The 4 Questions That Kill 90% of Startup Ideas

You probably skipped the hard part

You have an idea. You are excited. You can already see the landing page, the first users, maybe even the logo.

I know the feeling. I have been there more times than I want to admit.

But here is what I learned after watching dozens of startup ideas (including my own) crash into walls that were visible from the start: the idea is almost never the problem. The questions you never asked are the problem.

Most founders treat startup idea validation questions as a formality. Something you do after you have already decided to build. That is backwards. These questions should come first, before the wireframes, before the code, before you tell your friends about it.

Here are the 4 questions that consistently separate real opportunities from expensive hobbies.

Question 1: Am I the right person to build this?

This is founder-market fit, and it kills more ideas than bad markets do.

The test is simple: Do you have an unfair advantage in this space? Not “I think this is interesting” but something concrete. Domain expertise. A network of potential customers. Distribution access. Years of experience with the problem.

Real example: I once had an idea for a logistics optimization tool for restaurants. The market was there. The problem was real. But I had never worked in food service, I had no restaurant contacts, and I had zero distribution into that world. A restaurant owner with basic coding skills would have been 10x better positioned to build it.

I killed the idea. Not because it was bad. Because I was the wrong founder for it.

Ask yourself:

• Have I spent years dealing with this problem firsthand?
• Do I already know 10+ potential customers I could call today?
• Can I reach my target market without paying for every single click?
• Would someone look at my background and say “of course that person built this”?

If you answered no to most of these, you are probably the wrong founder for this idea. That does not mean the idea is worthless. It means someone else should build it.

Question 2: What are the real constraints?

Every idea sounds great in the abstract. Constraints are what turn abstractions into reality checks.

There are three types of constraints most founders ignore:

Technical constraints. Can this actually be built with today’s technology, at a cost that makes sense? “AI-powered” is not a magic wand. What specific AI capabilities do you need? Do they exist? Are they reliable enough for production?

Financial constraints. How much money do you need before the first dollar comes in? If the answer is “a lot,” how are you funding it? Bootstrapping a product that needs $500K in infrastructure before launch is a different game than bootstrapping a SaaS that needs a weekend of coding.

Market constraints. Is the market big enough to sustain a business, but small enough that you can actually reach people? A $50B market sounds great until you realize you need $10M in marketing just to get noticed. A $5M niche sounds small until you realize you can own 20% of it with a blog and some Reddit posts.

Real example: A friend wanted to build an AI tool that analyzed satellite imagery for farmers. Cool idea. The constraint nobody talked about? Satellite imagery APIs cost thousands per month, and the target customer (small farmers) had a willingness to pay of maybe $20/month. The unit economics were broken from day one. Six months of building before anyone did this math.

Question 3: What is the strongest argument against my own idea?

This one hurts. It is also the most valuable question on this list.

Most founders build a case FOR their idea. They look for evidence that confirms what they already believe. That is confirmation bias, and it is the silent killer of startups.

Flip it. Spend 30 minutes building the strongest possible case AGAINST your idea. Pretend you are a skeptical investor who gets paid to find holes.

The arguments usually fall into these buckets:

• “This already exists and nobody cares” (market saturation)
• “The people who need this will not pay for it” (willingness to pay)
• “You can build this, but you cannot sell this” (distribution)
• “This works for 100 users but breaks at 10,000” (scalability)
• “One regulation change kills this overnight” (external risk)

Real example: I was working on a tool that aggregated job listings from multiple platforms into a single dashboard. The argument against it? LinkedIn, Indeed, and Glassdoor actively fight aggregation. They block scrapers, change APIs, and send legal threats. The entire business model depended on the cooperation of companies that had every incentive to shut me down.

That argument was strong enough to kill the idea. And it took 30 minutes to find, not 3 months.

Question 4: Have you talked to actual customers?

Not your friends. Not your mom. Not your Twitter followers. Actual people who have the problem you think you are solving, and who currently spend money (or significant time) dealing with it.

Here is what most “customer research” looks like:

“Hey, I am building X. What do you think?”

That is useless. People are polite. They will say “sounds cool” and never think about it again.

Better questions:

• “Tell me about the last time you dealt with [problem]. What happened?”
• “What did you try? What worked? What did not?”
• “How much time/money do you spend on this today?”
• “If I could make [problem] disappear tomorrow, how much would that be worth to you?”

The pattern I have seen over and over: Founders talk to 0-2 people before building. Then they build for 3-6 months. Then they launch and discover that the problem they solved is not the problem people actually have.

Real example: A founder I know was building a project management tool “for creative agencies.” He talked to 5 agency owners before writing any code. Three of them said the same thing: “We do not need another project management tool. We need a tool that handles client approvals without 47 email threads.” He pivoted before building anything. The approval tool found paying customers in week one.

Five conversations. That is all it took.

The framework in practice

These 4 questions work as a sequential filter:

1. Founder-market fit filters YOU out of ideas you should not pursue
2. Constraints filter out ideas that cannot work in reality
3. Counter-arguments filter out ideas with fatal flaws you have not seen
4. Customer conversations filter out ideas based on assumptions instead of evidence

Most ideas die at question 1 or 2. That is fine. That is the point. Every idea you kill early is months of your life you get back.

Using this as a system

I got tired of running through these questions manually for every idea I had, so I built them into a structured process. It is an open source AI skill for Claude that walks through these questions (and a lot more) systematically, using research agents to pull real market data instead of relying on guesses.

If you want to try it: github.com/ferdinandobons/startup-skill

But honestly, even without any tool, just sitting down for an hour with these 4 questions and a blank document will save you more time than any framework or methodology out there. The hard part is not knowing the questions. The hard part is being honest with the answers.