Thin-niche research: what to do when your market has almost no videos
Scarce footage on your exact vertical is a routing problem, not a dead end. The workflow is documented somewhere; usually one layer up.
The short answer: stop searching the niche and start searching the layers around it — the tools it runs on, the trade next door that performs the same workflow at bigger scale, and the vendor content those tools produce. A vertical with five videos usually sits downstream of a tool with five hundred.
The first hour of research in a specialist market often produces the same discouraging result: a handful of videos, most of them promotional, one genuinely useful walkthrough from 2021, and a suspicion that the whole method does not apply here. That conclusion is usually wrong, and acting on it costs you the market rather than the method.
First, work out which kind of thin you have
Thin coverage comes in two flavours that look identical from the search results page and mean opposite things commercially. A small market has few practitioners, so it has little content, little forum traffic, and few job listings. A quiet market has plenty of practitioners who simply do not make videos — often because the work is confidential, regulated, or so unglamorous that nobody has thought to film it.
The diagnostic is whether demand leaves traces in places other than video. Job listings describing the workflow, vendor webinars, industry association sessions, and recurring questions in trade forums all indicate a quiet market. If those are thin too, you may genuinely be looking at a small one, which is not automatically disqualifying but does change the model — the arithmetic for that is in the market-size sanity check from creator signals.
| Layer | What to search | What it gives you |
|---|---|---|
| The niche itself | Role and workflow names | Ground truth, usually two to five sources |
| The tool layer | Software the niche runs on, by name | Where the workflow visibly breaks |
| Vendor content | Webinars, conference talks, release demos | Official workflow, roadmap direction, gaps |
| Adjacency | The same workflow in another industry | Problem shape, at scale, with more voices |
| Upstream and downstream | Who hands work in, who receives it | The real constraint, often outside the role |
The tool layer is where the footage lives
Almost every specialist workflow runs on named software: a practice management system, an ERP module, a scheduling tool, a CAD package, a billing platform. Those tools have documentation channels, consultants who publish tutorials, and users who record themselves solving problems with them. The workflow you cannot find is visible inside those videos as the thing the tool is being wrestled into doing.
This is the highest-yield move in thin-niche research because it inverts the search. You are no longer asking “who talks about this job” but “who talks about the software this job depends on”, and that second question always has answers. What you extract is the same material a normal pass extracts: the workaround, the export to a spreadsheet, the second system opened alongside the first — the tells covered in finding micro-SaaS ideas in tutorial videos.
Independent consultants who implement one platform for a living produce the single most useful content in thin markets. They demonstrate real client workflows, they name the platform’s limits candidly because their credibility depends on it, and they say out loud which requests they cannot fulfil — which is a product backlog spoken aloud.
Build an adjacency corpus deliberately
An adjacency is another industry doing a structurally identical job. If your niche reconciles two systems weekly, find a trade that reconciles two systems daily. If your niche has a manual approval queue, find one where that queue is the bottleneck of a much larger operation.
The value is scale: the adjacent industry has more practitioners, so it has more footage, so the pattern is visible in ten voices instead of one. The discipline is that everything you learn there is a hypothesis about your market rather than a fact about it. Write it down as a question to test, not as a finding, and carry it into interviews using the framework in turning video research into a customer-interview script.
- ✗Searched the vertical name only
- ✗Concluded the market is too small
- ✗One 2021 video treated as the whole picture
- ✗Adjacent findings used as facts
- ✓Vertical, tool layer, vendor and adjacency corpora
- ✓Small versus quiet diagnosed explicitly
- ✓Single-source claims labelled as hypotheses
- ✓Adjacent findings carried into interviews as questions
Conference talks do the heavy lifting
Quiet markets are quiet on YouTube search but loud at their own events, and most industry conferences now post sessions publicly. A recorded conference talk is unusually dense research material: the speaker is a practitioner, the audience is practitioners, the questions at the end are real objections, and the framing tells you what the industry currently believes about itself.
Association webinars and vendor user-group sessions behave the same way and are often the only public place a regulated workflow is described end to end. Building those into the corpus is a habit worth keeping even in rich niches, which is the argument made in using conference talks and podcasts as research sources.
Thin often means constrained
There is a common reason a workflow has no public footage: showing it would expose client data, patient records, financial detail, or a regulated process. That absence is itself a finding. It tells you the product will be bought under constraints, that a demo may need synthetic data, and that procurement will involve someone who was not in the room for the sales call.
Treat the silence as a prompt to go looking for the rules rather than as an information vacuum, using the approach in spotting compliance constraints in practitioner content. Vendor content is again the way in: platforms serving regulated markets publish compliance-shaped material constantly, because it is what their buyers ask about first.
Weight your sources honestly
A thin-niche corpus is unbalanced by construction, and the analysis has to know that. Two on-topic sources plus twelve adjacent ones is not a fourteen-source consensus; it is two observations and twelve analogies. Recording the layer each source came from keeps the synthesis honest and stops an adjacency pattern from being reported as a niche finding.
The practical form is a confidence label per claim: observed in the niche, observed in the tool layer, inferred from adjacency. Anything in the third category needs a human conversation before it earns a place on a roadmap, and anything in the first deserves disproportionate weight even though it came from one video. That labelling discipline is what stops the failure mode described in the failure modes of AI video research.
What a thin-niche pass should conclude
A good outcome here is not a roadmap. It is a three-part answer: whether the market is small or quiet, what the workflow actually looks like end to end, and the three questions that only a practitioner can settle. That is enough to justify the interviews, and the interviews are where a thin niche becomes a real one.
If the pass concludes small — thin everywhere, not just on video — that is a legitimate and cheap result. Killing an idea for the cost of an afternoon is the highest-return outcome the process produces, which is the case made in disqualifying a SaaS idea on early signals and in choosing between two ideas with evidence.
What the pass costs
Thin-niche work uses more sources than a rich pass because of the tool and adjacency layers — usually twenty to thirty-five in total. As of September 2026, Hobby covers 25 videos a month and 2 projects at $19, Pro covers 80 videos and 8 projects at $59, and Studio covers 250 videos, 20 projects and 3 seats at $199, all with a 7-day free trial. Pro is typically the right plan for a genuinely thin vertical; see the pricing page.
Run the vertical, the tool layer and an adjacency as one project, with per-source notes that keep track of which layer each finding came from. 7-day free trial.
Closing thought
The markets with no content are frequently the markets with no software either, and for the same reason: nobody outside them has looked. That is an obstacle for an afternoon and an advantage for a decade.
Frequently asked
My niche has almost no YouTube videos. Is video research still usable?
Yes, but you change what you search for. Instead of the niche, you research the shape of the workflow, the tools the niche runs on, and the adjacent industry that does the same job at larger scale. Thin niches almost always sit downstream of a well-documented tool or a well-documented neighbouring trade.
Is a thin niche a bad sign for the business?
Not on its own. Thin video coverage sometimes means the market is small, and sometimes it means the practitioners are busy people who do not make content. Distinguishing the two is the first job, and the tool and adjacency searches are what distinguish them.
How do I tell a small market from a quiet one?
Look for demand that leaves traces elsewhere: vendor conference talks, tool-specific tutorials with high views relative to the channel, job listings describing the workflow, and forums where the same question recurs. A quiet market has strong traces in those places and few videos; a small one has thin traces everywhere.
What is an adjacency corpus?
A set of sources about a different industry that performs a structurally identical workflow — the same handoff, reconciliation, or review queue — used to understand the problem shape when your own vertical has no footage. The findings transfer as hypotheses to be checked, not as conclusions.
How many sources do I need if the topic is thin?
Fewer on the niche itself and more on the surrounding layers: typically two to five directly on-topic, plus eight to fifteen across the tool layer and one or two adjacencies. Total corpus size ends up similar to a normal pass; the composition is what changes.
What is the biggest risk with thin-niche research?
Over-reading a single source. When only one person has documented your workflow publicly, their idiosyncrasies look like industry norms. Any conclusion resting on one voice should be labelled as a hypothesis and tested with an interview before it reaches a roadmap.
What does this kind of pass 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, each with a 7-day free trial. A thin-niche pass uses more sources than a rich one because of the adjacency layer, so Pro is often the practical plan.