Validate your business idea before spending time or money

Executive overview

Most founders build first and find customers second. That order is backwards and expensive. The minimum viable product (MVP) is the cheapest, fastest version of your idea that reveals whether customers will actually pay.

Find customers before you build anything. If nobody wants it, that's the win — you can redirect to something they do want.

What a real MVP looks like

• Tesla sold $1.5B of Model 3 pre-orders from a car no one could drive or touch
• Airbnb launched with no search, no map view, no card payments — cash only, San Francisco only
• Uber started as four links and a red logo, serving friends going to clubs
• Eventbrite was a plain HTML page; the team called customers for feedback and shipped changes
• Craigslist began as an email newsletter to Craig's friends — no website required

Two ways to validate before building

• Offer it as a service first — do manually what you plan to automate; SaaS means service before software
• Pre-sell it — go to people with the problem and ask when they'd pay; a customer with money is the only real validation
• Noah sold weights during COVID to friends, collected cash, then sourced the weights — delivery was the easy part
• Sold Imgur Pro on Reddit after emailing the founder and offering $7 per sale; 200 copies sold in a weekend
• If nobody buys, you've saved months and can ask what they actually want instead

Starting without a developer

• Identify where you already have an advantage — network, domain knowledge, a community
• No-code tools (Squarespace, WordPress, Carrd) and Google Docs can run real businesses
• No-code is a means, not the goal — the goal is proving customers want something
• Hire a developer only after you have paying customers; Noah paid $50 to launch AppSumo's first site

Common mistakes to avoid

• Building for 12 months before testing demand
• Waiting until the product is polished — the iPod and first iPhone both looked rough at launch
• Worrying about scale before you have any customers ("make it a problem")
• Confusing no-code tools with the actual work, which is finding customers

The only question that matters: can you find customers willing to pay before you build anything?