1.8x to 4.2x ROASon Meta Ads, While SpendMore Than Doubled.
A £1.4M DTC brand had been running its own Meta Ads on the same campaign structure for over a year. It was profitable, but almost every purchase came from retargeting people who already knew the brand — and a share of the sales it did make were never reaching Meta at all. We fixed the tracking, rebuilt the account by audience temperature, and only then scaled the spend.
Free, no obligation — we reply within 48 hours.
Client DTC ecommerce · £1.4M annual revenue · Meta Ads run in-house
Profitable, and Selling Almost Entirely to People Who Already Knew the Brand
The brand is a DTC ecommerce business generating £1.4M in annual revenue, running its Meta Ads in-house on a campaign structure that had not changed in over a year. Monthly ad spend was £28,000 and the account was profitable at 1.8x blended ROAS.
Growth had stopped, and the reason was in where the return was coming from. Almost all of it came from retargeting warm audiences — people who had already visited the site — rather than from acquiring anyone new. An account in that shape is steadily reselling to a pool it has stopped refilling, which is a ceiling rather than a plateau: it holds until the pool runs down.
Three Reasons the Account Could Not Take More Budget
The audit looked at what the account could measure before it looked at what it was bidding. All three findings were reasons that adding spend would have made performance worse rather than better.
Nothing separated cold, warm and retargeting spend
A handful of broad campaigns ran side by side, so there was no way to tell which budget was buying genuinely new customers and which was recycling the same warm audience. Every performance read was an average across both.
The same creative had been carrying the account for months
Three or four ads were doing all the work with no structured testing cadence behind them. Frequency climbed, CPMs inflated with it, and there was nothing in the pipeline to rotate in once a creative finally burned out.
A share of purchases never reached Meta at all
The Conversions API had not been fully implemented, so real sales were going unreported. The algorithm was optimising delivery against an incomplete picture of what was working — a problem that compounds rather than dilutes as more budget goes through the account.
Twelve Weeks to Rebuild, Then Scale
Tracking first, structure second, spend last. Scaling an account that cannot measure itself just buys more of whatever it was already getting wrong.
- Weeks 1–3
Tracking and account restructure
We implemented the Conversions API alongside the existing pixel to close the reporting gap, then cleaned up event deduplication so the same purchase was not counted twice from both sources. The campaign structure was rebuilt into distinct cold prospecting, lookalike and retargeting tiers, so budget could be allocated and judged by function instead of pooled into one undifferentiated number.
signal restored first - Weeks 4–7
Creative testing pipeline
A weekly testing cadence replaced ad-hoc creative changes. New concepts — UGC, static and short-form video — ran through a dedicated testing campaign, and only proven winners were promoted into the scaling campaigns. That gave the account a standing supply of fresh creative instead of a dependence on the handful of ads already fatiguing.
winners promoted, not guessed - Weeks 8–12
Controlled scaling
With clean signal and a creative pipeline in place, budgets were scaled progressively on winning ad sets and pushed into new lookalike and interest-based cold audiences. Dynamic product ads took over retargeting, so abandoned carts and browsers were recovered without spending twice against prospecting campaigns chasing the same people.
spend moved last, deliberately
Six Months On, Measured Against the Starting Baseline
| Metric | Before | After (6 months) |
|---|---|---|
| Monthly ad spend | £28,000 | £61,000 |
| Monthly revenue from Meta Ads | £50,400 | £256,200 |
| Blended ROAS | 1.8x | 4.2x |
| Cost per purchase | £34.20 | £18.50 |
| New customers as a share of purchases | 22% | 58% |
The rebuild took twelve weeks; the figures above are measured six months from the start. Spend more than doubled across that period and the return on it rose rather than fell, while the cost of each purchase dropped 46% — the opposite of what usually happens to an account as budgets scale. The account also stopped leaning on retargeting: new customers went from roughly a fifth of purchases to well over half.
The Foundation Was Fixed Before Any Budget Moved
None of this was a bidding trick. Accurate tracking gave the algorithm the signal it needed to optimise against real sales, a cold/warm/retargeting structure made performance readable by audience type instead of as one blended average, and a standing creative cadence stopped the fatigue that had stalled growth in the first place. The scaling worked because it came last.
Is Your Meta Account Scaling, or Just Spending More?
Accounts that stall at a decent ROAS usually have the same three things wrong underneath: incomplete tracking, no separation between cold and warm spend, and creative that has been running too long. A free audit is enough to tell you which of them applies to yours.