Beyond POAS: Why Profitable Brands Still Run Out of Cash

TL;DR

POAS told us what to bid on. It did not tell us what to bid for.
A profitable Google Ads account can still drain cash, build inventory you cannot move, and train Google to find your worst customers.
Profit is one of six commercial layers. Reading only one of them is how brands miss the others drifting.
The Six-Layer Read: Profit · Cash · Inventory · Customer Quality · Brand Position · Commercial Trajectory.

The client call that broke my model of "good"

A few years ago a client showed me something that broke my model of what "good" looked like in Google Ads.

Their account was profitable. POAS at 2.4. Every campaign making money on every order. By every metric we championed at the time, it was a healthy account.

Six months later they nearly went under.

Not because the ads stopped working. Because the things POAS does not see were drifting, and nobody was reading them. That call changed how we read every other account on our books.

What POAS does well, and why we still use it

POAS replaced ROAS as the more honest primary metric for ecommerce Google Ads. The case is straightforward. ROAS treats every pound of revenue as if it were worth the same. A £100 sale on a 12% margin product and a £100 sale on a 60% margin product look identical to ROAS. POAS factors in COGS, shipping and returns to show the actual contribution per pound of ad spend.

For most brands at £3M+ revenue, switching from a ROAS-first account to a POAS-first account is the single biggest unlock in performance we have seen. We have argued it for years. We still do. But POAS is one number. A business is not one number.

The five things POAS quietly misses

A profitable Google Ads account can still:

Drain cash you do not have if payback periods extend beyond your working capital cycle.
Build inventory you cannot move if the platform optimises around the easy conversions and starves the products that need to clear.
Train Google to find your worst customers. High first-order POAS often correlates with low retention, especially in subscription brands.
Erode brand equity you spent a decade earning when discount-driven conversions look profitable per order but condition the market to wait.
Hide a 90+ day payback period your CFO will only discover at year end.

Each one is invisible if you read the account through POAS alone. That was the realisation.

The Six-Layer Read

We now read every account against six commercial layers, not one. If any one is drifting, the others are lying to you.

Profit (POAS)

What it measures: Contribution margin per pound of ad spend, with COGS, shipping and returns factored in.

What it tells you: Whether the ad pound was worth spending in isolation.

Signal it's drifting: Blended POAS holds steady but individual SKU POAS variance widens. The average is hiding a split.

Cash

What it measures: How long the ad spend takes to come back as recovered cash, given your payment terms, refund window and replenishment cycle.

What it tells you: Whether profitable growth is also fundable growth. Profit at 90 days is not the same as profit at 30.

Signal it's drifting: Payback period extending while POAS holds steady. Common in subscription and BNPL-heavy brands.

Inventory

What it measures: Whether the ads are moving the stock the business needs to move.

What it tells you: Whether ad spend is aligned with commercial priority or fighting it. A profitable bestseller running out of stock while ad budget keeps chasing it is leakage. So is a slow-burn SKU ageing out while the platform ignores it.

Signal it's drifting: Stock-at-risk rising on slow-movers; bestsellers going out-of-stock without bid pull-backs.

Customer Quality

What it measures: Whether the customers the platform is finding will stick around.

What it tells you: Whether you're acquiring assets or liabilities. High first-order POAS often hides low retention. A 4× POAS click that becomes a one-and-done customer is worth less than a 2× POAS click that subscribes for 14 months.

Signal it's drifting: Cohort LTV by acquisition source diverging. Performance Max cohorts often underperform here without commercial constraint.

Brand Position

What it measures: Whether the ads are building or eroding the brand.

What it tells you: Whether the account is borrowing from future demand to hit this month's number. Discount-led conversions look profitable in the account and expensive on the brand equity line.

Signal it's drifting: Discount-attributed revenue rising as a share of total; full-price conversion rate softening; brand search volume flat or falling.

Commercial Trajectory

What it measures: Whether the account is set up for the next 90 days, not the last 30.

What it tells you: Whether profitability today is structural or fragile. An account dependent on a single PMax asset group, one SKU, or one promotional calendar is one platform update away from a problem.

Signal it's drifting: Spend concentration rising; campaign diversity falling; reliance on automation increasing without commensurate constraint logic.

How the layers interact, and why interactions matter more than numbers

The point of reading against six layers is not to track six numbers. It is to spot the interactions between them.

Cash and Inventory drifting together usually means the brand is building dead stock funded by overdraft. The ads look profitable. The warehouse fills up. The bank gets nervous. See zombie SKUs for the pattern.

Customer Quality and Brand Position drifting together usually means promotional dependency. Google has learned to find bargain hunters, and the brand has learned to tolerate them. Margins compress. Repeat rate falls. First-order POAS still looks fine. This is the failure mode we see most often in fashion & apparel and subscription brands.

Profit and Commercial Trajectory drifting together usually means an account that is profitable today on a structure that will not survive the next change. PMax consolidating into a single asset group. One SKU carrying 40% of the spend. One promotional calendar carrying the year. Endemic in high-SKU retail.

Read any one of these in isolation and you miss the signal entirely. That is the whole point of widening the lens.

POAS vs The Six-Layer Read

What POAS sees	What POAS misses	Which layer catches it
Profit per pound of ad spend	Cash recovery timing	Cash
Order-level contribution	Stock health and ageing	Inventory
First-order economics	Cohort retention and LTV	Customer Quality
Per-conversion profitability	Brand equity erosion	Brand Position
Account performance this month	Structural fragility	Commercial Trajectory
Whether the bid was worth it	Whether the bid was right for the business	All six combined

How to start reading your own account this way

The Six-Layer Read is not a tool. It is a discipline. The minimum data inputs:

COGS by SKU
Inventory levels and ageing by SKU
Returns data by SKU and customer cohort
LTV by acquisition source, 90, 180 and 365-day
Discount-attributed revenue as a percentage of total
Payback period by channel
Spend concentration by campaign, ad group and asset group

Most £3M to £10M ecommerce brands have most of this data. Almost none of them feed it back into the Google Ads decision loop. That is the gap. Our Commercial Risk Index is how we score the layers above for the brands we work with.

Why this matters now

Google Ads automation has compounded the cost of reading the wrong layer. Performance Max optimises hard for whatever signal you feed it. If you feed it ROAS, it finds revenue. If you feed it POAS, it finds profit. Neither is enough on its own, and the more autonomy you grant the platform, the faster a drift in any one layer becomes structural.

The brands quietly winning in 2026 are not the ones with the best automation. They're the ones reading their accounts against more than one layer, and feeding the read back into the bid logic, the budget logic and the campaign structure. That is the difference between an ads agency and a commercial decision layer. It's also why our best clients don't ring us about ROAS anymore. They ring us about working capital.

If your blended POAS looks healthy but your CFO keeps asking why cash is tight, that's the gap.

Book a 30-minute commercial audit. We'll read your account against the six layers and show you which ones are drifting.

Book a 30-minute discovery call

Frequently asked questions

Is POAS still the right primary metric for ecommerce Google Ads?

Yes. POAS remains the most honest single metric for bid decisions and the right primary metric for ecommerce brands with variable margins. It is not sufficient on its own. It needs to be read alongside cash, inventory, customer quality, brand position and commercial trajectory.

What's the difference between POAS and contribution margin?

POAS measures profit per pound of ad spend. Contribution margin measures profit per pound of revenue after variable costs. POAS tells you whether the ad pound was worth spending; contribution margin tells you whether the product is worth selling. A high-POAS product can have low contribution margin if its absolute volume is small. A high-CM product can have low POAS if the platform cannot find buyers efficiently.

How is the Six-Layer Read different from a normal commercial dashboard?

A dashboard reports the numbers. The Six-Layer Read interprets the interactions between them and feeds that interpretation back into bid, budget and structural decisions inside the Google Ads account. A dashboard tells you cash is tight. The Six-Layer Read tells you cash is tight because the ads are pushing the wrong inventory at a payback period your working capital cannot fund.

Can we run the Six-Layer Read in-house?

Yes, if you have the data feeds and someone whose job is to read them weekly. Most £3M to £10M ecommerce brands have the data and not the someone. That is the gap we close.

How quickly will switching to a Six-Layer Read change account performance?

Two effects. Short-term, 30 to 60 days, waste reduction by spotting the layers that are already drifting. Long-term, 90 to 180 days, structural change as the account is rebuilt around the layers, not just POAS. The brands that see the largest commercial swing are the ones where two or more layers were drifting at once and the interaction had gone unread.

Does this apply to brands using Performance Max?

Especially. Performance Max optimises hard for whatever signal you feed it and compounds drift in any single layer faster than manual campaigns. The Six-Layer Read is how you set the guardrails.

What size of brand does the Six-Layer Read apply to?

Brands spending £10k+ per month on Google Ads with variable product margins, established inventory data, and someone in the business who understands unit economics. Below that, the data inputs usually are not there and POAS alone is a reasonable working metric.

BeyondPOAS:WhyProfitableEcommerceBrandsStillRunOutofCash