Case study · Consumer membership marketplace · Google Ads, Meta and creative
Once they’d seen the ownership we took and the results we drove, handing us the creative was the obvious next step. The numbers have proven them right.
The short answer
A members-only consumer marketplace hired us in August 2024 for Google only. We earned the rest one channel at a time.
- Google, August 2024. We fixed the tracking, then grew activated members from Google from 24 a month to over 80 in our first eleven months.
- Meta media, July 2025. We cut the half of the budget that was buying members at about $9,700 each, and blended cost per member fell 23% in our first quarter.
- Creative, February 2026. They ended their creative agency and handed the ads to us. Within six months our creatives were taking 89% of the Meta budget.
+56%
Activated members a month since we started
−23%
Blended cost per activated member
2.5x
New creatives per dollar of spend
Chapter 1: Google, where we earned the trust
We started in August 2024 with search only. The audit found campaigns optimising on clicks and button presses, with no data on who actually became a member. We set up offline conversion tracking first, so Google could learn from accepted and activated members, not form fills. Then we scaled it.
In eleven months activated members from Google went from 24 a month to 81, and the share of Google registrations that became members nearly doubled, from 1.3% to 2.4%. By mid-2025 Google was producing half of all members on about a tenth of the paid budget.
Chapter 2: Meta, where the money was leaking
Meta was run by another team. In the first half of 2025 its spend more than doubled, to $1.08 million a month, and its cost per registration rose from $33 to $150. Each extra dollar was buying less.
After a year on Google, the client asked us to audit Meta, and we took it over in July 2025. We consolidated 70 campaigns into 7 so the algorithm had enough data to learn in each, then cut the spend that wasn’t turning into members.
In the quarter before we took over, the account spent $1.05 million a month and got 152 activated members at $6,890 each. In our first quarter it spent $512,000 a month and got 97 at $5,294 each. The $538,000 a month we cut had been buying members at roughly $9,700 each, nearly twice what the rest of the budget paid. From February to July 2026, cost per member averaged $5,927, still 14% below the quarter before we took over.
Chapter 3: Creative, replacing the agency
The ads still came from a dedicated creative agency. The work was well made and on brand, but it played it safe: small variations on whatever was already working, built around the same three themes. What the account needed was new angles to reach and convert new audiences, and those weren’t coming.
We didn’t wait to be asked. In September 2025 we shipped nine statics of our own, and one became the top-spending ad in the account within two weeks. In December the client told us they needed far more creative than the agency delivered. In January they ended the agency and handed creative to us.
From February to September 2026 we launched 23 new creatives a month, 20 of them new angles rather than versions of old ones. That’s 5.9 for every $100,000 of spend, against 2.4 in the agency’s final months.
We didn’t switch off the old ads. We ran ours next to them in the same campaigns and let Meta’s algorithm choose. Our creatives took 11% of spend in February and 89% by August, and brought in registrations 20% cheaper than the older ads they pushed out.
What made our creative work
Every concept started from one specific person and a hypothesis, built from the client’s best-performing search terms, community threads and competitor ads. Every batch carried safe concepts and wild ones, and a creative roadmap logged what was tested, why, and what we learned.
The breakthrough was the persona call-out. Generic ads kept reaching the same people, whatever the angle. Naming one specific kind of member in the hook reached people the account had never touched, at around half the account’s cost per registration. And because the same team ran media and creative, a test result became the next brief the same week.
The job isn’t to make each channel look good. It’s to put the next dollar wherever it buys a member.
What the client says
“Lean’s creative team has been a huge unlock for [us]. Their output is consistently strong, but the real value has been their understanding of our business and willingness to challenge our thinking. When our creative needed a refresh, the team brought angles we hadn’t considered and opened up new avenues for growth.”
Questions we get asked
Should the team buying media also make the creative?
When the account spends enough to burn through ads every week, yes. A media team sees a hook fading days before a creative agency hears about it. When the same people own both, the next brief is written from this week’s numbers, not last month’s deck.
Isn’t that marking your own homework?
The judge here wasn’t us. Old and new ads ran side by side in the same campaigns, and Meta’s auction moved the budget to whichever brought in registrations cheaper.
Why measure new creatives per dollar rather than per month?
Because spend changes. The account launched 26 creatives a month in early 2025, but on twice today’s budget. Per $100,000 of spend, that was 3.3. Per dollar is the only fair way to compare.
What is a persona call-out?
Naming one specific kind of person in the hook instead of selling to everyone. It tells the viewer the ad is for them, and helps the algorithm find more people like them.
Client name withheld. Activated members and cost per activated member come from the client’s back-end data, for members activated within 30 days of registering, across Google and Meta. Creative counts and cost per registration by creative come from every ad in the Meta account, November 2024 to September 2026, as reported by Meta; copies across ad sets, languages, formats and versions count as one creative, and a creative is new in the first month it spent $250 or more. USD.
