A Shopify supplement brand had 4.1 ROAS in Meta and declining revenue month over month. The founder was confused — the platform metrics looked strong. When we ran three diagnostic reports that don't appear in standard weekly updates, the problem was visible in 18 minutes: their high-ROAS campaigns were acquiring existing customers (who would have repurchased anyway), while new customer acquisition had quietly collapsed to 1.6% of total orders. Meta's dashboard reported 4.1 ROAS on repeat revenue. New customer growth was nearly zero.
Why account-level ROAS can look strong while revenue falls
ROAS measures revenue attributed to ads divided by ad spend. It doesn't measure: whether those customers are new or returning, whether the revenue is incrementally driven by the ads or would have happened anyway, or whether the attributed revenue is being double-counted across channels.
An account in decline often hides it in aggregate metrics. A campaign retargeting existing customers at 8.0 ROAS averages with a new customer prospecting campaign at 2.1 ROAS, producing a blended 4.1 that looks like a healthy account. But if the retargeting campaign is the majority of spend and is capturing people who were going to repurchase anyway — and if the prospecting campaign has quietly stopped acquiring new customers — the brand is paying to attribute existing repeat revenue to Meta while the top of funnel is empty.
Two to three months later: the cohort of existing customers who were being retargeted starts to shrink (buyers eventually churn). The 8.0 ROAS retargeting performance declines. The prospecting campaign, which was never fixed, can't replace those customers fast enough. Revenue falls. ROAS eventually falls. But by then, the business has been in decline for months and the metrics only now show it.
The 3 diagnostic reports that reveal the actual picture
Report 1: New customer rate from paid traffic (Shopify).
Pull Shopify orders for last 30 days. Filter by UTM source = facebook/instagram and UTM medium = cpc (or paid social). For each order, check whether the customer is new (first order) or returning (prior order exists in their history).
The target: 60–80% of paid traffic acquisitions should be new customers for a growing brand. If new customer rate is below 40%, paid campaigns are predominantly capturing existing buyers — retargeting with little or no new customer acquisition.
For the supplement brand: 1.6% new customer rate from Meta. They were spending $22,000/month on Meta primarily to recover credit for 98.4% repeat sales that would have happened via email or organic.
Report 2: Prospecting campaign performance isolated from retargeting.
Most accounts blend prospecting and retargeting into a single ROAS number. Separate them.
In Meta Ads Manager, filter by audience type. Find campaigns/ad sets targeting cold audiences (Lookalike, Broad, Interest) vs. warm audiences (Website Visitors, Customer Lists, Engagement). Calculate ROAS separately for each bucket.
For the supplement brand: prospecting campaigns (targeting Lookalike 1-5%) were running at 1.4 ROAS — below their 2.3 break-even. Retargeting was at 8.2 ROAS. The blend of 4.1 had completely obscured that prospecting was unprofitable and had been for 11 weeks.
Report 3: Creative frequency by audience segment.
Pull frequency data for each ad set (Meta Ads Manager → Columns → add Frequency). Check frequency separately for prospecting and retargeting audiences.
The supplement brand's prospecting Lookalike ad set: 8.9 frequency. The creative had been running for 14 weeks. Buyers in the Lookalike audience had seen it 8–9 times on average. The algorithm kept spending (strong historical CTR) but the audience had exhausted. New customer acquisition fell because the prospecting creative had fatigued and no new creative had been tested for 11 weeks.
What the supplement brand's account revealed — and what changed
The 18-minute diagnosis:
- Report 1: 1.6% new customer rate. Finding: prospecting isn't working.
- Report 2: Prospecting ROAS 1.4 (below break-even). Finding: confirmed — prospecting unprofitable for months.
- Report 3: Prospecting creative frequency 8.9. Finding: cause identified — creative fatigue, no testing in 11 weeks.
Changes made in the following 2 weeks:
- Prospecting creative refreshed: paused 8.9 frequency asset, launched 3 new creatives at $40/day each. Two cleared the thumb-stop threshold within 5 days.
- Prospecting budget isolated: retargeting reduced to 20% of spend (it wasn't driving growth, just capturing repeat purchases). Prospecting increased to 80%.
- New customer rate target set: 65% of paid acquisitions should be first-time buyers, measured weekly in Shopify.
Six weeks later: new customer rate from Meta 1.6% → 58%. Prospecting ROAS 1.4 → 3.2. Account-level ROAS dropped from 4.1 to 3.4 (retargeting contribution reduced), but revenue grew because new customers were entering the funnel again.
Run the diagnosis on your account today
Three reports, 20 minutes:
- Open Shopify, filter last 30 days orders by paid traffic UTM. Check: what percentage are first-time buyers? If below 50%, prospecting may not be working.
- Open Meta Ads Manager. Calculate ROAS separately for prospecting vs. retargeting ad sets. If prospecting is below your break-even ROAS, that's the finding — not the blended number.
- Check frequency on prospecting ad sets. If above 6.0 on any warm lookalike audience, creative fatigue is likely the cause of prospecting underperformance.
The account audit framework covers waste removal. This diagnosis covers structural failure — the difference between an account that's wasting spend and an account that's quietly stopping new customer growth.