POASforDTC:thedefinitiveguidetoprofitonadspend

1. Definition and formula

POAS (Profit on Ad Spend) is a paid-media efficiency metric expressed as a multiple. It compares the contribution profit generated by a campaign, ad group, audience or SKU to the advertising spend that generated it.

The numerator is sometimes called contribution margin 3 (CM3): the cash a sale leaves behind after every variable cost directly attributable to fulfilling that order. It is the most operationally honest definition of profit available to a paid-media team. It excludes fixed overheads (which paid spend cannot influence) and includes everything that scales with each additional sale.

A POAS of 1.0x means a campaign breaks even on contribution. Above 1.0x, the campaign is generating cash. Below 1.0x, the campaign is destroying it, regardless of how impressive the ROAS appears.

2. Why ROAS fails DTC brands

ROAS (Return on Ad Spend) is revenue divided by spend. It was designed for an era when ecommerce margins were uniform and supply chains predictable. Neither is true now.

Three structural failures of ROAS for DTC ecommerce:

1. Margin variance. A modern DTC catalogue can contain a 50-percentage-point gross margin spread between SKUs. ROAS treats a £100 sale of a 22% margin hero product identically to a £100 sale of a 65% margin accessory. Smart Bidding optimised on ROAS will allocate spend toward whichever SKU has the highest conversion rate, not the highest profit contribution.
2. Returns blindness. Apparel returns rates routinely exceed 30%. A 5x ROAS campaign on a 35% returning category is functionally a 3.25x ROAS campaign net of returns, before COGS is even loaded. Google has no native mechanism to reflect category returns rates in bidding.
3. Promotional decay. A 4.0x ROAS hit during a 25% sitewide discount is a different commercial event from a 4.0x ROAS hit at full price. ROAS hides the difference. POAS surfaces it because the discount is loaded into the contribution calculation.

The result, observed across audits of more than 200 DTC accounts: roughly 60-70% of brands tracking ROAS are profitable on paper and unprofitable in cash. The metric is not lying; it is answering a different question to the one finance is asking.

For the long-form treatment of this gap, see From ROAS to POAS and the audit walkthrough in Google Ads can turn £100 into £1 and still call it a win.

3. Calculating POAS in practice

POAS calculation requires four data inputs that most DTC brands hold in fragmented systems: ad spend (Google Ads), revenue (ecommerce platform), variable costs per SKU (ERP or spreadsheet), and returns by SKU (3PL or Shopify reports).

The minimum viable POAS calculation uses a single blended margin number applied across all SKUs. This is enough to expose accounts that are losing money, even if it cannot yet steer bidding precisely. The mature POAS calculation uses per-SKU contribution margin, refreshed weekly, passed into Google's conversion value layer at the moment of purchase.

A worked example. A skincare brand spends £40,000 on Google Ads in a month. Reported revenue is £180,000. Reported ROAS is 4.5x. Loaded:

• COGS at blended 32%: £57,600
• Fulfilment and packaging at £4.20/order, 3,200 orders: £13,440
• Payment processing at 1.9%: £3,420
• Returns at 8% of revenue, fully refunded: £14,400
• Promotional discount net of regular price: £9,000

Contribution profit: £180,000 − £57,600 − £13,440 − £3,420 − £14,400 − £9,000 = £82,140. POAS: £82,140 ÷ £40,000 = 2.05x. The brand is profitable, but at less than half the strength the headline ROAS implied. Any campaign in the account performing below 2.0x POAS is destroying contribution.

4. Implementing POAS in Google Ads

There are three viable mechanisms for getting profit values into Google's bidding stack. Each suits a different operational maturity.

Option A: Custom profit values via conversion adjustment API

Engineering-heavy. Requires a server-side process that calculates per-order profit (typically 24-72 hours after the sale, once returns and refunds settle) and writes the adjusted value back to Google via the Conversion Adjustment API. Suits brands with strong data engineering capacity.

Option B: Profit-as-conversion-value at checkout via GTM

The most common implementation. A server-side GTM tag fires on order confirmation and looks up margin per SKU from a maintained reference table (typically a Google Sheet or BigQuery view). The conversion value sent to Google is profit, not revenue. Requires monthly margin discipline. This is the standard JudeLuxe implementation for accounts spending £10k-£100k/month.

Option C: Third-party POAS platforms

ProfitMetrics, smec/Whisbi, Northbeam and similar. These ingest order, cost and ad data, calculate margin centrally, and feed Google's bidding via their managed integrations. Suits brands with complex returns mechanics, multi-currency operations or subscription components. Adds £400-£2,000/month in tooling cost.

In all three cases, Smart Bidding (Maximise Conversion Value with target ROAS) becomes the optimisation engine. The target ROAS field is repurposed as the target POAS. A target ROAS of 3.0 in a profit-fed account is a target POAS of 3.0.

5. POAS and Performance Max

Performance Max is the most aggressive value-chaser in the Google ecosystem. Feed it revenue and it will reshape asset group spend toward whichever SKUs convert most easily, regardless of margin. This is why so many DTC brands report rising PMax revenue and falling cash.

Switching the conversion value layer to profit changes PMax behaviour within 14-21 days. We have observed three consistent shifts post-implementation:

• Asset group spend redistributes toward higher-margin SKUs without manual segmentation
• Brand cannibalisation drops as PMax stops valuing repeat-buyer revenue at the same weight as new acquisition revenue (when LTV-adjusted profit values are used)
• Search term overlap with brand campaigns reduces, because PMax sees less profit value in conquesting brand intent

For the structural critique of PMax, see Performance Max is not a strategy, it's a slot machine.

6. Target benchmarks by mode

Target POAS is a function of business mode, not industry. The same brand will run different POAS targets across the year as cash position and growth priority shift.

7. SKU roles and POAS

POAS at account level is an average. The decisions live one layer down, in how each SKU contributes. The JudeLuxe SKU role framework classifies every SKU into one of five jobs:

• Scale: high-margin, high-velocity. Push spend.
• Profit: high-margin, moderate-velocity. Defend bid position, never starve.
• Protect: brand-defining. Maintain visibility, accept lower POAS.
• Recovery: ageing inventory or surplus stock. Run aggressively to clear at any positive POAS.
• Gateway: low-margin acquisition products. Loss-leader logic, measured at LTV-POAS not first-order POAS.

Account-level POAS targets only make sense once SKU roles are defined. Otherwise the metric averages out, and Smart Bidding is left to discover the wrong winners.

8. The limits of POAS

POAS is a powerful operating metric, not a complete commercial framework. Three honest limitations:

1. Data quality dependency. POAS is only as good as the margin data feeding it. Brands with chaotic costing or untracked supplier price changes cannot run POAS cleanly. Margin discipline is a precondition, not a side benefit.
2. Brand and upper-funnel undervaluation. POAS measures last-touch contribution. It under-weights assisted conversions, brand awareness lift and the long compounding of repeat purchase. Pair POAS with a separate measurement framework for upper-funnel investment, such as geo-experiment incrementality testing or media-mix modelling.
3. Subscription and LTV blindness. First-order POAS punishes acquisition-heavy categories. Brands with strong second-order economics (subscriptions, replenishment, premium retention) should run an LTV-adjusted POAS that loads expected 12-month contribution into the value signal.

9. The profit-first operating model

POAS is the bidding layer of a wider operating model. The full stack:

1. Margin truth. Per-SKU contribution margin maintained monthly.
2. SKU role classification. Every SKU assigned one of five jobs.
3. POAS bidding. Profit values fed to Google. Targets calibrated to mode.
4. Margin-aware budget pacing. Daily and weekly budget shifts based on POAS performance, not ROAS.
5. CFO-aligned reporting. Monthly view that maps ad spend to P&L impact, not platform metrics. Our CFO Board Pack is the public template.

We call this the Commercial Decision Layer: the layer of judgement that sits between the ad platform and the P&L. It is what JudeLuxe runs for every retained client.

FAQ