Marketplaces Get Built for the 90th-Percentile Seller. The Real Wins Live in the 50th.

Marketplaces Get Built for the 90th-Percentile Seller. The Real Wins Live in the 50th.

When you build partner-facing product at a marketplace, the loudest sellers — the ones with brands, ops teams, and dashboards on three monitors — are the ones who shape the roadmap. They have the time to file detailed bug reports, the bandwidth to attend partner roundtables, and the relationships to push features to the top of the queue. So the platform gets optimised for them.

The trouble is they're not the median seller. The median seller has a phone, a warehouse the size of a living room, and a single person doing fulfilment, returns, and customer service. Most of the levers that would actually move marketplace GMV live with this seller — and most platforms ignore them.

The default workflow is the actual product

A 90th-percentile seller will tolerate a clunky workflow because they've built around it — automated scripts, internal tooling, a person whose job is to massage exports. A 50th-percentile seller cannot.

The lesson I keep relearning: the default workflow on the platform *is* the product for most partners. If creating a listing takes seven steps, four toggles, and three tabs, half of your seller base is silently doing it less than they could. Not because they don't want to grow — because the cognitive cost is too high to do it twice in a day.

Onboarding is where churn is born

Every marketplace I've worked on has a partner onboarding flow that's been "good enough" for years. New partners trickle through, some thrive, most don't. The instinct is to look at the thriving ones and study what they did. The better question is to look at the partners who joined three months ago and never crossed five orders — and ask what we taught them, what we didn't, and where they got stuck.

Most of those answers don't live in product. They live in the gap between what onboarding implied and what the seller actually had to do on day one. Closing that gap is rarely glamorous. It's often the single highest-leverage thing on the roadmap.

The data hides where research surfaces

I came into product from a Teach For India fellowship, which means I learned to trust observation before I trusted dashboards. The classroom version: a kid who looks engaged might be deeply lost; a kid who looks disengaged might be the one quietly tracking. The marketplace version: a seller with high session counts might be confused, not active; a seller with low session counts might have a third-party integration that you don't even know about.

When I'm scoping a partner-experience feature, I now do at least five seller calls before any spec gets written — and at least two of them have to be from the bottom half of the GMV distribution. The pattern that surfaces from those calls almost never matches what the data alone is telling me.

What I'd tell a younger PM building marketplace product

- Define your seller personas by *capability*, not by GMV. A high-GMV brand and a high-GMV reseller behave nothing alike. - Write your PRDs assuming the seller has a phone, not a laptop. If your flow doesn't work on a phone, it doesn't work for half your base. - Treat onboarding as a product team's permanent backlog, not a one-time project. The lifecycle moves; so should it.

The brand of marketplace product that lasts isn't the one that wins the sharpest seller. It's the one that helps the median seller ship one more order than they did last week. Compounded across a base, that's where the real platform leverage lives.