Validating a Startup Idea Before You Write Any Code
When you have an idea for a product, the instinct is to start building. Building feels like progress, and it is the enjoyable part.
Silicon Lab Team30 Sept 20266 min read
When you have an idea for a product, the instinct is to start building. Building feels like progress, and it is the enjoyable part. But a working product for a problem nobody has is an expensive mistake. Validation is the work of finding out, cheaply and quickly, whether the problem is real and whether people will act to solve it. You can do most of it before writing a line of code.
Start with the problem, not the solution
Founders often fall in love with a solution: an app, a device, a platform. Customers do not care about your solution. They care about their problem. So begin by describing the problem clearly:
- Who has this problem? Be specific: "small bakery owners who take custom cake orders by phone" rather than "small businesses".
- What exactly is the problem, in their words?
- How do they deal with it today?
- What does it cost them now, in time, money, stress or lost customers?
If you cannot answer these, your first job is to find out, not to build.
Write down your assumptions
Every idea rests on assumptions. Make them explicit so you can test them:
- This group of people has this problem.
- The problem matters enough that they want it solved.
- Their current solutions are unsatisfactory.
- They would switch to, or pay for, a better solution.
- You can reach them.
Rank these by risk: which, if wrong, would sink the idea? Test the riskiest first.
Run problem interviews
The fastest way to learn about a problem is to talk to people who have it. The purpose of a problem interview is to learn, not to sell.
Guidelines:
- Do not pitch your idea. As soon as you describe your solution, people start being polite about it.
- Ask about past behaviour, not future intentions. "Tell me about the last time this happened" produces facts. "Would you use an app that…?" produces guesses.
- Ask for specifics. What did they do? How long did it take? What did it cost? What tools did they use?
- Listen more than you talk. Follow interesting details with "Why?" and "Tell me more about that".
- Talk to strangers, not only friends and family.
Useful questions:
- "Walk me through the last time you had to deal with this."
- "What was the hardest part?"
- "What have you tried to solve it? What did you like or dislike about those?"
- "How much time or money does this cost you?"
- "If you could change one thing about how this works, what would it be?"
Take notes immediately after each conversation. Look for patterns across interviews rather than drawing conclusions from a single enthusiastic person.
What counts as evidence
Not all positive signals are equal. Learn to tell strong evidence from weak.
Strong evidence — people giving something up:
- They are already spending money or significant time on a workaround.
- They have built their own solution with spreadsheets, paper or improvised tools.
- They agree to a follow-up meeting, introduce you to colleagues, or pre-order.
- They join a waiting list and respond when you contact them.
- They ask when they can have it.
Weak evidence — words that cost nothing:
- "That's a great idea."
- "I would definitely use that."
- "You should build it."
- Likes and positive comments on social media.
Compliments are pleasant but they are not demand. Commitments of time, money or reputation are.
Cheap tests you can run in a week
Once interviews suggest the problem is real, test whether people will act on a solution. None of these need a finished product.
- Landing page. Describe the solution and its benefit on a simple web page with a sign-up or pre-order button. Share it with your target group and see whether people act.
- Concierge test. Deliver the service manually to a few customers. If your idea is an app that plans meals, plan the meals yourself by email. You learn exactly what customers value.
- "Wizard of Oz" test. Present what looks like an automated product, but do the work by hand behind the scenes. It tests demand before you invest in automation.
- Clickable mock-up. Build screens in a design tool and watch potential users try to complete a task. You will find confusing steps before any code exists.
- Pre-sales or deposits. The strongest test of all: ask people to pay, or commit to pay, before it exists.
Decide in advance what result would count as success for each test. Otherwise, any result can be read as encouraging.
False positives to watch for
These signals can mislead founders:
- Friends and family. They want to support you, so they will be positive.
- Hypothetical answers. "Would you use…?" questions invite optimistic guesses.
- Excitement at a demo. People enjoy seeing something new; that is not the same as needing it.
- Interest in you rather than the product. People may be kind because they like you or admire your effort.
- Leading surveys. Questions that suggest the answer produce the answer you wanted.
- Counting sign-ups without follow-up. A sign-up that never responds to your next message is weak evidence.
Decide: continue, change or stop
Set your decision criteria before the tests, then be honest about the results.
- Continue if strong evidence supports your riskiest assumptions. Build the smallest version that solves the core problem.
- Change direction if the problem is real but your solution, audience or approach is not right. This is common, and it is progress.
- Stop if the problem is not real or not important enough. Finding this out in weeks rather than after a year of building is a success, not a failure.
A validation checklist
- The target customer is described specifically.
- The problem is written in the customer's own words.
- Assumptions are listed and ranked by risk.
- You have held problem interviews with people outside your personal circle.
- Interview notes are recorded and compared for patterns.
- At least one test asked people to commit time, money or reputation.
- Success criteria were set before each test.
- The decision to continue, change or stop is based on evidence, not enthusiasm.
The takeaway
Validation is about learning what is true before you invest heavily in what you hope is true. Start with the problem, talk to real people about their past behaviour, look for commitments rather than compliments, and run cheap tests with clear success criteria. The code can wait until you know it is worth writing.