Marketplace
Supply first. Demand in week nine.
How a services marketplace launched two German cities in a quarter by refusing to spend on demand for the first six weeks.
CASE / 08
A services marketplace
Two-sided marketplaces in a new vertical have one reliable failure mode. Nobody shows up on either side until the other side is there. The client had burned through their first three months trying to buy demand in both cities at once. We spent the first six weeks on supply, and said no to paid demand until something specific was true.
GEO
Berlin · Munich
Setup
Supply-first launch
Duration
14 weeks
Shipped
Q3 2024
The demand-first playbook had run out
The demand-first playbook had run out
We started by modelling unit economics against real data, not plan. Paid demand ad spend did lift listings-search volume, and it did create sign-ups. LTV per acquired demand user, with empty supply, was strongly negative. The money spent to acquire them was being burned teaching them that the marketplace did not work yet, and a material share of them would not come back once supply existed. The model we built showed the break-even point was seven weeks out if we paused paid demand entirely.
Hand-signed supply, 120 providers in six weeks
Hand-signed supply, 120 providers in six weeks
This was not a CRM programme. A three-person operations team ran direct outreach, with a prioritised list built from Instagram business accounts, trade-association registries, and one scraped Google Maps dataset we verified by hand. Top thirty providers got in-person onboarding in Berlin and Munich. The remaining ninety went through a guided video call. The conversion rate from first contact to first listing was 31%, which the team hit by calling warm before emailing cold.
What we refused to build
What we refused to build
The product team had a self-serve supply onboarding flow half-built. We paused it for the engagement. Self-serve would have let low-quality listings flood the marketplace before there was enough demand to filter them. It shipped four months later, after the supply bar had been set by hand. That sequencing mattered: self-serve flows inherit the quality norms set in their first six months. Setting the norm manually was the whole point.
Paid demand turned on in week nine
Paid demand turned on in week nine
The threshold we set in week one was supply density: average 4.2 providers per 5km radius in both cities, with at least 70% of providers having completed two or more bookings. Berlin cleared it in week eight. Munich in week ten. Paid demand went live in Berlin with clear CAC targets per postcode, and in Munich two weeks later. First-booking CAC came in 22% below target in Berlin, at target in Munich.
Launch With Us