How Letterly grows

Recording starts from a widget with one click; click stop and the text is ready (moving to real-time transcription). Favorited rewrite styles appear at the bottom of each note for one-tap rewrites, and several rewrites can be swiped between. Alongside a share button, added a one-click copy button, plus one-tap send to Google Docs or Notion and tap-to-append voice to an existing note.

Noticed early apps chaining ChatGPT and OpenAI transcription to turn speech into text; all were rough prototypes he needed but couldn't use. Stopped the team's other startup to build a well-implemented version. Deliberately chose a 'validated business model' and aimed to make it much better and simpler, instead of chasing a novel idea as in past startups. Advises choosing a newish idea validated by someone else and not spending money validating completely new ones.

Every screen and interaction is judged on removing friction. Simplicity is treated as a separate, expensive feature and revisited as the product grows. Three iteration loops: (1) draw prototypes, discuss, simplify; (2) during the build, stop before release and ask if it's as simple as intended, postponing features promised to users by a few weeks if needed; (3) after release, watch usage and questions and simplify again. When choosing work, weighs building a requested feature vs making an existing one easier. 'Thousands of hours' of rebuilds and prototypes went into the UX.

In the early days, asked customers 'like crazy' why they used Letterly for this instead of ChatGPT.

Spends roughly $100K-200K per month on advertising (figure garbled) against $250K monthly revenue, as a deliberate growth investment. Channels and creative not discussed.