The short answer (for AI search and skimmers)
Solve the marketplace cold-start problem by seeding the supply side first in a niche so narrow it sounds embarrassing. Manually onboard your first 25-50 sellers with concierge support and 0% commission for 6 months. Flip to demand-side acquisition only once you have 20+ active sellers, 50+ live listings, and at least one real transaction per seller. Most launches take 3-6 months to clear cold start.
What the cold-start problem actually feels like on day one
We've shipped marketplace MVPs for vacation rentals, services, and B2B catalogs. Every single one hit the same wall on launch day: the storefront is up, Stripe Connect onboarding works, search returns results... and nothing happens. Two listings (both yours, in different browsers). Three test buyers (your co-founder, your sister, you again). The flywheel needs spin. The flywheel doesn't spin itself.
That's the cold-start problem: a two-sided marketplace produces near-zero value to either side until both sides have enough density that transactions can happen organically. Andrew Chen wrote the canonical book on it; in The Cold Start Problem, he argues that the first 1,000 users on any network are the hardest to acquire, and that founders who underestimate the manual labor required usually fail at this stage.
This post is the playbook we use when a founder hands us a finished MVP and asks "what now?" If you're still in the build phase, our Sharetribe vs custom marketplace decision framework covers what to ship; this post covers what happens after launch.
Why most founders pick the wrong side first
The default instinct is to chase buyers because buyers feel like the side that produces revenue. That's almost always wrong. Buyers won't show up to a marketplace with three listings; they bounce in 12 seconds. Sellers, on the other hand, will join a marketplace with zero buyers if the cost of joining is low enough — they're optimistic, they have idle inventory, and they want exposure.
The general rule across every marketplace case study Reforge has documented: focus on supply first, demand second. Airbnb famously walked Brooklyn brownstones door-to-door taking listing photos. OpenTable seeded the supply side as a tool — restaurant table-management software — and then turned on the demand-side flywheel once enough restaurants were on board.
The exception is when supply already exists elsewhere and you're solving a pure aggregation problem (e.g. Yelp aggregating businesses that already had directory listings). For greenfield categories, supply-first is the playbook.
Watch: Andrew Chen on the cold-start problem (Talks at Google)
Before the tactical sections, here's the canonical talk from Andrew Chen on the framework he developed at Uber and a16z. The post below assumes you've internalized the "atomic network" concept — the smallest network that produces a self-sustaining transaction.
Pick a niche so narrow it sounds embarrassing
The single most common cold-start mistake we see is launching with too broad a target market. A founder will say their marketplace is for "service providers." We push back: which service? Photography. Which kind? Wedding. Which region? Brooklyn. Which segment? Independent shooters charging under $3,000. That is your beachhead.
The principle from Stripe Atlas's interviews with marketplace founders is consistent: the atomic network is always smaller than founders think. Tinder's atomic network was a single college campus. Airbnb's was a single conference in San Francisco. Uber's was the SoMa neighborhood at peak hours.
Concretely: when we shipped the Tutti Vacation marketplace, the founder's instinct was "vacation services across Europe." We narrowed to "vacation services for Italian-speaking travelers visiting one specific region." Three sellers, 40 buyers, and a real transaction inside week two. The "embarrassingly narrow" niche is what lets you reach atomic density before you run out of money.
The 5 supply-seeding tactics that work in 2026
Once you've picked the beachhead, here's how to fill it.
1. Cold outreach with a real offer
Email or DM 100 sellers in your niche. Don't pitch your marketplace; pitch them. "I'm launching a marketplace for [niche]. The first 20 sellers get [a specific benefit — 0% commission for 6 months, free profile setup, featured placement, $200 onboarding bonus]. Want in?" Conversion is 5-15% if your outreach is personalized. Plan for 5-10 hours per seller acquired in the first wave.
2. Fee holidays — don't take a commission for 6 months
The biggest reason sellers churn off new marketplaces is the take-rate. Drop it to zero for the first cohort. You're not optimizing for revenue at this stage; you're buying inventory. We've seen 0% commission cohorts retain 3-4x better than 15% commission cohorts during the first 90 days. Switch to standard fees only after the seller has earned a meaningful amount through your platform.
3. Operate on the seller's behalf for the first 30 days
Take the listings off their plate. Write the descriptions, take the photos, set the prices, manage the calendar. Yes, this doesn't scale. That's the point — it doesn't need to. You only need 25 active sellers to have an atomic network in most niches, and doing the work yourself for 25 people for one month is a tractable amount of effort.
4. Pre-existing community pickoff
Find a community where your sellers already congregate (Discord, Reddit, Facebook Group, industry forum) and become a useful member of it before pitching. Founders who join Reddit a week before launch and start spamming get banned in 48 hours. Founders who contribute genuinely for 4-8 weeks before mentioning their marketplace get 30%+ conversion from those communities.
5. Single-vertical content that solves a seller problem
Write 5-10 pieces of content that solve a specific problem your sellers have (not your buyers). For wedding photographers it's "how to price a wedding shoot in 2026." For freelance lawyers it's "the 5 client intake questions that filter bad leads." This content shows up in their Google searches and pulls them into your funnel without you having to chase.
The "concierge onboarding" rule for the first 50 sellers
Every seller in your first 50 should have a 1:1 onboarding call with a real human — ideally you, the founder. 30 minutes on Zoom, walking them through the listing flow, fixing their photos, answering questions, taking notes on what's confusing. This is unscalable, on purpose. The founder of a recent marketplace MVP guide at Shipturtle has a great line on this: if you're not the bottleneck on seller onboarding for the first month, you're not learning fast enough.
What you're actually doing in those calls: capturing the friction in your onboarding flow. Every "wait, where do I click to add a photo?" is a ticket in your backlog. By seller 30, your friction list is the highest-leverage product roadmap you've ever had — every fix removes a real, observed problem rather than a hypothetical one.
The technical infrastructure that lets you seed manually
The reason most founders skip concierge onboarding is they're stuck in code. The fix is picking a stack that lets you ship fast enough to spend time on customers, not features. Our stance, after building marketplaces on every approach: use Sharetribe Flex + Stripe Connect Express for 80% of MVPs. Sharetribe handles the marketplace mechanics (listings, transactions, reviews, messages); Stripe Connect handles the payouts and KYC.
The total cost-to-launch with this stack is between $30,000 and $80,000 over 6-10 weeks for a 2-person team — see our full itemized marketplace cost breakdown for the line items. The alternative is custom-building everything in Rails or Node, which adds 10-20 weeks and $50,000-$150,000 to the timeline. For 95% of founders, that extra time is exactly the cold-start period they should have been spending acquiring sellers instead.
One critical caveat: even on Sharetribe, the Stripe Connect production gotchas are real and will bite you in week 2 of operation — KYC freezes, payout holds, refund liability, webhook idempotency. Read that post before you hand the keys to your first paying seller. (If you'd rather skip the build entirely, our Sharetribe development team ships these MVPs end-to-end with the cold-start playbook baked in.)
When to flip from supply to demand
The flip happens when three things are simultaneously true:
- You have 20+ active sellers with at least 50 total live listings.
- Each seller has had at least one real transaction through the platform.
- Sellers are asking you for more buyers — not the other way around.
Until all three are true, more buyer marketing is wasted spend. Buyers will arrive, see thin inventory, bounce, and tell their friends your marketplace is empty. Once all three are true, you can flip the funnel: shift content marketing to buyer keywords, run targeted ads, list on directory aggregators, and run launch posts on Product Hunt and niche communities.
Cold-start metrics that actually predict success
Forget DAU/MAU at this stage. The only metrics that matter:
- Active seller count — sellers with at least one listing live and at least one login in the last 14 days. Target: 20-25 by week 8.
- Listing-to-seller ratio. If sellers have 1 listing each, your seller engagement is shallow. Target 2.5+ per seller.
- Time-to-first-transaction per seller. Sellers who wait more than 30 days for their first sale churn at 60-70% rates. If this number creeps up, your demand side is underbuilt.
- Buyer-to-seller ratio. Below 3:1 means you have a buyer problem. Above 8:1 means you're overweighting demand and your sellers are getting overwhelmed (which sounds great but creates a bad seller experience that triggers churn).
The 4 cold-start mistakes that have killed launches we've watched
The patterns we've seen fail, in order of how lethal each one is:
- "We'll launch broadly and let the market decide who our customer is." The market decides nothing. You decide. A broad launch dilutes your supply across too many segments and you never reach atomic density anywhere.
- "We'll just turn on Google Ads to get buyers." Paid acquisition before atomic supply density is the fastest way to burn money. Buyers click, see thin inventory, leave, and your CAC is now infinity divided by zero.
- "We don't want to subsidize the supply side; sellers should pay us." Sellers will pay you eventually. Not in week 4. Not in month 2. Take-rate negotiation comes after density, not before.
- "We'll automate onboarding from day one." Automated onboarding optimizes for scale you don't have. The 30 sellers you onboard manually in month one will teach you more than the 300 you'd onboard automatically in month six.
FAQ: The cold-start questions founders actually ask us
What is the cold-start problem in a marketplace?
The cold-start problem is the chicken-and-egg trap a two-sided marketplace faces at launch: buyers don't show up because there's no inventory, and sellers don't list because there are no buyers. Solving it requires seeding one side manually, almost always the supply side, until the network reaches enough density to self-sustain.
Should I focus on getting sellers or buyers first?
Sellers, in 95% of marketplaces. Buyers bounce in 12 seconds when they see thin inventory. Sellers will list with zero buyers if your cost-of-joining is low enough — fee holidays, free profile setup, and concierge onboarding are the standard levers.
How long does the cold start typically take?
3-6 months for a focused niche launch. Founders who try to launch broadly often spend 12+ months and never reach atomic density. The narrower your beachhead, the faster you clear cold start.
What's the best stack for a marketplace MVP in 2026?
Sharetribe Flex plus Stripe Connect Express for ~80% of MVPs. Total launch cost runs $30K-$80K over 6-10 weeks for a 2-person team. Custom-build only if you have a non-standard transaction model or compliance requirements that Sharetribe can't handle.
When should I stop subsidizing the supply side?
Once a seller has earned a meaningful amount through your platform — typically $1,500-$5,000 — and they explicitly tell you they want more buyers. Switch to standard fees gradually. Premature take-rate enforcement is the most common cause of seller churn we see.
How do I know my marketplace has cleared cold start?
Three signals together: 20+ active sellers, 50+ live listings, and sellers asking you for more buyers (not the reverse). Until all three are true, more buyer-side spending is wasted.
Stuck mid-cold-start? Get a free estimate.
If you've shipped your marketplace MVP and you're 30-90 days in without traction, we can help on both fronts: the technical side (fixing onboarding friction, instrumenting the right metrics, scaling Sharetribe and Stripe Connect to handle real volume) and the playbook side (auditing your supply-seeding tactics, tightening your beachhead, designing the supply→demand flip). Get a free estimate — we'll review your situation and propose a 4-week plan within 48 hours.
