Staarm started as a personal frustration and a car sitting in a driveway. In June 2021, my co-founder and I launched a peer-to-peer car rental marketplace in Nigeria — no deck, no roadmap. We rented our first car manually, doing KYC in person and handling every handover ourselves. Those first months without a product turned out to be the most valuable thing we ever built.
The problem
Renting a car in Nigeria was slow, documentation-heavy, and limited in selection. For what should have been a simple transaction, the process consistently got in the way. If the available options didn't suit you, there wasn't much else.
Building a two-sided marketplace meant solving two fundamentally different problems simultaneously. Demand was never the constraint — supply was. Getting enough hosts onboarded with enough cars to meet renter demand shaped nearly every early product and business decision we made.
01 / 03
Browse without a login wall
A user can open the app, search by area, and compare available cars by price or type — all before being asked to create an account. Only at the moment of reservation does the platform ask anything of them. Every barrier removed before that moment of commitment was friction that wasn't protecting anything.
Area-based search — available cars visible before login
02 / 03
Streamlined renter flow
Pick a car, reserve it, chat directly with the host, pay, and go. We stripped the booking flow to its minimum viable steps. KYC asked for NIN, driver's licence, and full legal name — enough to confirm identity without feeling like a bank application. Post-launch feedback surfaced a segment of renters who didn't want to drive at all, so we added a driver-included option alongside self-drive.
Booking flow — reservation, host chat, payment, and self-drive vs. driver toggle
03 / 03
Host listing and inventory tools
Hosts could photograph their car, list it, set their own pricing, and manage their availability calendar — a lightweight inventory tool built for someone with a car and free time, not a fleet manager. We also built a host sponsorship program, covering parts of onboarding costs for hosts who needed support getting listing-ready, directly accelerating supply growth.
Host listing flow — photo upload, pricing, and availability calendar
Design decisions
Run the business manually before building the product
For the first two to three months, Staarm had no app. Bookings came through direct conversation, KYC was done in person, and every handover was managed by us. That manual period meant we understood what made hosts comfortable listing a car with a stranger, what made renters hesitate, and what the real operational friction was — before we had a product to hide it behind.
No login wall before browse
Every other rental product asked you to create an account before you could see what was available. We removed that entirely. A user who can see cars they want has a reason to sign up. A user who hits a login screen first may never find out.
Solve supply asymmetry directly
In a two-sided marketplace, equal investment in both sides is a mistake when one side is the binding constraint. Demand was healthy. Supply was the bottleneck. We built the host sponsorship program specifically to remove the cost barrier between 'I have a car' and 'my car is earning money' — treating host acquisition as a product problem, not just a sales one.
Outcomes
Reflection
Staarm taught me that you don't truly understand a market until you've operated inside it manually, without a product to hide the friction. Those first months doing bookings by hand weren't a delay before the real work — they were the real work. Every assumption we would have baked into the app got tested against actual human behaviour first. I also learned how much a two-sided marketplace lives or dies on asymmetry. Renters and hosts were never the same problem wearing two different UIs — they had different motivations, different anxieties, and different definitions of a good experience. Designing for both meant resisting the urge to build one generic flow and call it done.