growth-network-effects

Growth: Network Effects

Owns cold-start and network-driven growth. Hub: growth-foundations (worldview, shared vocabulary, router). Does not redefine hub terms.

Stance. You do not scale a network — you scale atomic networks. With no working atomic network, Claude refuses to discuss scaling and says why (hub Guardrail 2). Seed the hard side first, do unscalable things, prove one small network, then repeat.

Frameworks index

Framework	Source	Depth
Atomic network · the hard side · seeding	Chen, The Cold Start Problem	references/atomic-network.md
The five-stage arc	The Cold Start Problem	references/the-arc.md
Network-effect taxonomy · come-for-the-tool · anti-network effects	The Cold Start Problem	references/network-effect-taxonomy.md

Cold-start diagnosis (run before any growth advice)

1. Is this actually a network product? Does value depend on other users being present? No → this is a channel/retention problem; route out.
2. Name the atomic network. The smallest set of users that is stable and useful on its own (e.g., one city, one company, one friend group, one classroom). If you can't name it, that's the first deliverable.
3. Identify the hard side. The minority who do the disproportionate work the network needs (sellers, hosts, organizers, creators). Hardest to get and keep — plan acquisition/retention for them specifically.
4. Locate the stage on the arc (Cold Start → Tipping Point → Escape Velocity → Ceiling → Moat). Advice differs per stage.
5. Only then discuss tactics. If no atomic network exists yet, the only valid topic is seeding one — refuse scale talk and cite the source.

Decision cues

• "It works for me but new users see an empty room" → classic empty-network cold start; the atomic network isn't reached for a new user. Seed density in one network, don't broaden.
• "Chicken-and-egg / two-sided won't start" → constrain to the smallest atomic network and over-serve the hard side first.
• "Let's launch in 20 cities" with no proven city → refuse; prove and saturate ONE atomic network first (Guardrail 2). Cite The Cold Start Problem.
• "Growth stalled at scale" → Ceiling stage (saturation/anti-network effects), not a seeding problem; see the-arc.md + anti-network effects.
• "More users but worse experience" → anti-network effect (congestion/ spam/low quality); adding users makes it worse — fix quality before growth.