Andrew Chen

A dense graph you can actually fill: a city, a campus, a profession. 'Everyone' is how networks freeze. The wedge is the niche.

1. 01Pick the smallest network that is still useful when complete.
2. 02Map who must be present for the product to work at all.
3. 03Prefer graphs you can reach on foot, not 'global from day one'.
4. 04If there is no atomic network, you may not have a network product.

Build the hard side first, or the side that creates liquidity. Features that do not add to the loop are hobbies.

1. 01Draw the loop on paper before the roadmap.
2. 02Make the first successful circuit possible for a tiny group.
3. 03Instrument the cycle time of the loop.
4. 04Delay anything that does not help the first graph come alive.

Seed the atomic network by hand. Concierge supply. White-glove demand. Pretend scale is a later movie.

1. 01Recruit the hard side one by one.
2. 02Be present in the analog version of the market.
3. 03Do not announce a network that is empty — it advertises death.
4. 04Write the city-by-city (or graph-by-graph) playbook.

Launch a wedge, not the universe. Uber did not 'launch transportation'. It filled a map.

1. 01Turn on a geography or a vertical that can succeed alone.
2. 02Use unscalable ops until the loop kicks.
3. 03Celebrate completed matches, not downloads.
4. 04Keep the story local and specific.

Copy the playbook to the next graph. Defend density. Paid is a way to fill a loop, not a replacement for one.

1. 01Hire or write SOPs for the seeding work.
2. 02Watch for the loop slowing — that is the real churn.
3. 03Add viral artifacts only if they serve the same job.
4. 04Read retention as a network health metric.