How To: Find Good Startup Ideas

The third in this series.
HOW came first. Then WHY. Now: WHAT.

So. You want to start a company… you just don't have an idea.

You're in good company. Most aspiring founders get stuck right here — whether you've decided you want to build and you're now hunting for what, or you've spent years inside an industry watching problems go unsolved without ever considering that you might be the one to solve them.

Here's the slightly annoying truth: the best startup ideas aren't really ideas at all. They're problems with a name. So this is about how to find good problems. 

Good news: problems are much more findable than you think. 

Bad news: you might not recognize them.

Stop Waiting for the Perfect Idea

The number one reason aspiring founders never get started is that they're waiting for a brilliant idea to strike them. They're at their day job, in their commute, in the shower, hoping inspiration shows up.

It doesn't work like that.

‍The founders who find good ideas are actively looking. They practice noticing problems — every meeting, every clunky workflow, every annoyed conversation becomes data. The problem-spotting muscle gets stronger the more you use it.

Problems vs. Annoyances

To do that, you’ve got to get clear on what counts as a real problem. Most people initially identify "problems" that are actually just annoyances — there's a difference.

‍An annoyance is something that bothers you sometimes. You'd rather it didn't exist, but you mostly live with it. You don't pay to make it go away. You don't spend time looking for workarounds. If a free tool appeared that solved it, you might try it and forget about it.

A problem is something that costs people real time, real money, or real frustration on a recurring basis. They've tried to solve it. They have workarounds. They may already be paying for current solutions that don’t work well. They'd pay more for one that did.

So when you’re seeking a problem, the simplest test is: would this person pay real money to make this go away?

If they tell you "I built a spreadsheet to handle this," or "we've been looking for something to fix this for years," or "we tried Company X and it was awful" — that's a problem.

Problem Value = Frequency × Acuteness

Not all real problems are equally valuable. Two factors determine how much a problem is worth:

• Frequency: how often does the problem actually happen?
• Acuteness: how painful is it when it does?

Multiply them and you get the rough value of solving the problem.

The jackpot is high on both axes — frequent and excruciating. That's where customers happily pay good money on a recurring basis to make their lives less painful.

Plenty of great businesses are built when one axis is high and the other is moderate — either a problem that hits constantly but isn't excruciating each time, or one that hits less often but really hurts when it does. Both can produce real businesses at scale.

What's much harder is when both are just moderate. That's no-man's-land — not painful enough to make anyone reach for their wallet, not frequent enough to add up across a customer base. 

And if either dimension is low — if the problem hardly ever happens, or barely registers when it does — you don't have a business at all. 

Boring is where the Opportunity Hides

When most people imagine a startup idea, they picture something consumer-facing and exciting. The next Uber. The next Airbnb. A viral app.

That ground is brutally competitive. The customers are fickle, the gravity pulls toward winner-take-all, and breaking through is hard.

The fertile ground most aspiring founders overlook is in the opposite direction. Boring, specific, often industrial. Insurance back-office. Trade contracting. Compliance reporting. Healthcare billing.

These industries are fertile because they're full of real, recurring problems that someone is already paying real money to solve poorly. There's far less competition, because tech founders don't naturally want to go there. And the customers, when you find them, are often desperate for someone to take their pain away.

If you already work in or around one of those "boring" industries, you have an enormous advantage. You see the problems. You speak the language. You know which workarounds are universally hated. You know which problems cost the most money. Founders parachuting in from outside cannot quickly replicate any of that. Use that advantage.

If you don't have that kind of industry exposure yet, the move is to go get it. Pick a boring industry you find interesting. Take meetings with people who work in it. Read trade publications. Listen to podcasts. Talk to operators. The depth of understanding you gain compounds fast, and there's no shortcut. The founders without a starting industry who succeed are the ones who picked one and went deep. For example:

ServiceTitan was started in 2007 by Ara Mahdessian and Vahe Kuzoyan, both children of immigrant home-service contractors in California. They watched their fathers — a plumber and a residential contractor — spend their evenings doing paperwork by hand after full days in the field. They built a desktop tool to help their dads. That tool became ServiceTitan, now the dominant software platform for the trades. Public on Nasdaq since December 2024. Multi-billion-dollar business.

Calendly is a different kind of lesson. Tope Awotona spent seven years in enterprise sales — IBM, Dell, others — while trying to be a founder on the side. He tried three startups before Calendly: a dating site, a projector business, a grill business. All three failed because he was chasing what he thought would make money. After yet another day wasted on scheduling back-and-forth as a sales rep, he decided to build the tool he wished existed. Calendly launched in 2013, bootstrapped for eight years, and is now a multi-billion-dollar business. 

None of these founders sat around waiting for inspiration to strike. They put themselves in proximity to real problems — through their families or through their own painful daily experience — and paid attention. That's the move. Get close enough to a real problem that you can feel it. 

Use AI as a Thinking Partner

Modern AI is genuinely good at helping you find problems. It's a tireless thinking partner that can help you gain this exposure.

Use it to brainstorm. Challenge your assumptions. Research industries you don't know well. Read between the lines of customer reviews. Surface insights on complaints. A prompt:

"I have experience in [industry / role]. Help me identify the 10 most common recurring problems people in this industry face — the kinds of things that cost them real time, money, or frustration on a weekly basis. For each, briefly note who feels the pain, how often it happens, what current workarounds exist, and why the existing software solutions (if any) aren't solving the problem well."

Then dig in. Ask AI to push back; to play devil's advocate on whether the problem is actually big; to identify the worst current solutions on the market and why they're bad. Ask which problems on the list might be solved by an AI-native product specifically.

You can also point AI at one-star reviews of existing software and ask it to extract the top complaints, or to find the forums where industry insiders complain about their work. It can research the typical workflow of any role at any company so you can spot the cracks.

None of this used to be possible without a research team. It is now. 

Go Talk to People

Once you've spotted something that looks like a real problem, you need to talk to the people who have it. Not to pitch them on a solution — to understand the problem deeply.

The fundamentals of validation are the same as they've always been: get to people who have the problem, ask the right questions, listen carefully to what they actually do (not what they say they'd do). Be persistent, be creative, and respect their time when you get it.

The thing most people get wrong: they ask leading questions, like "would you use a product that did X?" People say yes to almost anything when asked like that. The answers mean nothing.

What you actually want is to ask about past behavior, not future intentions. "Tell me about the last time this happened." "What did you do?" "What workaround are you using now?" "How much does this cost you each month — in time, money, or frustration?" Those questions get real information out of people. There's a short, useful book called The Mom Test by Rob Fitzpatrick that walks through this in detail — worth a read.

And don't get stuck here. The point of validation isn't to be perfectly certain. The point is to move quickly from "I think this is a problem" to "I'm pretty sure this is a problem, and these are the people who have it." Then start building.

Some Resources to Go Deeper

This post is shaped by other people's good thinking. If you want to keep digging, a few places worth starting:

On the art of finding startup ideas:

• How to Get Startup Ideas — Jared Friedman, Y Combinator. A 32-minute video that is basically the foundational talk on this topic. If you take one next step after this post, watch this.
• How to Get Startup Ideas — Paul Graham. The essay that shaped how the startup world thinks about idea generation. Required reading.
• Schlep Blindness — Paul Graham. A short essay on why people systematically overlook the painful, unsexy problems that turn out to be the best ones to solve. Perfect companion to the boring-is-fertile-ground argument above.
• The Mom Test by Rob Fitzpatrick. A short, practical book on how to talk to people about their problems without leading them on. You can read it in an afternoon and it will change how you do customer conversations forever.‍

‍Communities and people worth following:

• IndieHackers.com. A community of founders building bootstrapped businesses, with detailed revenue numbers and origin stories. Especially valuable if path 2 from the previous post is what's calling you.
• Greg Isenberg (@gregisenberg on Instagram). Greg posts startup ideas, market observations, and "boring business" angles constantly. If you want a regular stream of candidate problems and trends to react to, his feed is one of the best around. He also runs a podcast and newsletter that are easy to find from there.

Practical places to go spot real problems:

• One-star reviews of existing software on sites like G2 or Capterra. Each one is a person describing, in detail, a real problem that current solutions don't fix.
• Industry-specific subreddits — find the one where your industry's professionals complain to each other. There's one for almost every field.
• Job postings — when companies keep hiring for the same role over and over, there's often a broken process underneath. Look at what they're trying to fix.

Ed Martin is the President & CEO of Genesis, a tech startup incubator in St. John's, Newfoundland & Labrador. He previously co-founded Clockwork Fox Studios (Zorbit's Math Adventure), a venture-backed edtech company that was acquired in 2021. Learn more about what we do.

‍