS cannot correct a conversion value that was wrong at collection. A proposal recorded as closed revenue, a duplicated order or an incorrect product key remains wrong after division by ad spend.
2. Applying one blended margin percentage. Multiplying every conversion by the same percentage preserves the existing ranking. The low-margin offer still looks disproportionately valuable because the variation POAS needs has been averaged away.
3. Raising the revenue-based target ROAS. A stricter target may reject more auctions, but it still tells Google Ads to prefer revenue. It controls aggressiveness without correcting the objective.
4. Rebuilding campaigns or relabelling dashboards. New campaign names, additional asset groups and a “POAS” reporting column create no margin data. Excessive splitting can also scatter the profit-valued conversions the bid strategy needs to evaluate.

No bidding strategy can repair corrupted commercial inputs.

Bid strategy implications of profit-based ROAS

Do not change the conversion value, target and budget on the same morning. Simultaneous changes make a bad target, broken import and ordinary demand movement indistinguishable.

Maximise conversion value will maximise whichever value is supplied. Once gross profit becomes primary, a target ROAS constrains return on gross profit rather than return on revenue. The old revenue target therefore has no valid numerical relationship with the new profit target.

Use these four operating gates, which are Actualyse rules rather than Google Ads minimums:

1. Shadow gate: Calculate profit values alongside revenue for at least 14 days before allowing them to influence bidding.
2. Integrity gate: Require profit values on at least 90% of eligible conversions and a finance-to-platform reconciliation gap no greater than 5% across the trailing 30 days. Otherwise, repair the feed.
3. Volume gate: Fewer than 30 profit-valued conversions in 30 days → keep economically similar offers consolidated instead of creating smaller campaigns.
4. Change-control gate: Hold daily budgets within ±10% for the first 14 days after switching. A conversion-volume decline greater than 30% versus the preceding 14 days, while integrity remains sound, should trigger a target review before a budget increase.

Our central claim is falsifiable: after 60 days and at least 100 accurately profit-valued conversions, no movement towards higher expected gross profit under unchanged targeting and availability would prove it wrong.

Implementation crosses finance definitions, CRM states, site tracking and campaign controls. Our Google Ads management work for B2B companies treats those as one operating system rather than four disconnected jobs.

Budget follows verified profit, not dashboard confidence.

FAQ

Four edge decisions commonly block an otherwise workable implementation.

Should VAT be included in the conversion value?

Exclude VAT from both revenue and cost calculations. An ex-VAT feed-to-ledger discrepancy above 1% should block activation until the tax treatment is consistent.

How should refunds and cancellations be handled?

Net them from profit using conversion adjustments or the agreed import process. If reversals exceed 2% of monthly attributed revenue, update weekly rather than waiting for a month-end correction.

What should a multi-currency account upload?

Choose the finance reporting currency and use the booking-date conversion rate. An exchange-rate variance above 2% between the feed and ledger should trigger transaction-level currency conversion.

How can margin data remain commercially restricted?

Upload the final profit value without itemised cost components. Limit the underlying cost table to two accountable roles: the finance owner and the advertising operations owner.

Clear operating rules prevent metric theatre.

Summary

Five operating rules survive the arithmetic:

• Margin spread of at least 10 percentage points → replace blended revenue values with offer-level profit values.
• Revenue target below 1 ÷ gross margin on an offer receiving at least 20% of spend → freeze budget growth.
• Profit-value coverage below 90% or reconciliation error above 5% over 30 days → keep profit bidding inactive.
• Fewer than 30 profit-valued conversions in 30 days → consolidate economically similar campaigns.
• During the first 14 days, keep budgets within ±10%; a volume fall above 30% → review the target first.

Margin discipline keeps paid growth economically honest.

Actualyse builds measurement and attribution setups that tie B2B ad spend to real revenue. Book a call to talk through where yours stands.