Gamma

Gamma's shared decks and sites bring in free search and LLM traffic. Gamma's presentations and websites get shared, indexed by search engines and referenced in ChatGPT and Claude answers, which gives it personal and social viral loops. It grew with little paid acquisition and raised an a16z Series B at a reported valuation above $2 billion.

Phil Carter's summary: "all of that is free user acquisition that the company doesn't have to spend money on." He also noted Reddit over-indexed as an LLM source, though less than it did 6 to 12 months earlier. The lesson is to build a product whose output is a shareable artifact.

Lenny's Podcast▶ 1:13:00

Gamma chose cheaper long-tail models and was profitable six months after going viral. Gamma went viral in March 2023 and was profitable six months later, helped by model choices that kept costs low. A cheaper, faster model was good enough for its price-sensitive users early on.

Carter's framing: the product experience depends on output quality, speed and cost. If peak performance matters and users pay premium prices, use the best model instead. Because API prices are falling, plan for 6 to 12 months out. Host David Barnard pushed back that optimizing cost too early can cost you the best experience.

Lenny's Podcast▶ 1:17:16