Cost per acquisition has climbed for 3 months. The agency says the creative is tired, so you test 12 new ads. Nothing moves. Someone says the landing page is the problem, so you rebuild it. Nothing moves. Someone says the targeting is wrong, so you restructure. Nothing moves.
Here is the account checked in 3 places (example, redrawn, not a screenshot):
The margin per order is £28 (£100 order × 28% contribution margin). Each new customer costs £35. Every customer loses £7. No ad test, page rebuild or restructure fixes that.
Ads fail in 1 of 3 places: creative, funnel or economics. Most brands start with creative because it is visible. Checking all 3, in order, takes 30 minutes.
Layer 1: creative
Are enough people clicking? Check click-through rate (CTR) on prospecting campaigns.
| CTR | What it means |
|---|---|
| Above 1.5% | Creative works. Move to layer 2 |
| 0.8 to 1.5% | Borderline. Check layer 2 before blaming creative |
| Below 0.8%, consistently | Creative is the limit. The hook, audience match or format is off |
Before you decide creative has failed, check 2 things:
- Budget. Meta’s guide asks for about 50 results a week to leave learning. Some accounts settle earlier, some never do, so judge on the trend, not 1 week.
- Frequency. Above 2.5 over 7 days, the audience has seen the ad too often. Make new concepts, not new versions. See why Meta ads stop working so fast.
Layer 2: funnel
Are enough clickers buying? Pull click-to-purchase conversion rate for the last 30 days from Shopify or your analytics, for Meta traffic only. Do not use the blended rate.
| Conversion rate | What it means |
|---|---|
| Above 2.5% | Funnel works. Move to layer 3 |
| 1.5 to 2.5% | Acceptable. The best place to improve before scaling |
| Below 1.5% | Funnel is the limit. More spend is wasted on a page that cannot convert it |
The usual causes, most common first:
- Product page friction. Reviews are thin or vague. The first screen does not continue the ad’s promise. The objection is not answered before the add-to-cart button.
- Checkout drop-off. Surprise shipping costs, forced account creation, too few payment options. A checkout completion rate below 60% (checkout started to purchase) is a leak no creative change will fix.
- Slow mobile pages. In the accounts we audit, most paid traffic lands on a phone. Google rates a page good when its main content loads within 2.5 seconds (Core Web Vitals report). Slower pages lose visitors before they see the product. Desktop tests hide this.
Layer 3: economics
Does the model pay for the acquisition cost? Run 3 steps:
- Contribution margin per order = average order value × (1 − COGS% − fulfilment% − returns% − payment fees%)
- New customer CAC = total ad spend ÷ new customers (not total orders)
- Compare. If margin per order is above CAC, scaling is viable. If CAC is above margin per order, every customer loses money.
With the example above: £100 × 28% = £28 margin, against a £35 CAC. Scaling makes it worse faster. The fixes are margin (price, cost of goods, returns) or CAC (repeat purchase, channel mix, organic growth). See breakeven ROAS for the full calculation. Try your own numbers:
Try it with your numbers
Your break-even ROAS
Where each order goes
- Product
- Shipping
- Fees
- Returns
- Left for marketing
Left for marketing: £ an order
Aim for a target ROAS of ×, 20 to 30% above the floor.
Add your fixed costs (monthly view)
Staff, rent and software don't rise with each sale, so they stay out of the floor. To see the MER you need to cover them too in a month, add them here.
To cover fixed costs as well, this month needs an MER of ×.
The 30-minute check
- 5 minutes: pull 30-day CTR by campaign type from Meta. Flag any prospecting campaign below 0.8%.
- 10 minutes: pull 30-day conversion rate for Meta traffic from Shopify, plus checkout completion. Flag conversion below 1.5% or checkout completion below 60%.
- 15 minutes: work out margin per order and new customer CAC for the same 30 days. Compare them.
Start with the first layer that fails. Only that layer.
What to do next
- Creative fails: test different concepts (not variations) for different buyers. Give each concept 2 to 4 weeks before you judge it.
- Funnel fails: run a product page CRO pass before raising spend. Check the first screen, main image, proof and checkout. See checkout conversion optimisation.
- Economics fails: do not scale. Raise price, renegotiate cost of goods, cut returns, add a higher-margin product or build repeat purchase.
In the accounts we audit, most brands sit in layer 2 while most agencies prescribe layer 1.
Related reading
- Product page CRO: 12 changes that lift conversion: fixing the funnel layer
- Checkout conversion optimisation: the second most common funnel leak
- Breakeven ROAS for ecommerce: the economics layer in detail
Layer 3 is a number you can work out today. Check yours.
If the data itself is in doubt, see our tracking and reporting service, or book a free call and we will run this diagnosis on your real account data.
Quick answers
Why are my Meta ads not working even though I have tested lots of creative?
Creative is only 1 of 3 possible causes. If you have tested creative widely and cost per acquisition is still above target, the problem is likely the funnel (page and checkout conversion) or the economics (your contribution margin cannot support the acquisition cost). Many brands iterate creative for months when the real limit is a 1.2% page conversion rate or a 28% margin.
How do I know if my Meta ads problem is creative, funnel or economics?
Work through 3 layers in order. First check click-through rate: above 1.5% on prospecting means creative is working. Then check click-to-purchase conversion: below 1.5% means the funnel is the limit. If both are healthy and you still lose money, run a contribution margin calculation. Fix the layer that fails before touching the others.
What is a good click-through rate for Meta ads in ecommerce?
For cold prospecting, we treat above 1.5% as creative working and 0.8 to 1.5% as borderline. Below 0.8% consistently points to a creative problem. Retargeting runs higher, at 2 to 5%, because the audience already knows you.
What conversion rate should I expect from a Meta click to a purchase?
For cold traffic, 1.5 to 3% is typical in most ecommerce categories. Above 2.5% is strong. Below 1.5% points to a funnel problem. Warm retargeting traffic should convert at 3 to 6%.
