I built a two-sided marketplace and never once asked anyone to use it
I'm 75, I've been writing software since the 1970s, and I just spent three months building a marketplace that got zero customers. Posting the numbers because I had to say the diagnosis out loud before I believed it.
Brief Work (briefwork.io) — short-burst expert gigs. One deliverable, under $5K, under two weeks. Clients pay $50 to post a brief, experts apply free, no cover-letter theater. The bet was that charging the client to post, rather than charging workers to apply like Upwork does, was the honest version of this business.
Built solo: Next.js, Supabase, Stripe live mode, auth, RLS, messaging, direct hire, transactional email. Real payments, end to end. Went live in May.
Three months later, from the production database:
Briefs posted: 4 — all four by me
Applications received: 3
Registered users: 5, two of which are mine and one is junk
Revenue from strangers: $0
One real transaction ever: I paid a freelancer $50 for a GTM plan. He did good work. I was the client.
What I got wrong. I assumed both sides were cold and I'd have to seed supply first — that's the standard advice, and it's what every marketplace article tells you. It was wrong for my situation. Three freelancers found the site and applied with zero supply-side marketing, and the one I hired delivered. Supply seeded itself.
Demand never appeared. And the honest reason isn't mysterious: nobody ever asked for it. In three months, client acquisition was a launch post, a couple of social posts, and an email to three friends. No outbound. Not one human being was ever personally asked to post a brief.
I kept building instead of selling, because building is the part I'm good at and selling is the part where people say no.
The actual lesson, which cost me three months: a two-sided marketplace is a distribution business wearing a product costume. The product is the cheap part — I proved that by building a working one solo in weeks. Manufacturing demand needs at least one of: an existing audience, money for ads, a network to mine, or appetite for sustained cold outreach. I had none of the four, and the fourth I didn't want.
That's not a product problem you can engineer your way out of. No feature fixes it. I could have built escrow, ratings, and dispute resolution and still had zero customers, just later and more tired.
If you're about to build a marketplace, the question that would have saved me the three months isn't "can I build this" — it's "name the first ten people who will pay, and what makes you able to reach them." If you can't finish that sentence, the build is the expensive way to find out.
The aside: it's for sale. Everything works and it's live. I'd rather it went to someone with an audience and no product — the exact inverse of me — than sit here as a monument. It's listed on Acquire and Little Exits, or just reply here and ask me anything.
Happy to answer technical questions about the build too. I won't oversell it — the numbers above are the numbers.
That's a bold approach given how much marketplace advice pushes the opposite (manually recruiting both sides early). What actually got the first users in without you asking, was it organic discovery, or did something else do the work for you?
The honest numbers are the useful part — thanks for posting them. Since you're running live Stripe from Brazil with buyers mostly abroad, I'm curious how the settlement side has held up: are payouts and FX the smooth part, or the part you'd redo? I've been mapping payment setups for indie founders selling cross-border and happy to share notes if useful.
Straight answer first: I can't tell you how settlement held up at volume, because there wasn't any volume. The posting fees that cleared Stripe were mostly my own money going into my own account, and the one real transaction — paying a freelancer for the GTM plan — settled off-platform via Wise. So treat anything I say as research, not operating experience.
Two things I did learn the hard way, both of which cost me time:
Sharetribe doesn't support payouts to Brazil or India. I'd been evaluating it as a shortcut for the whole payments/escrow chunk and got that confirmed directly by their team. Brazil kills it for me as the recipient; India mattered because that's where most of their build-partner agencies are. If you're mapping options for cross-border founders, that one's worth having on the list — it isn't obvious from their marketing.
Stripe Connect does support Brazil, but wants a BRL bank account matched to a CPF or CNPJ. Workable, just not the frictionless thing people assume when they picture "Stripe handles payouts."
The reason none of this got stress-tested is the point of the post: I built the settlement plumbing before I had anyone to settle with. Would genuinely like your notes — and if you want the Sharetribe correspondence for your own map, happy to send it.
Great post. I have spent the last 10 years building a 2 sided marketplace.
We've managed to get to profitability, but it's a tough and slow road.
Ours has been complicated by 2 additional factors: (a) it's a hyper local marketplace - ie supply and demand have to be in the same place. So building liquidity pools in different geographies is an added challenge; and (b) the "buy side" are businesses and the sell side are consumers. You want that the other way round so you can stack up sell side businesses and then go live big bang to consumers who make buying decisions faster.
The funny thing is - if I knew what I know now, I'd never had started this business - it's not the big growth business I thought it could be. Yet, 10 years in, we're profitable and have cracked lots of the hard problems. So, I'm not sure what that tells you about marketplaces? Maybe, build them if you can, but don't bet the house you'll pull it off.
This is the most useful comment I've had, and the inverted-sides point is one I hadn't framed that way.
That maps onto my own numbers oddly well. Supply seeded itself — three freelancers found the site and applied with zero recruiting, and the one I hired did good work. Demand never showed up once. So the side that stacks passively was free for me and the side that had to be manufactured was the whole business. You're describing the same asymmetry with the sides swapped, and paying for it in geography instead.
On your closing question, my honest read is that it isn't really about marketplaces. Ten years and profitable, having cracked the hard problems, and you'd still not start again — that's an opportunity-cost verdict, not a failure one. The business works; you're saying the decade could have bought more elsewhere. That's a much harder thing to know in advance than "will this work."
The part I'd defend about my own version is only the clock. I found out in three months instead of ten years. I don't think I was smarter — I just had less at stake and quit earlier, which is easier when nobody's depending on it.
"Build them if you can, but don't bet the house" seems about right.
Well, there's lots of things I wouldn't do again if I knew how hard they were and they're all the best things I've ever done. So I don't know about the opportunity cost verdict from a personal perspective. Economically of course, you're right - investment in this product/team/market/timing would have been better placed elsewhere.
You're also right about this being about distribution. My job now is to get the platform into the hands of someone who can solve that problem better - if they can, we've proven the business works and is profitable.
The issue I have is liquidity for businesses at our scale is pretty average so getting a deal away that will get shareholder approval won't be easy.
The honesty in the numbers makes this unusually interesting. “I kept building instead of selling” is probably a familiar story for a lot of technical founders.
Thanks. The part I'd add is that building has a feedback loop and selling doesn't. Code either works or it doesn't and you know inside a minute. Outreach gives you silence, and silence reads identically whether you're wrong about the market or just early. So you drift toward the activity that answers you. Knowing that mechanism didn't stop me doing it for three months.
Also — are you the Aryan who wrote the GTM plan for Brief Work back in June? If so, you're owed a note: the plan was fine. I never ran it. Three months on, the outreach section was still untouched, which is the actual story of the post rather than anything you got wrong.
Fair enough — and I appreciate you clarifying that. The execution gap is actually the more interesting part of the story.