Section 1 · The Problem
Audrey sells professional massage tables and aesthetic equipment to clinic owners, aesthetic professionals, and fitness operators. Established brand, real product, real demand. When they came to Loocro in May 2021, they had one channel, one problem, and zero predictability.
Google Ads was running. Results were "okay," but every month felt like a gamble. Revenue fluctuated with no clear pattern. No second channel. No systematic approach to audience building. No way to forecast what the next 30 days would look like.
The founder, Guilherme Aiup, wasn't burned out on paid traffic. He'd seen enough to know it could work. He just needed it to stop being a coin toss every Monday morning.
Section 2 · The Diagnosis
Three things stood out when we audited the account.
01The "okay" wasn't a Google problem. It was a system problem.
Single-channel dependency is fragility. If Google's algorithm shifted, if CPCs spiked, if a key audience saturated, there was no second engine to pick up the slack. Audrey wasn't underperforming on Google. Audrey was overexposed to a single point of failure.
02Meta was a wide-open opportunity.
Guilherme had never run a campaign outside Google. Zero pixel history. Zero audience signal. From a platform perspective, that looks like a problem. From a strategic perspective, it's a green field: a buyer profile (clinic owners, aesthetic professionals, fitness operators) that fits Meta's targeting model almost too well, never exposed to a single Audrey impression.
03The reporting layer was blended.
The monthly ROAS number that Audrey was reading was a single average across all campaigns. That number couldn't tell anyone which campaign was carrying the account and which was bleeding. The insight to scale a winner or cut a loser lived one layer beneath the report. We needed to pull it up.
The diagnosis wasn't "add Meta." It was "build a system where channel diversification, segmented ROAS reading, and weekly reallocation reinforce each other."
Section 3 · What Changed
Two moves in month one.
First: We rebuilt the Google Ads account from scratch.
New campaign structure, tighter audience signals, sharper bidding logic. The account that was producing "okay" results closed May 2021 with a 10.19x ROAS.
Second: We launched Meta cold.
Zero historical data, zero pixel events, zero audiences. We built it all from creative one. By end of month, Meta closed at a 7.87x blended ROAS, with the Mother's Day campaign hitting 31.19x on its own.
That was month one.
Over the next five years, both channels scaled together. Google captured intent (people searching for specific equipment), Meta built reach (clinic owners and aesthetic professionals who matched the buyer profile but hadn't searched yet). We tightened the feedback loop between them quarter after quarter: which Meta audiences fed Google's branded search, which Google keywords signaled high LTV cohorts worth retargeting on Meta.
Reporting moved to segmented ROAS by channel, by campaign, by audience, by creative. Blended ROAS stayed on the monthly summary as the headline. Behind it, the segmented numbers drove the weekly reallocation.
The compounding came from the discipline of reading the numbers at the right granularity, not from a single big optimization.
Section 4 · The Metrics
| Year | Blended ROAS |
|---|
| 2021 (year 1) | 7.0x |
| 2022 | 7.0x |
| 2023 | 9.5x |
| 2024 | 8.9x 1 |
| 2025 | 11.3x |
1 In 2024, Audrey's main supplier had a factory fire. Stock shortages set in with no predictable deadline for new product arrivals. A full supplier transition was underway, with delays in imported equipment. Budget pulled back significantly. ROAS held above 8x throughout the year.
| Metric | Value |
|---|
| Peak Meta ROAS (Aug 2025) | 16.37x |
| Peak Google ROAS (Aug 2023) | 15.70x |
| Mother's Day Meta campaign (month 1, May 2021) | 31.19x |
| Total attributed revenue | R$20M+ (~$4M USD) |
| Engagement | 5 years and counting |
Read the table this way: ROAS improved every single year, with one dip in 2024 directly caused by a supply-chain crisis outside our control. Even that dip held above 8x. Year 1 blended: 7x. Year 5 blended: 11.3x. The system compounds because the inputs that feed it (segmented reading, weekly reallocation, two-channel resilience) compound.
Section 5 · The Business Impact
What changed for Audrey isn't only in the ROAS table. It's in what the founder gets to do with his month.
Guilherme stopped opening the ad manager on Monday morning to figure out if the week would be good or bad. He started planning inventory, supplier relationships, new product lines, retail expansion. The paid media stopped being the thing the business depended on him to fix.
When the 2024 supplier fire hit, that compounded too. A business with a fragile growth engine would have collapsed under stock shortages with no clear replenishment timeline. Audrey absorbed the shock, pulled spend back, and the system held. The founder spent 2024 fixing supply, not fixing ads.
The deepest result of a five-year paid-media partnership isn't the ROAS curve. It's the operator headroom the curve buys.
