Where great startup ideas come from: lessons from Airbnb, Coinbase, and Stripe

Executive overview

The best startup ideas rarely look good on paper. Airbnb, Coinbase, and Stripe each entered markets with existing players, faced near-universal skepticism from investors, and turned out far larger than their founders anticipated.

Three patterns connect all three: the existing product was bad enough that a 10x improvement was achievable, the founders personally experienced the pain, and most experts told them it was a terrible idea.

Contrarian founders who lived the problem — not market researchers — find the best ideas.

Airbnb: solving a real problem, not inventing a category

• VRBO and Couchsurfing already existed; Airbnb was seen as a third-rate entrant to a crowded space.
• VRBO didn't facilitate payments — hosts and guests had to exchange cash or wire transfers between strangers.
• VRBO charged hosts to list, which throttled the supply side; inventory is everything in a marketplace.
• Brian Chesky discovered the payments insight only after using his own product and forgetting to bring cash.
• Charging for home-sharing was seen as crass — the Couchsurfing ethos treated commercial models as impure.
• Investors advised building the network first and monetising later; doing the opposite was considered suicidal.

Coinbase: buying Bitcoin shouldn't require a money order

• Circa 2011–2012, buying Bitcoin meant sending money orders to foreign countries or trusting Mt. Gox — which routinely lost funds.
• Mt. Gox had a reputation for getting money in but never letting it out; it was originally a Magic: The Gathering card exchange.
• Brian Armstrong's YC application focused on P2P transfers; the simple insight — buy Bitcoin with a debit card without getting hacked — was enough to underpin a hundred-billion-dollar company.
• He didn't fully believe in the buying-Bitcoin angle himself at the start; the core insight emerged after launch.
• Coinbase's early product was a single page: enter amount, press buy. No graphs, no social features.
• Bitcoin was seen as a tiny, bubble-prone, fraud-adjacent market; a US-based exchange required a bank deal that seemed impossible to obtain.
• Armstrong had a near-impossible time raising on demo day; quitting on that signal would have killed the company.

Stripe: making payments developer-first

• Pre-Stripe, getting a merchant account involved paper applications, faxes, personal guarantees, fraud checks, and 30-page contracts — "applying for a mortgage."
• The Stripe founders saw every YC peer struggling with the same Authorize.net pain and recognised a pattern worth solving.
• Their insight: the real customer was the developer, not the business owner who signed the contract.
• They built obsessively developer-friendly documentation and a beautiful website — "if Steve Jobs designed a payments company."
• An invite-only beta made access a status signal; every developer wanted in before they'd even used it.
• Stripe launched as the most expensive option in the market and was still in desperate demand — the clearest signal of product-market fit.
• Objections: two teenagers wanting to handle hundreds of millions of dollars; required a Wells Fargo banking deal; PayPal was one step away from copying it.

Three takeaways for founders

• Ignore existing products in the market if you and your peers are users who hate the experience.
• Discount expert skepticism when those experts have no personal experience with the problem.
• Don't anchor on your initial market size estimate — transformational products create use cases nobody predicted, and all three founders consistently revised their market sizing upward after launch.