Here is a number that should terrify every aspiring founder: 90% of startups fail, and the number one reason is building something nobody wants. The Y Combinator library is full of post-mortems confirming this pattern. Not bad engineering. Not poor marketing. Building the wrong thing entirely.
The temptation is understandable. You have an idea that feels brilliant. You want to start building immediately. But spending $15K–$50K on an MVP before validating demand is like printing 10,000 copies of a book before anyone has read the first chapter. At Meld, we have watched dozens of founders make this mistake—and we have helped dozens more avoid it entirely.
Before you write a single line of code or hire a single developer, run through these seven validation steps. They cost almost nothing, take two to four weeks, and dramatically increase your odds of building something people actually pay for. We cover the financial side in detail in our complete breakdown of AI development costs, but first you need to know whether the idea is worth funding at all.
Step 1: Problem Interviews (Talk to 20 Real Users)
The single most valuable thing you can do before building anything is talk to people who have the problem you want to solve. Not your friends. Not your co-founder. Actual potential customers in your target market.
The goal is not to pitch your solution. The goal is to understand the problem deeply:
- How do they currently solve this problem? Every problem has a current solution, even if it is a spreadsheet or a sticky note.
- How much time or money does the current solution cost them? This tells you the ceiling for what they will pay.
- What have they tried that did not work? This reveals failed approaches you should avoid.
- How often does this problem occur? Frequency determines urgency and willingness to pay.
Aim for 20 interviews minimum. The first five will surprise you. By interview fifteen, you will start hearing the same patterns. By twenty, you will know whether the problem is real, painful, and frequent enough to build a business around.
When we helped validate AeroCopilot, the team spoke directly with Brazilian general aviation pilots before writing any code. Those conversations revealed that pilots were spending hours on manual flight planning calculations that could be automated—a clear, quantifiable pain point with regulatory urgency behind it. That validation shaped every feature decision that followed.
Step 2: Solution Sketch (Not a Wireframe—A Story)
Once you understand the problem, sketch your solution as a user story, not a wireframe. The Lean Canvas framework is excellent for structuring this thinking on a single page. Wireframes are premature at this stage. Instead, write out the experience:
"Maria is a property manager with 47 units. Every Monday she spends three hours chasing late rent payments. With our tool, she opens her dashboard at 9 AM and sees that automated reminders were already sent on Friday, two tenants set up payment plans overnight, and she only needs to personally follow up with three accounts."
This narrative forces you to think about the transformation, not the features. What does the user's life look like before and after? If you cannot write a compelling before-and-after story, the problem may not be painful enough.
Share this story with five of the people you interviewed. Watch their reaction. If they say "I need that," you are on to something. If they shrug, go back to step one.
Step 3: Landing Page Test (Measure Real Demand)
Words are cheap. Credit cards are not. The landing page test bridges that gap.
Build a simple landing page that describes your solution and includes a clear call to action—a waitlist signup, a "request early access" form, or even a pre-order button. Drive traffic to it using $200–$500 in targeted ads (Google, LinkedIn, or Meta depending on your audience).
The metrics that matter:
- Click-through rate on ads: Above 2% suggests the problem resonates
- Landing page conversion rate: Above 5% for email signups is strong; above 2% for paid commitments is exceptional
- Cost per lead: This tells you whether your customer acquisition economics can ever work
You do not need a product to run this test. You need a clear value proposition, a professional page, and a small ad budget. If your bilingual digital strategy targets both English and Portuguese-speaking markets, test both languages separately—conversion rates often differ dramatically.
Step 4: Competitor Analysis (Find the Gap, Not the Void)
Founders often say "we have no competitors" as if it is a good thing. It is not. No competitors usually means no market. What you want is competitors with clear weaknesses.
Map every existing solution in your space across these dimensions:
- Price point: Where are they clustered? Where is the gap?
- Target customer: Who are they ignoring?
- Feature depth vs. simplicity: Is the market over-served with complex tools and under-served with simple ones, or vice versa?
- Technology generation: Are incumbents using legacy tech while AI-native approaches could deliver 10x better results?
The best startup opportunities exist where incumbents are too expensive, too complex, or too slow to adapt. AI-native development has created an entirely new category of competitive advantage—you can build in weeks what would have taken incumbents months. We explore this dynamic in depth in what AI-native development actually means.
Step 5: Pricing Validation (Will They Pay, and How Much?)
This is where most founders get squeamish. Asking people to commit money—even hypothetically—feels uncomfortable. Do it anyway.
The Van Westendorp pricing model works well for new products. Ask your interview subjects four questions:
- At what price would this be so cheap you would question its quality?
- At what price would this be a great deal?
- At what price would this start to feel expensive but you would still consider it?
- At what price would this be too expensive to consider?
Plot the responses. The intersection points reveal your optimal price range. More importantly, if people cannot answer these questions, they probably do not care enough about the problem to pay for a solution.
For SaaS products, also validate your pricing model—monthly subscription, usage-based, per-seat, or hybrid. The model matters as much as the number. Our SaaS starter guide covers pricing model selection in more detail.
Step 6: Technical Feasibility (Can It Actually Be Built?)
Not every good idea is technically feasible at a reasonable cost. Before committing to development, get an honest technical assessment:
- What are the core technical risks? AI accuracy, data availability, integration complexity, regulatory requirements.
- What is the minimum viable architecture? You do not need microservices on day one. You need something that works, scales to your first 100 customers, and can be refactored later.
- What third-party dependencies exist? APIs, data providers, compliance frameworks. If your product depends on a single API that could change its terms tomorrow, that is a risk.
- What is the realistic timeline and cost? Not the optimistic estimate—the realistic one.
This is where working with an experienced technical team pays for itself. Our CTO has spent 20 years architecting enterprise systems—including platforms for Banco Itaú, Ambev, and Walmart—and the difference between a senior architect's feasibility assessment and a junior developer's guess can save you six figures. We walk through the true cost of building an MVP in 2026 with real numbers, not hand-waving.
Step 7: Unit Economics (Will the Business Work?)
The final validation step is mathematical. Even if the problem is real, the solution is compelling, and the technology is feasible, the business must make economic sense.
Calculate these numbers honestly:
- Customer Acquisition Cost (CAC): Based on your landing page test data, what does it cost to acquire one paying customer?
- Lifetime Value (LTV): Based on your pricing validation and estimated churn, what is each customer worth over their lifetime?
- LTV:CAC Ratio: Below 3:1 is a warning sign. Below 1:1 is a death sentence.
- Payback Period: How many months until you recover your acquisition cost? For startups, under 12 months is the target.
- Gross Margin: After hosting, API costs, and support, what percentage of revenue is profit?
If the unit economics do not work at scale, no amount of brilliant engineering will save you. Better to discover this before spending $30K on development than after.
When to Move Forward
You are ready to build when:
- At least 15 of 20 interviewees confirm the problem is real and painful
- Your landing page converts at 5%+ for signups or 2%+ for paid commitments
- Competitors exist but have clear, exploitable weaknesses
- People will pay a price that supports 3:1+ LTV:CAC
- A technical assessment confirms the product can be built within your budget
- Unit economics work at realistic (not optimistic) assumptions
If you hit five of six, move forward with confidence. If you hit fewer than four, iterate on the idea or pivot entirely. The validation process is not a gate—it is a compass.
From Validation to Velocity
Once you have validated demand, the goal is speed. Every week between validation and launch is a week where the market can shift, competitors can move, and your insights can go stale. This is exactly why our eight-week idea-to-revenue process exists—it compresses validated ideas into production software with paying customers in under two months.
The founders who win are not the ones with the best ideas. They are the ones who validate fastest, build fastest, and learn fastest. Validation is not a delay—it is an accelerant. It gives you the confidence to move decisively and the data to make every development dollar count.