What Validation Means
Many startups fail, and a common reason is building something nobody wants. Validation helps you avoid this fate by testing assumptions before investing significant time and money.
Validating a startup idea means testing the assumptions it depends on before you build it. It is how you find out early, when changing course is still cheap.
The Validation Mindset
Your idea is a hypothesis, not a fact. Your job is to prove or disprove it quickly. Be prepared to pivot based on what you learn. Seek disconfirming evidence, not just validation.
Write Down Your Assumptions First
Every idea rests on a handful of assumptions. Write them as sentences that could turn out false:
- Problem: enough people have this problem, often enough to care
- Customer: a specific group has it, and you can reach them
- Value: they would pay for a solution, or switch from what they use now
- Channel: you can reach them at a cost that makes the business work
- Feasibility: you can build it at the quality it needs. For an AI product, that includes whether the model can do the job on real inputs
Rank them by risk. The assumption that would end the idea if it were wrong goes first.
Six Validation Methods Compared
Each method answers a different question. All six follow the same shape: what you do, what it tells you, and what it doesn't.
Method 1: Customer discovery interviews
Talk to potential customers. Ask about their problems, not your solution. Use open-ended questions: 'Tell me about the last time you experienced X.' Never pitch during discovery interviews.
What it tells you: whether the problem is real and how people cope today. What it doesn't tell you: whether they would pay. Keep interviewing until the answers start to repeat.
Method 2: Landing page test
Create a simple landing page describing your solution. Drive traffic with small ad spend. Measure signup rates and engagement.
What it tells you: whether your message interests the audience you chose. What it doesn't tell you: whether people will pay, or whether the product works. Set the target before you run the test.
Method 3: Concierge MVP
Manually deliver your service before automating. This 'Wizard of Oz' approach tests demand with minimal development. Learn exactly what customers need before building.
What it tells you: what customers need and value, in detail. What it doesn't tell you: whether you can automate the work at a profit.
Method 4: Pre-sell or crowdfund
Ask customers to pay before you build. Successful pre-sales prove real demand. Platforms like Kickstarter or simple Stripe checkout work well.
What it tells you: real willingness to pay, which is stronger evidence than interest. What it doesn't tell you: whether customers will keep using the product, so deliver what you sold.
Method 5: Competitor analysis
Study existing solutions. If competitors exist, that validates market demand. Identify gaps in their offerings that you can fill.
What it tells you: that demand exists, and where the gaps are. What it doesn't tell you: that you can win. A crowded market with no gap you can fill is not validation.
Method 6: Build a rapid MVP
Create the simplest version that solves the core problem. Launch to a small audience and measure engagement. Use tools like Bubble or Webflow for no-code prototypes, or hire SpeedMVPs for a 2-3 week professional MVP.
What it tells you: what users do, not what they say: whether they finish the job and come back. What it doesn't tell you: how the product scales.
Match the Method to the Riskiest Assumption
- The problem may not be real: discovery interviews
- The message may not land with your audience: a landing page test
- The value may be unclear: a concierge MVP or a pre-sale
- Willingness to pay is unknown: a pre-sale
- The market may already be served: competitor analysis
- You may not be able to build it well enough: a rapid MVP, tested on real inputs
Decide the Result Before You Run the Test
Write down what result would make you continue, change direction or stop before you collect any data. Without that, any number can be read as success. You don't need an industry benchmark. You need a threshold that makes your own business work, such as how many paying customers would justify building.
Validation Metrics That Matter
Signup/conversion rates, willingness to pay, user engagement frequency, referral rates, and qualitative feedback intensity. Trust actions over words: what people do matters more than what they say.
A Two-Week Validation Sprint
- Days 1-2: write your assumptions and the thresholds
- Days 3-7: run discovery interviews and publish the landing page
- Days 8-12: run the cheapest test of willingness to pay: a pre-sale, a deposit or a concierge pilot
- Days 13-14: compare results with the thresholds and decide
Signs to Stop, Change or Continue
- Stop when people shrug at the problem or won't commit anything, even after you've fixed the message
- Change direction when the problem is real but the customer, the offer or the price is wrong
- Continue when people describe the problem unprompted, commit time or money, and come back without being asked
When an MVP Is the Right Test
Interviews and landing pages show interest. Only working software shows behavior. When the remaining risk is whether people will use the product, build the MVP and read the results against your thresholds. MVP vs prototype explains why a mockup isn't enough. SpeedMVPs builds MVPs in 2-3 weeks at a fixed price, which makes a working product a realistic test instead of a late commitment. For AI products, see how to validate an AI MVP idea, or talk to the team.
What You'll Get
Validation Checklist
Step-by-step validation process
Interview Script Template
Questions for customer discovery
Metrics Dashboard
Key metrics to track during validation


