Validating a two-sided marketplace when only one side makes videos
Supply publishes constantly because content is how it gets found. Demand is almost silent — so the corpus you gather by default answers the wrong half of the question.
The short answer: assume your corpus is lopsided, because it is. Supply publishes; demand does not. Read the silent side second-hand through supply-side content, and settle one question before anything else — which side is scarce, since that is the side the marketplace has to solve. Almost every wrong marketplace decision traces back to getting that one backwards.
Marketplace validation has a structural sampling problem that single-sided SaaS does not. Freelancers, sellers, operators and agencies publish relentlessly because visibility is how they acquire work. The people who hire them publish nothing, because they have no reason to. Gather videos on the category and you will end up with a rich, confident, one-sided picture.
Which side is scarce decides the product
A marketplace is a liquidity business, and liquidity is won by acquiring the scarce side. Everything downstream — which side you charge, which side gets the good interface, where the growth budget goes — follows from that single determination.
The trap is that the loud side looks important. A category full of supply-side creators hustling for visibility feels like a vibrant supply market, and it is, which means supply is abundant and demand is what you actually have to find. Loudness is evidence of competition, not of scarcity.
| What supply-side creators say | What it means | Scarce side |
|---|---|---|
| How to get more clients, pricing yourself competitively | Supply competing for work | Demand |
| Turning work away, waiting lists, raising rates | Demand competing for supply | Supply |
| Vetting clients, avoiding bad ones | Matching quality problem, not volume | Neither — trust is the product |
| Everyone names the same two channels | Matching is already solved | Probably no marketplace here |
If practitioners consistently name one or two existing channels that work well enough, you are proposing a nicer version of a solved problem. That business exists, but it is a distribution fight rather than a liquidity opportunity, and it should be entered knowingly.
Reading demand second-hand
The silent side is not unobservable, it is just quoted rather than speaking. Supply-side creators describe their buyers in detail because buyers are the subject of their working lives: how the lead arrived, what the brief said, which requests were unreasonable, why the deal died.
Treat those descriptions as demand evidence with a known bias. The bias is consistent and therefore correctable — supply narrates buyers unsympathetically, so complaints about clients are usually accurate about behaviour and unreliable about motive. What you want is the behaviour.
The same correction technique applies to any interested narrator, and the general method is in vendor content versus independent creators. Where the vocal side is genuinely thin on the ground, the fallbacks in what to do when your niche has few videos apply to each side separately.
Run two corpora, not one
- ✗Supply voices outnumber demand ten to one
- ✗Findings read as consensus when they are one side speaking
- ✗Pricing designed around what sellers wish they could charge
- ✗Scarcity judged by who was easiest to find
- ✓Each side gathered and synthesized separately
- ✓Disagreements between the sides treated as the finding
- ✓Second-hand demand evidence labelled as second-hand
- ✓Scarcity determined from competitive language, not volume
Keeping the sides apart is the single highest-value structural choice in marketplace research, because the interesting result is always the disagreement between them: what sellers think buyers want versus what buyers reportedly ask for. Blend the corpora and that contrast disappears into an average nobody holds.
Handling those disagreements deliberately rather than averaging them away is the general discipline in resolving contradictions in video research.
Trust and payment are usually the actual product
Ask what a marketplace sells and the honest answer is rarely matching. Matching is often achievable with a spreadsheet and a group chat. What is hard is trust: knowing the counterparty will deliver, that the money will arrive, and that a dispute has somewhere to go.
Practitioner content is unusually rich here because payment failures make memorable stories. Non-payment, scope creep, deposits, chargebacks and the specific moment people ask for money up front all surface unprompted, and each one is a product surface. Where those rules are regulated rather than customary, they need separating out — the reading in spotting compliance constraints in practitioner content.
Seeding the scarce side is a distribution problem
Once you know which side is scarce, the launch plan is mostly about reaching that side in a place where they already gather. The abundant side will arrive on its own and should be deliberately throttled early, because a marketplace with visible unmet supply looks worse than one with less supply and better fill rates.
Where the scarce side actually congregates is answerable from the same research rather than from channel-mix guesswork — the approach in reading distribution signals to decide where to launch, with the sizing sanity check in a market-size sanity check from creator signals.
Two sides means two ICPs and two onboardings
The practical consequence founders underestimate is duplication. Two sides need two value propositions, two onboarding flows, two sets of objections and often two pricing models, and the research burden roughly doubles rather than splitting.
Defining each side as its own customer profile, with its own evidence, is the way to keep that honest — the method in defining an ICP from video research, applied twice.
Liquidity is usually local, even when the market is not
A marketplace can be nationally viable and locally dead, and video research obscures this because creators publish to the whole internet. A corpus that proves demand exists across a country says nothing about whether it exists densely enough in any one place to produce a match.
Density is set by whatever makes a match feasible. If the service is delivered in person, the boundary is a metro area or a drive time. If it requires a licence, a language or a specific certification, that requirement is the real boundary and it can be far narrower than geography suggests. The question to ask of the research is not how many people want this, but how many want it inside the smallest unit where a match can actually happen.
Practitioner content answers this obliquely but usefully. When supply-side creators describe their catchment — how far they travel, which regions they will not serve, what makes a job too small — they are describing the boundary of a liquidity pool. Several creators drawing similar boundaries is a strong signal about the unit you have to fill.
The practical consequence is that launch is a sequence of small markets rather than one big one, and the first market should be chosen for density rather than size. A thin national marketplace looks larger and performs worse than a dense single-city one, and only the second teaches you anything about whether the matching works.
What this costs
Two corpora means two projects, which the entry plan covers. As of September 2026, Hobby is $19 a month with 2 projects and 25 videos, Pro is $59 with 8 projects and 80 videos, and Studio is $199 with 20 projects, 250 videos and 3 seats, each with a 7-day free trial — see the pricing page.
Run one project per side and compare the syntheses: what supply says buyers want against what buyers are reported to ask for. 7-day free trial.
Closing thought
The side that makes the videos is the side that needs customers. That sentence is worth reading twice before you decide which side of your marketplace to build for.
Frequently asked
Why is video research lopsided for marketplaces?
Because one side of almost every marketplace makes content and the other does not. Supply — freelancers, sellers, operators — publishes constantly because content is how they get found. Demand rarely publishes anything, so a naive corpus tells you a great deal about sellers and almost nothing about buyers.
How do I research the silent side?
Indirectly, through the vocal side. Supply-side creators describe their buyers constantly: how they were found, what they asked for, why deals fell through, what they complained about. That is second-hand demand evidence and it is more available than most founders assume.
What is the single most important thing to establish first?
Which side is actually scarce. Marketplaces are won by solving the scarce side, and founders routinely guess wrong because the loud side looks like the important one. Supply that is visibly hustling for work is usually abundant, not scarce.
What signals show which side is scarce?
Listen for who is competing for whom. When supply-side creators describe undercutting each other, chasing leads and marketing hard, supply is abundant and demand is scarce. When they describe turning work away, raising prices and waiting lists, demand is abundant and supply is the constraint.
Can research tell me whether I have a marketplace at all?
Often, yes. If practitioners describe finding each other through one or two dominant channels that already work, you are proposing a better version of a solved matching problem, which is a much harder business than a genuine liquidity gap.
What about trust and payments?
Those are the real product in most marketplaces, and practitioner content is unusually explicit about them because bad payment experiences generate memorable stories. Disputes, escrow, deposits and non-payment come up unprompted.
What does this research cost?
As of September 2026, Hobby is $19 a month for 25 videos and 2 projects, Pro is $59 for 80 videos and 8 projects, and Studio is $199 for 250 videos, 20 projects and 3 seats, with a 7-day free trial on every plan. A marketplace read usually wants two projects — one per side — which fits the entry plan.