Meta Ads Campaign: 4.2x ROAS, Up From 1.1x, in 60 Days
A fashion e-commerce brand was spending on Meta Ads with a 1.1x ROAS — barely breaking even. After a full campaign restructure, creative testing and a proper retargeting funnel, ROAS reached 4.2x in 60 days and revenue from paid social increased 180%.
A fashion e-commerce brand was running Meta Ads at a break-even 1.1x ROAS with no audience segmentation or retargeting. A 60-day campaign restructure — Cold/Warm/Hot funnel, Custom & Lookalike Audiences, 12 tested creative variants, and Conversions API — took ROAS to 4.2x, cut cost-per-purchase by 38%, and grew paid social revenue by 180%. Retargeting alone returned 7.8x ROAS.
Where Things Stood at the Start
What We Actually Did
- Rebuilt full campaign structure: Cold Awareness → Warm Engagement → Hot Conversion
- Built Custom Audiences from website visitors, video viewers and past buyers
- Created Lookalike Audiences from top 5% of customer purchase list
- A/B tested 12 creative variants — video, carousel and single image hooks
- Set up Conversions API alongside Pixel for accurate attribution
Why This Order
The funnel restructure came first because a single undifferentiated campaign asks the algorithm to optimize for everyone at once, which is exactly why the starting ROAS was stuck near break-even — cold, warm, and hot audiences convert at very different rates and need different creative and bids. Custom and Lookalike Audiences were built next so the Warm and Cold stages had real, high-quality audience pools to draw from rather than broad interest targeting.
Conversions API was set up alongside the Pixel specifically because browser-side tracking alone increasingly under-reports conversions (iOS privacy changes, ad blockers, cookie restrictions) — without server-side confirmation, Meta's algorithm is optimizing on incomplete data, which quietly caps performance regardless of how good the audiences and creative are. See the retargeting setup guide for the Pixel/CAPI process in detail.
Measurable, Verified Outcomes
| Metric | Before | After |
|---|---|---|
| ROAS | 1.1x | 4.2x |
| Cost per Purchase | Baseline | -38% |
| Retargeting ROAS | — | 7.8x |
| Paid Social Revenue | Baseline | +180% |
Common Questions About Meta Ads ROAS
It depends heavily on margin — a low-margin category needs a higher ROAS to be profitable than a high-margin one. As a rough industry reference point, many e-commerce brands treat 3x-4x as a solid, sustainable target, with anything below 2x usually needing scrutiny on margins and costs. The 1.1x starting point here was effectively break-even once ad spend was accounted for.
Retargeting audiences (past site visitors, video viewers, cart abandoners) have already shown interest, so they convert at a fundamentally higher rate than cold prospecting. The 7.8x retargeting-only figure here reflects that warm-audience advantage — it's not representative of what cold acquisition alone would return, which is why the overall blended ROAS (4.2x) sits well below it.
Browser-side Pixel tracking increasingly under-reports conversions due to iOS privacy changes, ad blockers, and cookie restrictions. Conversions API sends conversion events directly from the server, filling in gaps the Pixel misses on its own — without it, Meta's optimization algorithm is working from incomplete data, which caps performance regardless of audience or creative quality.
There's no universal number, but testing across formats (video, carousel, single image) and hooks — as the 12 variants here did — matters more than testing many near-identical versions of the same creative. The goal is covering genuinely different angles, since Meta's algorithm needs meaningfully different options to find what resonates with each audience segment.
Key Takeaways
- Segment by funnel stage (Cold/Warm/Hot) before scaling spend — one audience for everyone caps performance.
- Retargeting warm traffic consistently outperforms cold prospecting; don't skip it even on a tight budget.
- Set up Conversions API alongside the Pixel — browser-only tracking under-reports in the current privacy landscape.
- Test genuinely different creative angles, not minor variations of the same concept.
Want results like these?
SEO, Meta Ads, or both — let's build your next case study together.