Finding and validating profitable startup ideas in overlooked niches

Fire at the first serious doubt.

If you ever think "should I fire this person?", the answer is yes; superstars never trigger that thought.

Why: Early decisive cuts prevent months of drag and culture damage.

"If I ever think, should I fire this person? Even once, I should fire them immediately."

Choose niches with real demand and little competition.

Look for problems that few others are solving—funeral homes, pest control, government-form software—rather than crowded spaces like restaurants or generic SaaS.

Why: Competition equals lower margins and higher failure rates; monopolistic niches allow pricing power and easier wins.

"Fish where the fish are... you actually want to walk off into the forest and find a small fishing hole with lots of fish and very little competition."

Start with "baby weights" before attempting 300-pound lifts.

First-time founders should avoid highly regulated, capital-intensive, or hyper-competitive markets (AI moonshots, new banks) and instead pick simple, cash-flow-positive models.

Why: Early wins build confidence and capital; complex markets magnify rookie mistakes.

"You don't want to walk into the gym on day one and try and deadlift 300 pounds... take the baby weights and start slowly building your muscle."

Boring is beautiful.

Unsexy businesses (form-filling for government grants, kitchen-exhaust cleaning) often generate $20-30 M ARR with minimal competition.

Why: Few founders are willing to work on dull problems, leaving high-margin opportunities for those who are.

"Nobody wakes up and goes, I want to make form-filling software. But I think they would if they could make $20 million a year."

Avoid markets littered with dead bodies.

If many have repeatedly failed in a space (bars, local news, pizzerias), assume hidden complexity rather than assuming you’re smarter.

Why: Business-model forces usually trump individual talent; structural problems persist.

"The biggest mistakes I've made have been going into business models where other people have repeatedly failed and thinking, I can do this better."

Use AI to simulate the business before spending capital.

Feed Claude or ChatGPT a concept and ask for P&L, regulatory hurdles, payroll, and realistic customer willingness to pay.

Why: Rapid modeling reduces expensive trial-and-error and surfaces deal-killers early.

"You should be able to do it with AI... hey, I'm thinking about starting a Botox clinic. Can you break down the numbers?"

Build moats: brand, network effects, or high switching costs.

Prefer businesses where customers stay because of loyalty (Tylenol), community size (Letterboxd), or pain of switching (Salesforce).

Why: Moats protect long-term cash flow and make the business hard to kill even with mediocre management.

"I'm looking for a business though, where it is so good that it's hard to mess up... something that has staying power that is hard to compete with."

Separate business from job by reaching scale.

A solo pressure-washing gig is a job; a 10-crew operation with delegated labor is a business. Design for scale from day one.

Why: Without scale, the founder is trapped doing grunt work; with scale, they can focus on highest-value activities.

"If the business can get big enough where you can have employees, he can just focus on sales or digital marketing... there's a big difference between a business and a job."