How to validate your MVP idea before writing a single line
I've built products nobody wanted. More than once. The first time, I spent three months on it. The second time, six weeks. By the third time, I'd learned to check whether anyone cared before I wrote a single line of code.
Validation is the least exciting part of building a startup. It's not glamorous. Nobody tweets about their validation process. But it's the one step that separates founders who ship successful products from founders who ship things into a void.
Here are five methods I've used. Each one costs little to nothing and takes less than a day. Do at least two of them before you open your code editor.
Method 1: The landing page test
This is the most reliable method. Build a simple landing page that describes your product as if it already exists. Headline, subheadline, a few bullet points explaining what it does, and an email signup form. No product behind it—just the page.
Then drive traffic to it. Share it on social media, post it in relevant communities, or run a small ad spend ($20-50 on Google or Reddit ads targeting your niche).
The metric that matters: conversion rate on the email signup. If 5-10% of visitors give you their email, you've got a strong signal. If it's under 2%, the positioning needs work or the problem you're solving isn't painful enough.
You can build this landing page in under an hour using the Landing Page Builder skill. Tell Claude Code to create a landing page with your headline, value props, and an email form. Deploy it on Vercel. Done.
I used this method for adworthy.ai. Put up a landing page describing an AI ad creation tool. Got 47 email signups in 48 hours without spending a dollar on ads. That was enough signal to start building.
Method 2: The waitlist
Similar to the landing page test, but with more commitment from the user. Instead of just collecting emails, ask people to join a waitlist and tell you what they'd use the product for.
The key difference: a waitlist form with a "What would you use this for?" text field gives you qualitative data. You learn not just that people want it, but why they want it and what specific problem they're trying to solve.
This changes what you build. I've had waitlist responses that completely shifted my feature priorities because users wanted something I hadn't considered. That kind of insight is worth more than any amount of market research.
Set up a simple waitlist with a form tool like Tally or Typeform, or just add a form to your landing page. Ask three questions: their email, what they'd use the product for, and how much they'd pay for it.
Method 3: The Reddit post
Reddit is brutally honest, and that's exactly what you want during validation. Find the subreddit where your target users hang out and post about the problem you're solving. Not a product pitch—a genuine post about the problem.
"I've been struggling with [problem]. I'm thinking about building a tool that [solution]. Would anyone actually use this? What would it need to have?"
Two things happen. Either people respond with "yes, I need this" and start describing their version of the problem, or they tell you why it won't work. Both outcomes are valuable. The first validates demand. The second saves you from building something flawed.
I've gotten some of my best product ideas from Reddit threads where someone described a problem and fifty people agreed it was painful. That's validation before you even create the post.
Rules for this method: be honest that you're exploring the idea. Don't fake it as market research. Redditors can smell insincerity instantly. And don't get defensive if people criticize the concept. That's the entire point.
Method 4: Cold DMs
This method takes more effort but gives you the highest-quality signal. Find 20-30 people who match your target customer profile and message them directly. LinkedIn, Twitter, or email—whatever channel they're active on.
The message is simple: "Hey, I'm building [one-sentence description]. I noticed you [reason you think they'd be interested]. Would you have 10 minutes to tell me if this sounds useful?"
Most people won't respond. That's fine. You need 5-8 conversations, not 30. And the conversations you do have will be incredibly informative. You'll learn language your customers use to describe the problem. You'll hear about solutions they've already tried and why those solutions failed. You'll get a gut sense for whether this is a vitamin or a painkiller.
If people are enthusiastic enough to schedule a call with a stranger about your concept, that's a strong signal. If you can't get anyone to respond after 30 outreach messages, the problem might not be painful enough to build a business around.
Method 5: Presell it
This is the nuclear option for validation. It's also the most convincing. Before you build anything, try to sell it.
Create a simple sales page—again, the Landing Page Builder skill makes this fast—with a "Buy now" or "Pre-order" button that charges a real amount. Not $1 as a test. An actual price you'd charge for the product.
If people pay for something that doesn't exist yet, you've validated demand as definitively as it gets. If nobody pays, you refund the few who did (if any) and move on without investing days of build time.
This method isn't appropriate for every product. It works best for tools and software with a clear value proposition. It doesn't work well for marketplaces or products that need network effects to be useful.
But when it works, it gives you something no other validation method does: customers before you write a single line of code.
Combining methods for stronger signals
I don't rely on just one method. My typical validation sprint looks like this:
Day 1 morning: Write a Reddit post about the problem. Start 5-10 cold DM conversations. Build a landing page with the Landing Page Builder skill.
Day 1 afternoon: Share the landing page in relevant communities. Set up tracking to measure signups.
Day 2: Check Reddit responses. Follow up on DMs. Review landing page conversion data.
By end of Day 2, I have enough data to make a decision. If all three channels show positive signals, I start building. If the signals are mixed, I refine the positioning and test again. If the signals are negative, I move to a different idea.
This two-day process has saved me from building at least four products that would have gone nowhere. That's months of reclaimed time.
The validation mindset
Here's the thing most founders get wrong: validation isn't about proving your idea is good. It's about finding out if your idea is bad before you invest in it. The goal is to kill weak ideas fast, not to confirm what you already believe.
If you catch yourself interpreting lukewarm responses as positive signals, stop. Lukewarm means no. You're looking for people who light up when they hear your concept. People who ask when it'll be ready. People who offer to pay before you've asked them to.
Anything less than that, and you should keep iterating on the idea or try a different one. With AI skills, building is the easy part. The hard part is building the right thing. Validation ensures you do.
Use the Guerrilla Marketing skill to amplify your validation efforts if you need more reach. And once you've validated, jump into building with the right skills from our catalog.
Frequently Asked Questions
How do I know if my MVP idea is validated?
Your idea is validated when strangers take real action on it—signing up for a waitlist, paying for a presale, or explicitly saying they'd use it and asking when it's ready. Friends and family saying "cool idea" is not validation.
How long should MVP validation take?
One to three days at most. Put up a landing page, post in relevant communities, and measure the response. If you can't generate interest in 48 hours of effort, either the idea needs work or the positioning does.
What's the cheapest way to validate an MVP idea?
A Reddit post. It costs nothing and takes 15 minutes. Describe the problem, ask if people would use a solution, and read the responses. If multiple people ask when it'll be available, you've found demand.
Should I validate before or after building?
Before. Every time. Building first is the most expensive way to discover nobody wants what you've made. Even with AI skills that let you build in days, those days are better invested in a validated idea.