How to Validate a Business Idea Before You Build

Most failed startups didn't fail because of bad code. They failed because they built something nobody wanted.

Validation is the process of testing whether a problem is real, whether people will pay to solve it, and whether you're the right person to solve it — before you invest months building a product. Done right, validation takes 2–6 weeks and saves you from 6–18 months of wasted effort.

Here's the framework we use at IdeaRunway, and the one that underlies our AI scoring pipeline.

Step 1: Define the Problem Precisely

Vague ideas fail. "Better project management" is not a business — "project management specifically for construction subcontractors who need to track daily progress logs and weather delays for insurance claims" is.

Your problem statement should answer:

Who has this problem? (Be specific — not "small businesses" but "2-person HVAC shops")
How often do they experience it? (Daily, weekly, per client project?)
What does it cost them in time or money to work around it today?
What are they currently doing about it? (Spreadsheet, duct tape, paying someone, ignoring it?)

If you can't answer these questions specifically, you don't understand the problem well enough yet.

Step 2: Talk to 10 People in the Target Audience

This is the step everyone skips. Don't.

You don't need a survey. You need 10 real conversations with people who have the problem you think you're solving. Ask:

"Walk me through how you currently handle [X]."
"What's the most frustrating part of that process?"
"Have you ever tried to fix it? What happened?"
"How much would you estimate this costs you per month?"

What you're looking for: Do people describe the problem without you prompting them? Do they light up when you describe a solution? Do they ask "when can I use it?"

If 7 out of 10 people shrug and say it's "not really a problem," that's signal. If 8 out of 10 say "oh god, I've been trying to fix this for years," that's also signal — the right kind.

Step 3: Audit the Competition

People often skip this because they're afraid of finding competitors. Don't be. Competition means there's demand.

How to audit:

Google the problem, not your solution. ("roofing contractor software" not "AI CRM for roofers")
Check G2, Capterra, and Product Hunt for existing tools
Read the 1-star and 3-star reviews — that's where unmet needs live
Check Reddit and niche forums for complaints about existing tools

What good competition looks like: Expensive tools with bad UX, enterprise software ignoring SMBs, outdated incumbents with no recent updates, tools that solve 70% of the problem but not the core job.

What bad competition looks like: Well-funded startups with strong product-market fit and a loyal user base. You can still compete, but you need a sharper wedge.

Step 4: Score the Opportunity

Before building, quantify the opportunity. IdeaRunway's AI does this automatically across 5 stages, but you can do a basic version manually:

| Signal | Score (1–10) | |--------|-------------| | Problem frequency (daily = 10, monthly = 3) | | | Willingness to pay (enterprise = 10, consumer = 3) | | | Market size (large = 10, niche = 4) | | | Competition level (blue ocean = 10, saturated = 2) | | | Your unfair advantage (deep domain expertise = 10, none = 2) | |

If you score below 30/50, think hard about whether this is the right idea.

Step 5: Run a Pre-Code Test

This is the fastest signal. Before writing a line of code, test whether people will actually do something:

Options (choose one):

Landing page test: Put up a page describing the product, add a "Join waitlist" CTA, buy $100 of targeted ads. If your CVR is above 15%, the concept resonates.
Fake door test: Add a button/link in an existing product or community. Count clicks.
Pre-sell: Offer the product at a discount to early adopters. Ask for payment info. "I'm interested" costs nothing — entering a credit card is real signal.
Manual concierge: Do the thing manually for 3 customers before automating it. Learn exactly what they need.

The goal isn't to build a perfect prototype — it's to answer "will anyone give me money for this?" before you spend months finding out.

Step 6: Define Your Success Metric

Validation is only useful if you define what "validated" means before you run the test. Otherwise, you'll rationalize weak results.

Good success metrics:

10 people on a paid waitlist at $X/month
3 customers willing to pay in advance for beta access
20% click-through rate on a landing page
5 customers who described the problem in their own words without prompting

Set your threshold before you start. If you hit it, build. If you don't, iterate the problem statement or move on.

How IdeaRunway Automates This Process

Our AI scoring pipeline runs every idea through:

Heuristic scoring — market size, competition level, complexity (from enrichment)
Claude deep analysis — opportunity score, competitive landscape, differentiation angles
Devil's advocate gauntlet — stress-testing for fatal flaws and survivability
Market research — real-world competitor discovery, market signals, funding activity
Business plan synthesis — go-to-market, pricing, validation strategy, unit economics

When you browse ideas on IdeaRunway, you're looking at ideas that have already passed steps 1–5. The free tier shows you the problem, market, and verdict. Unlock full access to see the complete validation breakdown, competitor analysis, and ready-to-use MVP prompt.

The Most Common Validation Mistakes

Asking friends — friends will be supportive and useless. Talk to strangers who are your target customer.
Building first, validating later — your sunk cost makes it hard to hear negative feedback honestly.
Validating the solution, not the problem — ask about the problem, not whether they like your product idea.
One good conversation = validated — you need enough signal to distinguish luck from pattern.
Ignoring the "why not?" — the people who say no often have the most useful insights.

What to Do When Validation Fails

If your tests don't hit your success metric:

Check the targeting — did you talk to the right people?
Check the framing — is the problem statement resonating?
Check the timing — is this something they need to solve now?
Check the price — is the willingness to pay where you assumed?

Often, a failed validation points to a refinement opportunity, not a dead end. The problem is real but the customer segment, use case, or pricing model needs adjustment.

Browse validated SaaS ideas → — every idea on IdeaRunway has passed our AI validation pipeline before you see it.