2 views of the same quarter. The numbers are an example:
| Meta dashboard | Bank account |
|---|---|
| ROAS 4.2 | Cash -£18,000 |
| Revenue +12% | Down for the quarter |
The ads are working. They are working at a job that was never going to make the business profitable.
Rising iOS attribution gaps, more competition, creative fatigue and platform volatility are real. In the accounts we audit, it is usually one of 2 structural problems.
Cause 1: the retargeting pool dried up
Retargeting audiences (site visitors, add-to-carts, customer lookalikes) convert well and report ROAS of 6x, 8x, sometimes 12x. The instinct is to push more budget to what the dashboard says works.
Those people were going to buy. The retargeting ad intercepted a journey already under way and claimed the credit.
Meanwhile cold prospecting, which reaches people who have never heard of you, loses budget because its reported ROAS is lower. Fewer new people enter the funnel. Then the retargeting pool shrinks, those high-ROAS campaigns weaken and revenue drops. The explanation becomes “Meta got more expensive”. The real one is that the top of the funnel was starved while the dashboard looked fine.
Cause 2: the margin never supported the ads
The second is simpler. The contribution margin cannot pay for paid acquisition, and nobody modelled it.
Say you run Meta at 3.5x ROAS. Your contribution margin (after COGS, fulfilment, returns and transaction fees) is 28%:
You spend £1 to keep 98p: a 2p loss before salaries, warehouse or software. Break-even at 28% is 1 ÷ 0.28 = 3.57, so 3.5 loses money. The ads work exactly as designed. The margin cannot support them.
Meta is a traffic channel. If the margin and prices cannot pay for paid traffic, campaign optimisation will not fix that.
4 signals to check
| Signal | How to check | What it tells you |
|---|---|---|
| 1. New customer rate | Pull 90 days of Shopify orders. What share are first-time buyers? | If it is falling while order volume holds flat, the brand is not growing |
| 2. Blended ROAS | Total Shopify revenue ÷ total ad spend, all channels | A 4.2x Meta ROAS means little if blended ROAS is 2.8x and your break-even is 3.2x |
| 3. Retargeting share | Meta spend by audience type | Above 35 to 40% on retargeting is bottom-heavy, and healthy accounts run 60 to 75% prospecting |
| 4. Contribution margin per order | Revenue minus COGS, fulfilment, returns and transaction fees | If your current CPA leaves nothing after this, no ROAS number fixes it |
What the fix looks like
In rough order of impact:
- Rebalance toward prospecting. Platform ROAS will dip in the short term. Track MER during the change. It shows whether total efficiency is improving while Meta’s own number looks worse.
- Review margin and pricing. Sometimes the answer is “the product needs to be £6 more expensive” or “returns need to fall by 5 points”. Ads cannot fix a broken margin.
- Set a separate target for new-customer CPA. Track it apart from retargeting, so you can see whether prospecting works.
- Run MER weekly. MER is total revenue ÷ all marketing spend. Let that number, not the Meta dashboard, decide whether marketing is on track.
Platform ROAS measures the sales Meta can claim. It does not measure whether the business model works.
Related reading
- Platform ROAS is a lie: why the dashboard always looks better than the business
- What is blended ROAS?: the metric that would have caught this sooner
- What is MER?: the top-level measure that shows when paid growth is working
Check your own margin before you spend more on Meta.
If the numbers still disagree, our Meta ads agency team can audit the account, or book a free call.
Quick answers
How can Meta ads show strong ROAS if the business is losing money?
Platform ROAS measures attributed revenue per ad pound spent, not whether that revenue is profitable. If your contribution margin is 35% and your real blended ROAS is 2.5x, you spend £1 to generate £2.50 and keep 87.5p of it, so every £1 of ads loses 12.5p before fixed costs. The business loses money even with a strong ROAS number.
What is the most common cause of strong Meta ROAS but declining business health?
Over-indexing on retargeting. Retargeting campaigns generate high reported ROAS because they reach warm audiences who were going to buy anyway: they harvest existing demand rather than creating new demand. As the brand stops growing its prospecting reach, the addressable retargeting pool shrinks and the high ROAS hides the declining top-of-funnel until revenue starts dropping.
What should I look at instead of Meta ROAS to gauge business health?
3 metrics: (1) New customer acquisition rate: is the proportion of new vs returning customers stable or declining? (2) Marketing Efficiency Ratio (MER), total revenue divided by all marketing spend. (3) Contribution margin per order after all costs including returns and fulfilment. These 3 together tell you whether the business is profitable and growing.
Can I fix this while still running Meta ads?
Yes. The fix is usually rebalancing the campaign mix, reducing retargeting budget and increasing prospecting, even as platform ROAS drops in the short term. You also need to check whether the product margin supports the current acquisition cost, because sometimes Meta is working fine and the unit economics are the problem.
