The Marketplace Trap
Your supply is not your customer. But it's the one that leaves first.
Two-sided markets lie to you twice.
The first lie is about which side is harder to acquire. Most marketplace founders think demand is the scarce resource because demand is what creates revenue. Wrong, usually. Demand is abundant in almost every category where a marketplace could work — there are buyers for almost everything. Supply is harder to get, harder to keep, and harder to replace. But it's not the one writing the check, so it's not the one who gets designed for.
The second lie is about who the customer actually is. Revenue comes from one side. Value is produced by the other. The mistake is treating 'revenue source' and 'customer' as synonyms. They're not. Both sides of a marketplace are your customers. Only one of them pays.
Watch what happens when a marketplace optimizes relentlessly for its demand side. A recruitment platform I studied closely ran this playbook. Every sprint, the feature roadmap was shaped by recruiter feedback — they were the paying side, they had seats, they had contracts. The candidate experience got worse every quarter: harder to apply, less feedback on rejections, slower communication loops. It felt like the right tradeoff. Revenue was up.
Then applications slowed. Not from all candidates — from the best ones. High-signal candidates, the ones with options, started skipping the platform. They had better options: other platforms with cleaner experiences, direct outreach, referrals. The recruiter-facing metrics looked fine until the downstream quality metric — hire rate on the roles recruiters actually cared about — started moving. By the time the platform noticed, the candidate supply problem was already structural.
They had built for their revenue source and destroyed their product. The revenue source is a lagging indicator on a marketplace. Supply quality is the leading one.
Supply doesn't pay you. But supply leaves first.
The design question that clarifies everything is this: which side of your marketplace is harder to replace?
On a craft marketplace, the answer is unambiguous. Buyers are replaceable — there's a large, growing market of people who want authentic goods, handcrafted products, products with provenance. Lose a batch of buyers, run a promotion, they come back. Specialized suppliers are not replaceable. Each one has a specific craft tradition, a specific design vocabulary, a specific production cadence tied to the materials and skills they have. Lose a master craftsperson and you lose the specific product that made that category of buyer come to your platform.
Everything downstream follows from this. Your product roadmap should serve whoever is harder to replace. Your operations model should protect whoever is harder to replace. Your pricing should not alienate whoever is harder to replace. For a craft marketplace, that means supplier-facing features — transparent pricing, payment reliability, low friction on listing — are not nice-to-haves. They are the product. Buyer-facing features are how you scale the market. Supplier-facing features are how you keep what you have.
Most marketplaces get this exactly backwards because the harder-to-replace side is often the one that doesn't pay, doesn't complain loudly, and doesn't show up in a sales call.
Before greenlighting any feature for a marketplace roadmap, run one test: does this make the harder-to-replace side more likely to stay?
Most roadmaps fail this test badly. Feature requests come predominantly from the demand side. They're louder, easier to reach, and they have contracts with leverage. The supply side stays quiet — not because they're satisfied, but because they don't have the same vocabulary or the same access. They signal dissatisfaction by leaving. By the time supply is loud about a problem, you've usually already lost the best part of them.
This creates a systematic bias in every marketplace roadmap that isn't deliberately corrected for. You fix it by proactively researching supply health — churn rate by supply cohort, supply quality distribution over time, supply-side satisfaction separate from demand-side satisfaction — and weighting those metrics in prioritization decisions even when supply has no seat at the table.
The marketplace trap is not a growth problem or a cold-start problem. It's a definition-of-customer problem. Fix the definition and the rest of the roadmap starts making sense.