Written by Simon, founder who shipped 4 products nobody wanted.

The 5-Step Validation Framework for Hardware & Deep-Tech Startups

Most hardware founders spend 18 months and $500,000 building a product before they talk to a real customer. Then they discover the customer doesn't have the problem they assumed, or doesn't value the solution enough to pay for it. Startup idea validation isn't a nice-to-have for hardware and deep-tech companies. It's the difference between a funded, scaling company and an expensive lesson in humility.

Software founders can ship a prototype in a week and gather feedback by Friday. Hardware founders don't have that luxury. Every design iteration has a cost. Every wrong assumption burns runway. The MIT Sloan research on deep-tech innovation makes this explicit: hardware intensity means validation and scaling carry compounding iteration costs that software simply doesn't face. That changes how you validate, not whether you validate.

If you're building hardware, biotech, or any deep-tech product, this framework gives you a sequenced, practical path to proving your idea before you've committed to a manufacturing run. Validate your idea before the bills get irreversible.

Step 1: Problem Validation, The "Why" Test

Before you think about components, materials or certifications, you need to know one thing: does this problem actually hurt people enough that they'll pay to solve it? Your core assumption isn't "I can build this." It's "enough people have this problem, feel it acutely and can't solve it with existing options."

Customer discovery interviews are your tool here. The goal isn't to pitch your solution. It's to listen. Structure interviews around the customer's past behavior, not hypothetical future behavior. Ask them about the last time they experienced the problem. Ask how they handled it. Ask what it cost them in time, money or risk. The red flag to watch for is when founders spend 70% of the interview talking. If you're explaining instead of listening, you're pitching, not discovering.

Market sizing matters here too, but do it bottom-up first. Don't pull a $50 billion TAM from a Gartner report and work backwards. Count the number of potential customers you can actually name, estimate what each would pay and multiply up. Cross-check with search volume for related terms (Google Keyword Planner and Ahrefs both work). The Harvard Business School validation framework specifically flags search intent analysis as a signal of organic demand, not just manufactured interest. You've validated the problem when at least 8 out of 10 interview subjects describe the same core pain without prompting from you.

Step 2: Solution Verification, The "How" Test

Now you can think about your solution, but not the full product. Hardware validation runs on the concept of minimum viable prototypes, and the key decision is fidelity. How much fidelity does your prototype need to generate a real signal? The answer is almost always less than you think.

Technology Readiness Levels (TRL) give you a useful map here. The Equidam framework on deep-tech valuation describes TRL 1-3 as concept and proof of concept, TRL 4-6 as prototype and pilot production. Most founders rush to TRL 5 or 6 before they've validated the problem properly. For early validation, you often only need TRL 3: a lab demonstration that the underlying technology works. You don't need a production-ready device to confirm that a customer would use it.

Digital twins and simulation tools have changed this game dramatically. If you're building an AI hardware product, you can simulate performance characteristics, show customers predicted outputs and collect intent data without fabricating a single unit. An anonymized AI hardware company in the semiconductor space used digital twin simulations to run 60 customer conversations and collect 12 letters of intent before committing to their first silicon tape-out. That tape-out cost $800,000. The validation cost $40,000 and three months. The math is obvious.

Design thinking principles push you toward rapid iteration cycles even when the physical constraints of hardware slow you down. Use Wizard of Oz prototypes, mockups and service-based simulations wherever you can. Your goal is to answer the falsifiability test: what result would prove this solution does NOT work for customers? If you can't answer that, you're not doing validation, you're doing product development theater.

Step 3: Market Viability, The "Who" Test

You know the problem is real. You have a plausible solution. Now you need to know if the market will actually move. Landing pages with a clear value proposition and a waitlist signup are still one of the highest-signal low-cost experiments available. Run paid traffic to a page that describes the outcome your product delivers (not the technology) and measure conversion rate on the email capture. Anything above 15% on cold traffic suggests real demand.

Pre-sales conversations are even better. Your goal in this step is to collect Letters of Intent (LOIs). An LOI isn't a purchase order, but it's a signed document that says a real buyer has reviewed your solution concept and intends to purchase. Getting 5-10 LOIs from credible buyers before you've built anything is one of the strongest validation signals you can show an investor.

For deep-tech, regulatory considerations belong in this step, not after product completion. If you're building a medical device, FDA pathway selection (510k vs. PMA vs. De Novo) affects your timeline by 12 to 36 months and your budget by millions. If you're building industrial hardware, CE marking, UL certification or FCC compliance need to be mapped to your go-to-market plan from day one. Founders who discover these requirements at late stage often find their unit economics are broken and their timeline is unbuildable. Map the regulatory path during market viability, not after.

Step 4: Unit Economics and Go-to-Market Validation

This is where many hardware startups fail quietly. The product works. Customers want it. But the unit economics don't work. Cost of Goods Sold (COGS) at prototype scale is always higher than at production scale, sometimes by a factor of 10x. Your validation here is confirming that a credible path to target COGS exists, and that customers will pay a price that generates a viable margin.

Run your Customer Acquisition Cost (CAC) against projected Lifetime Value (LTV) with brutal honesty. At founder-led sales stage, your CAC is essentially your own time, but that changes the moment you hire a sales rep or run paid acquisition. Map out what channels you'll use to reach customers at scale and run small tests on each. Direct sales, distribution partners and platform integrations all have radically different economics. Test them during validation, not after raising your Series A.

The 90-day founder-led sales roadmap works like this. Month one: reach out personally to 50 potential customers using your network and cold outreach. Month two: run 20 demos or site visits and push for LOIs or pilot agreements. Month three: close at least 2 paid pilots. If you can't close 2 paid pilots in 90 days of focused effort with your personal credibility behind the product, you have a signal problem that no marketing budget will fix. Get started with a structured approach to this process before you burn more runway.

Step 5: Disciplined Scaling, Validation at Scale

You have signal. Now you need to move from anecdote to data. Statistically, 5 customers telling you they love the product is encouraging. It's not proof of a market. The IESE research on European deep-tech scaleups highlights that companies which formalize their validation frameworks before scaling have significantly higher survival rates than those operating on founder intuition alone.

Securing design partners is the bridge between validation and scale. A design partner is a real company that commits to co-developing with you, provides access to their environment for testing and often contributes cash or in-kind resources. These relationships also generate the case study evidence that institutional investors require. VCs investing in deep-tech want to see at least one design partner agreement with a recognizable company, a credible TRL progression roadmap and evidence that the regulatory path is understood. Raw technology demonstrations without customer commitment aren't enough anymore.

Build a metrics dashboard from day one: number of customer interviews conducted, LOI conversion rate, pilot close rate, COGS versus target, TRL stage and regulatory milestone progress. This dashboard isn't for investors. It's your instrument panel for deciding when to pivot and when to push forward.

Common Pitfalls in Hardware Validation

Overestimating TAM without checking customer willingness to pay is the most common error. A $10 billion market means nothing if every buyer in it tells you they'd pay $50 for what you need to sell at $500 to break even. Willingness-to-pay research belongs in step one, not step four.

Confusing technical feasibility with market fit is the second trap. The fact that you can build something says nothing about whether someone will pay for it. Founders with deep technical backgrounds are especially vulnerable here because building is familiar and comfortable while customer conversations feel uncertain and exposing.

The third pitfall is delaying customer interviews until the MVP is perfect. There is no perfect. Every week you wait to talk to customers is a week you're flying blind. A sketch on paper is enough to have a productive customer conversation if you're asking about their problem, not demoing your solution.

Your 30-90 Day Timeline

Month one is problem and market validation. Conduct 20 customer discovery interviews. Complete a bottom-up market sizing exercise. Research regulatory requirements for your specific product category. Define your core falsifiable assumption.

Months two and three are solution verification and customer commitment. Build a minimal prototype or simulation at the appropriate TRL. Run 20 solution feedback sessions. Collect at least 5 LOIs. Close 2 paid pilots. Conduct a go/no-go assessment against your unit economics model.

The go/no-go criteria are simple. If customers can articulate the problem in the same language without prompting, if LOI conversion is above 30% and if at least one pilot is generating real usage data, you go. If any of those three signals are absent, you pivot the assumption, not the entire company. Startup idea validation done properly tells you exactly which assumption broke, so you can fix it without starting over.

Validation isn't about proving you're right. It's about finding out where you're wrong while you still have money to fix it. The founders who treat validation as risk reduction, not as a box-ticking exercise, are the ones who show up to Series A with real evidence instead of compelling slides. Go do the interviews. Read more on how to structure that process and then get out of your lab and in front of real buyers.