The best source of ideas is the industry you already worked in. You are not looking for a gap in the market — you are looking for work you have watched people do badly.
The best source of SaaS ideas is the industry you already worked in. Domain knowledge is the one advantage a competitor cannot acquire in a weekend, and it is the reason vertical SaaS keeps working long after the horizontal version of the same tool exists.
You are not looking for a gap in the market. You are looking for work you have watched people do badly.
Three advantages compound, and none of them is available to someone entering your niche from outside.
You know the vocabulary. Vertical buyers recognise instantly whether a product was built by someone who understands their work. Getting the words right does more for conversion than any design decision.
You know where they gather. Trade associations, certification bodies, regional groups, specific conferences. That is a distribution channel you can name on day one, which is the single strongest predictor of whether a solo product survives.
You know what is actually painful. Outsiders build for the problems that are visible from outside. Insiders build for the ones that consume a Thursday afternoon every month and never appear in any job description.
The hated spreadsheet. Every industry has one — a file passed around, versioned by filename, breaking constantly, and central to something important. It has survived because no horizontal tool fits the workflow closely enough to be worth switching to.
What your employer outsourced. An agency or contractor invoice is demand with a number attached. If a firm pays £800 a month for something routine, you already know the budget exists and roughly what it is.
The bent generic tool. Watch for people using a general product in a way it was clearly not designed for — a CRM tracking court dates, a project tool tracking clinical sessions. The workaround is the specification.
The recurring obligation. Regulatory filings, renewals, audits, certifications. These arrive on a schedule, nobody enjoys them, and the deadline creates urgency you do not have to manufacture.
Domain knowledge tells you the problem is real. It does not tell you whether enough people will pay, and vertical niches can be genuinely too small.
| Check | What good looks like |
|---|---|
| Countable population | You can estimate how many exist, and it is thousands not dozens |
| They already buy software | Existing tools in the space, however bad |
| A place they gather | Association, forum, conference, publication |
| Decision requires a committee | Kills self-serve. Different business. |
The last row is the one that surprises people. A niche can be perfect in every other way and still be wrong for a solo founder if buying requires procurement — the validation checklist covers why the buyer deciding alone matters so much.
Smaller is usually safer than it feels. A few thousand potential customers at a meaningful monthly price is a substantial solo business, and small markets are ignored by funded competitors precisely because they cannot support a sales team.
The real risk is not size but concentration — if twenty organisations control the whole market, losing two is a crisis. Fragmented niches with many small buyers suit a self-serve solo product far better.
A useful sanity check: if you can name fifteen specific businesses or individuals who would benefit, and you could contact all fifteen this week, the niche is reachable enough to start. Whether it is large enough is a question for after the first ten customers.
Common, and worth taking seriously rather than dismissing. People leave industries for reasons.
The distinction that matters is whether you disliked the work or the environment. Building software for an industry is not the same as working in it — you get the domain knowledge without the parts you left. But if the subject itself bores you, two years of it will be hard, and that is a real cost.
The compromise most people miss: adjacent industries. You often understand the workflows of the people your old industry sold to or bought from almost as well as your own, without the same associations.
Do not build. Your advantage is access, so use it — talk to ten people in the niche before writing code. Your former colleagues are the warmest possible research list and asking about the past rather than the future keeps the answers honest.
Then score it properly with the five tests, and set a price before launch rather than after — the wider method for finding ideas covers the other three sources if this one does not produce a candidate.
Look for four things: the spreadsheet everyone hates and nobody replaced, work your employer outsourced to an agency, a generic tool people bend to fit a workflow it was not built for, and the recurring compliance task nobody enjoys.
Three things compound: you know the vocabulary, which vertical buyers use to judge whether a product was built by an insider; you know where they gather, which is a distribution channel on day one; and you know which problems are actually painful rather than merely visible.
Size is less risky than concentration. A few thousand buyers at a meaningful monthly price is a substantial solo business, and small markets are ignored by funded competitors. The real risk is a market where twenty organisations control everything.
Distinguish between disliking the work and disliking the environment — building software for an industry is not the same as working in it. If the subject itself bores you, consider adjacent industries whose workflows you also understand.
Talk to ten people in the niche before writing code. Your former colleagues are the warmest research list you will ever have, and your access is the advantage — using it for validation rather than launch is what makes it pay off.
Tell Marcus what industry you know and what you watched people do badly. You get a specific candidate and the test it has to pass.
Try GhostCoach free →14-day free trial · cancel anytime · 30-day money-back on Lifetime