Lore

Network Effect Taxonomy: Direct, Cross-Side, and Data

Three distinct kinds of network effect, useful for diagnosing what actually strengthens (or doesn't) as a product scales:

• Direct network effect — every added user increases the value of the product for existing users on the same side (e.g. WhatsApp: more contacts on the network makes it more useful for everyone already on it).
• Cross-side network effect — two different sides of a marketplace increase each other's value: more supply attracts more demand, and more demand attracts more supply. This is the core retention/defensibility engine behind most marketplaces.
• Data network effect — engagement from users improves recommendations/matching for other, similar users, without those users directly interacting (e.g. Pinterest's recommendations improving as similar users engage). See Data Network Effect: Interest Graph vs. Friend Graph for the specific interest-graph application.

A caution on cross-side effects specifically: they are real but not unbreakable. A well-funded competitor can still buy its way past an incumbent's cross-side network effect by subsidizing both sides simultaneously — Casey cites DoorDash overtaking GrubHub despite GrubHub's earlier marketplace liquidity.

Contradictions

Cross-side network effects are often treated as a durable moat by default; this material narrows that claim — the effect protects against slow-moving competitors but not against a well-capitalized one willing to buy liquidity on both sides at once.

Из тем: Growth, PMF, and Business Model Design