If you want the fastest useful path, start with "Write your riskiest assumption as a falsifiable statement" and then move straight into "Conduct problem interviews, not solution interviews". That usually gives you enough structure to keep the rest of the guide practical.
Know your actual use case
This guide is written for startup idea validation fails when founders rely on interview enthusiasm instead of behavioral evidence. This guide focuses on getting observable evidence—signups, pre-sales, or observed behavior—that a problem is real and the solution is valued., so define the real problem before you try every step blindly.
Keep the scope narrow
Focus on customer discovery and founder first instead of changing everything at once.
Use the guide as a sequence
Apply one or two ideas first, then keep only the ones that improve your results in real usage.
Write your riskiest assumption as a falsifiable statement
Step 1Before any customer conversations or experiments, write down the three assumptions your business model requires to be true. Rank them by: (a) how certain you are they're true, and (b) how catastrophic it would be if they're false. The intersection of 'uncertain' and 'catastrophic' is your riskiest assumption—and the first thing you test. Most founders test convenient assumptions and avoid existential ones.
Conduct problem interviews, not solution interviews
Step 2The purpose of early customer interviews is to understand the problem space—how people currently solve the problem, how painful it is, what they've tried, and how much time and money the problem costs. Never describe your solution in these interviews. If you pitch your solution, you get social feedback on your pitch, not honest data about the problem. Ask: 'Tell me about the last time you dealt with X' not 'Would you use a product that did Y?'
Build a landing page and measure actual conversion, not reactions
Step 3Create a one-page description of your value proposition with a call-to-action—email signup, waitlist join, or pre-order. Drive 200–500 targeted visitors through Reddit communities, industry forums, or small paid ads. A 5%+ conversion to your CTA is a meaningful positive signal. 1% or less suggests your problem framing or solution description isn't resonating. Visitor behavior on the page (time spent, scroll depth) gives secondary signal.
Attempt a pre-sale before building MVP features
Step 4Ask your most enthusiastic early interviewees to pre-pay for the product before it exists—even a small amount ($50, $200, whatever fits your pricing) in exchange for a discount or early access. The number who convert from 'this sounds great' to 'I'll pay now' is the most honest signal you can get at zero engineering cost. Three to five pre-sales from a cohort of 20 target customers is strong enough to justify building.
Define your validation threshold before running experiments
Step 5Set success criteria before you start each experiment: 'I'll build if we get 100 waitlist signups in 2 weeks' or 'I'll continue if 3 of 10 interviewees have this problem acutely enough to describe it unprompted.' Defining success before the experiment prevents motivated reasoning—the natural human tendency to interpret ambiguous results as confirmation. Founders who evaluate results before defining thresholds almost always find what they hoped to find.
What if no competitor exists for my idea—does that validate it?
Not necessarily—absence of competition can mean a large unmet opportunity, or it can mean the market is too small, the problem isn't painful enough to pay for, or previous entrants tried and failed invisibly. Do a thorough research pass: search for forum discussions, look for failed companies in the space (PitchBook, Crunchbase), and ask potential customers directly what they've tried. 'Nobody has done this before' requires more validation rigor, not less.
How many customer interviews do I need before I can validate a problem?
You're looking for convergence, not statistical significance. If your first 5 interviews produce the same core problem description, unprompted, in the same terms, you're getting signal. If you've done 15 interviews and each person describes a different primary problem, you haven't found a focused segment yet. Most experienced founders can reach meaningful conviction with 15–25 interviews in a well-defined segment.
Is a letter of intent (LOI) from a business a valid form of validation?
It's a meaningful signal but not strong validation. LOIs are non-binding and require no financial commitment—they reflect genuine interest but not confirmed willingness to pay on delivery. The stronger the LOI (specific terms, decision-maker signature, stated budget), the more useful it is. LOIs combined with a small deposit or pilot agreement are substantially stronger. Treat LOIs as 'qualified interest' rather than confirmed revenue.
What's the minimum viable product versus a prototype for validation purposes?
A prototype tests whether your solution concept solves the problem. An MVP is the minimum you need to deliver that solution to real users at real stakes. For validation before building, you don't necessarily need either—a landing page, a concierge service (manually doing what the software will automate), or a pre-sell attempt can validate demand before you write a line of code. Build only what's necessary to get behavioral signal.
