Give the product first to the best-connected people in your niche, and let their recommendation do the launch.
On launch day, 5 May 2003, LinkedIn's co-founders sent their first 350 invitations to people in their own networks, and the network grew from there through members' invitations. LinkedIn had 4,500 members at the end of its first month and 81,000 by December 2003 (Vator, 26 August 2016).
Google did something similar with Gmail in 2004, opening the beta on 1 April to about 1,000 opinion leaders who could invite friends and family. Within two months, invitations were selling on eBay for up to $150 (Spokesman-Review, 6 June 2004).
B2B tools and services sold into tight professional communities, and consumer products with a clear set of tastemakers.
Who joins first decides how a product spreads. The Cold Start Problem shows that LinkedIn seeded well-connected professionals rather than celebrities, and that a launch works when it starts in one small, dense network. Granola, the AI notepad, spread early among venture investors, who are among the most connected recommenders in technology, and raised $125m at a $1.5bn valuation in March 2026 (Tech.eu, 25 March 2026).
Reported. LinkedIn from Vator (26 August 2016), Gmail from the Spokesman-Review (6 June 2004) and Granola's Series C from Tech.eu (25 March 2026). Background. Andrew Chen, The Cold Start Problem (2021), on atomic networks and invite-only launches.
Cost to test. Free product for 50 people and the founder's time for two months.
Kill it if after 60 days fewer than one in five of them use it every week, or none has made an introduction.
Slow and unscalable by design: it is the ignition, not the engine. Connectors talk to each other, so a bad experience spreads as fast as a good one. Never present a free user as a paying customer or an endorser without their agreement.
The Waitlist Launch, The Roundtable Series, Podcast as Business Development
A 30-minute call first. If it fits, two calls in one week and a written growth plan you keep.