How To Get Your First Users

Most people are not early adopters, and few want to be a startup’s first paying customer. The first product needs to work for a small group and be easy to adapt. Some people actively try new products, while others will take a chance when they have a pressing problem, as the speaker did when paying a startup to handle billing and a public endpoint for an inference API.

Finding early users is a search problem: look for people who try new things or have a problem your product can solve. Charge real money early because paying customers tend to give more useful feedback than free users. Use direct outreach, launch early to help users find you, and study how they make decisions and why they trust you.

Test pricing, landing pages, onboarding, and features quickly while talking with early users. Churn is not necessarily a crisis: when relationships are personal, founders can address problems, and many potential users have not heard of the product yet. Startups can run imperfect experiments without attracting the attention that a large company might.

The speaker contrasts personal software spending, which they estimate at about $150 per month, with corporate tools that can each cost more. They say AI costs can make consumer apps difficult to sustain through ads or subscriptions. Some AI founders therefore begin with prosumers, businesses, or users such as doctors who have high advertising value.

The speaker compares a startup to an evolutionary tree: a simple starting product encounters market pressures and develops in different directions. Tesla’s Roadster attracted buyers willing to pay $150,000 for an impractical early electric car, and their preferences helped shape later vehicles. The first product should be simple enough to adapt, since its eventual form depends in part on where and with whom it begins.