Supply first, then demand: building a new marketplace

The first question that confuses you when launching a marketplace: do you start with suppliers or with customers? Most founders try to do both in parallel — both end up half-done. On ATP we decided up front: supply engine is priority #1. Demand follows supply.

The cold-start dilemma

The real problem with a marketplace: no suppliers means no customers, no customers means no incoming suppliers. The textbook answer — 'start small, stay local' — is correct but insufficient, because a marketplace only delivers value at scale. Five transfer companies aren't a marketplace; fifty are, five hundred definitely are.

The answer: a supply engine that runs independently of customer demand. A new operator starts shipping inventory hours after they sign up. The output of the supply engine grows every day — even before any customer demand exists.

What we actually did on ATP

We built a self-serve supplier panel. New transfer-company onboarding used to take 1-2 weeks — manual paperwork, calls, email back-and-forth. With the panel it dropped to hours. The supplier manages their own data; we just verify.

An outreach engine ran in parallel: local transfer companies in 80+ countries. Zero to 370 LIVE suppliers in four months. March 2026 added 170 new suppliers in a single month (2× the prior month). 82 countries, 514 airports, 7,383 routes, 9,641 active tariffs. 953 vehicles, 711 drivers.

When demand finally turned on

Demand showed up at its own pace. March 2026: 182 bookings (2.2× February). Both the supply side and the demand side hit scale in the same month. Onboarding system and outreach engine matured in lockstep.

Two demand headings turned on at once: the SEO surface grew as suppliers came online, and B2B channels (agency, affiliate, API) were already in place to absorb capacity. Today: 85 affiliates, an active agency portal, multi-currency international payments.

Three lessons

First: in a new marketplace, supply is priority #1. No debate — demand follows onboarding. Second: without a self-serve panel, you can't scale. Manual onboarding kills marketplaces. Third: prepare the SEO surface architecturally before the supply lands, so pages open automatically as inventory ships. Retrofitting is brutal.

The pattern repeats across every new marketplace. Going to customers before you know how supply will scale is the fastest way to lose them on empty inventory pages later.