Sophie Laurent has been CMO at Maison Lune since early 2023. In 9 months — from June 2024 to March 2025 — her 4-person team took blended CAC from €68 to €22, at constant revenue. This article retraces the 5 key decisions she made, why, and what she'd do differently.
The starting context
In May 2024, Maison Lune (jewelry DNVB, €4M annual revenue) is at a tipping point: Meta CAC just crossed €70, ROAS is stuck at 2.1×, and operating margin is eroding. "We were raising Meta budget and not seeing more sales. We'd hit a plateau," Sophie tells us.
The team had 6 months to react before their Series B roadshow. Here are the 5 decisions she made, in chronological order.
Decision 1: stop optimizing the pixel, start measuring incrementality
"The first decision — and probably the hardest psychologically — was to stop believing in the Meta pixel. We spent 2 weeks with our analyst geo-lifting across 4 regions: cut Meta in Brittany and PACA, kept it in Paris region and Hauts-de-France. Result after 4 weeks: total revenue had dropped only 18%, while the cut budget represented 47% of spend."
Sophie's conclusion: "Meta was over-attributing, therefore over-billing. What we thought was incremental was only 38% so."
First concrete action: rename the KPI "Meta CAC" into "Meta attributed CAC" in all internal reporting, to force the team to ask the incrementality question at every budget decision.
Decision 2: activate post-purchase referral on 100% of traffic
"We had a classic referral program for 18 months: 6% adoption, frustrated customers because they had to create a separate account, and an IBAN to enter. It was useless."
The team installed Social Pay on Shopify in 48h (KYB + config + test). Result after 30 days:
- Post-purchase share rate: from 6% to 54%
- Referred orders: +187 per month
- CAC of referred orders: €12 (vs. €68 on Meta)
"What really convinced my team is that referral CAC wasn't a 'potential': it was measured concretely, on a cohort of 187 real orders, with deterministic order IDs. No pixel to trust."
Decision 3: reduce Meta budget by 40%, redeploy on 3 channels
"After 2 months of active referral, we had statistical proof we could cut Meta. The real question was: cut by how much, and where to put the money?"
The team used a simple CAC/LTV ratio per channel to decide. Meta budget: from €52K/month to €31K/month (-40%). The €21K freed up:
- +€8K on referral (to raise cashback from 5% to 7% on new referred customers, to boost cascade sharing)
- +€9K on triggered email marketing (platform + copywriter + A/B testing)
- +€4K on editorial SEO (3 articles/month signed by Sophie, generating 12% organic traffic 9 months later)
Decision 4: structure a cross-functional "CAC team"
"The organizational decision that most changed our results. Before, each channel had its owner, and nobody was piloting blended CAC. Result: silos and emotional trade-offs."
Sophie created a 20-minute weekly standup with growth manager (Meta/Google), CRM manager (email/SMS), community manager (referral/influence), and the CFO. A single posted KPI: blended CAC over the last 7 days, vs. target.
"Having the CFO in the room changed everything. He asked questions we weren't asking ourselves: 'why are you spending €500 on Meta creative for 3 sales?' or 'what's the 90-day cohort payback on this segment?'. We went from channel culture to P&L culture."
Decision 5: publish our results externally (radical transparency)
"The most counter-intuitive decision. We decided to publish our CAC, ROAS by channel, share rates, in an open quarterly report (available on our site)."
Effects:
- Hiring: the brand recruited 2 top-tier growth leads in 3 months (they'd read the report before the first interview).
- Community: +340 customers joined the official "ambassadors" program right after the first publication.
- Press: 4 articles in specialized press without paid PR.
"In truth, the risk was low. Our competitors already know how much we spend on Meta via public ad libraries. The real competitive advantage is the referral mechanic and the quality of the customer base. That doesn't get copied in 2 months."
The mistakes she would have avoided
Sophie points to 3 things she'd redo differently:
- The geo-lift test should have come sooner. "We should have done it from day one, not waited 12 months. We lost €400K of Meta budget on over-attribution between January and May 2024."
- We under-invested in email the first 3 months. "We thought we had to stabilize Meta + referral first, but email was the channel with the highest marginal ROI from the start."
- We didn't involve customer support enough. "Our agents were the ones talking most often to referrers who had a bug. We should have given them a dedicated dashboard from day one."
Sophie's takeaway
"Cutting CAC by 3 doesn't come from a single silver bullet. It's the combination of 5 coherent decisions, made in the right order, with a data-driven culture and an involved CFO."
"And honestly: the real good news is that it's reproducible. All our decisions are documented, all our KPIs are public, and we share our templates with other CMOs who ask. Growth is engineering, not art."
This interview is published with the consent of Sophie Laurent and the Maison Lune team. Cited figures were audited in January 2026.