a16z Podcast: Getting Network Effects

Episode Summary

Executive Summary: This A16Z episode explains how to distinguish true network effects from mere growth or viral adoption, using case studies from Facebook, OpenTable, WhatsApp, Airbnb, eBay, Medium, and others. The speakers emphasize bootstrapping, localized experimentation, retention, and value creation as the real signals of defensibility, while warning that paid acquisition, branding, or high engagement alone do not prove a network effect.

Main Topics: Defining network effects vs. growth (Priority: 5/5): Network effects mean the product becomes more valuable to existing users as more users join. The speakers contrast this with viral growth, which is simply fast adoption, and argue the distinction is often confused in startup evaluation. Retention as the clearest signal (Priority: 5/5): They repeatedly use retention metrics—especially DAU/MAU—as evidence that a network is compounding value rather than merely attracting users temporarily. Bootstrapping and solving chicken-and-egg problems (Priority: 5/5): For marketplaces and two-sided networks, companies need explicit hacks to seed supply or demand first; examples include OpenTable, Facebook, Uber, and Instacart. Market-by-market and cohort-by-cohort validation (Priority: 4/5): The speakers stress small experiments and local proof points before scaling. For local marketplaces, success in one city or segment must be replicated elsewhere to indicate scalability. Different business models create different network dynamics (Priority: 4/5): The discussion distinguishes local marketplaces, global platforms, point-to-point services, data networks, and brands, arguing that not all scaling businesses have the same type of network effect. Monetization timing and marketplace economics (Priority: 4/5): They argue monetization is not necessarily in conflict with growth. In marketplaces, revenue can often be turned on earlier because transactions already exist, while consumer networks may delay monetization longer. External shocks and resilience (Priority: 3/5): The episode closes with a discussion of how macro events like 9/11 or the financial crisis can temporarily disrupt network formation, but durable networks tend to recover if they have real product-market fit and adequate capital.

Key Arguments: True network effects are about increasing user value as the network grows, not just user growth or engagement. Viral growth reflects speed of adoption; network effects show up in retention and repeated usage. DAU/MAU is a useful proxy for whether users keep returning because the product gets more valuable over time. Bootstrapping matters because two-sided businesses need a starting point on one side of the market to create the flywheel. OpenTable succeeded by starting with restaurant-side tools, then capturing diners once enough restaurants were available. Facebook’s early discipline came from launching in a tight cluster, driving high engagement, and forcing early social density through friend suggestions and directory hacks. Local marketplaces often need city-by-city rollout, while global networks like eBay can scale faster because the value is location-independent. Airbnb and WhatsApp both show that apparent overnight success often followed years of iteration, targeted marketing, and product-market-fit work. Paid acquisition can coexist with network effects, but if growth disappears when spending stops, the business is probably not network-effect-driven. Monetization should be evaluated in context: marketplaces can often monetize earlier because transactions already flow through them. Brand strength alone is not necessarily a network effect; it may create demand, but it does not always create user-to-user value compounding. Data can create a newer form of network effect when the aggregate corpus improves recommendations or operational decisions for users over time.

Data Points: Facebook retained activity among cumulative historic users: 53% - Jeff described an early Facebook metric showing 53% of all historic users were active the previous day, indicating strong network retention. Facebook DAU/MAU early chart: 52% to 55% to 57% - Anu cited Facebook’s rising daily-active-to-monthly-active ratio over the first 18 months as evidence of network effect. Facebook user growth chart: 800M+ MAUs - Used as an example of viral growth, but the speakers noted this alone does not prove a network effect. Facebook retention chart: 45% to 57% - Another retention metric discussed as evidence that users stayed engaged in the network. Facebook early retained-user metric cited by Mark Zuckerberg: 53% - Jeff recalled Zuckerberg saying 53% of cumulative users were active yesterday, surprising him at the time. Facebook first launch cluster: Harvard - Facebook launched in a tightly controlled cluster before expanding to other campuses. Facebook target penetration at Harvard: 80% - The team aimed to get at least 80% of Harvard signed up before broader rollout. Facebook engagement target: 50%+ daily return rate - The company wanted more than half of users returning daily to validate product-market fit and network value. Friends added in Facebook onboarding: 10 friends in 14 days - A key early heuristic: users who connected with 10 friends in 14 days were much more likely to return. OpenTable early restaurant sales pace: 3-4 new restaurants per month - A salesperson could add only a few restaurants monthly in early city-by-city rollout. OpenTable later sales pace: ~20 restaurants per month - Years later, the same reps were adding restaurants much faster due to network pull. OpenTable restaurant economics example: $50 reservation value; 30% margin; $15 restaurant earnings; $1 platform fee - Illustrated how the service created strong economic value for restaurants once reservations flowed through the network. OpenTable initial pricing for tools: $200 - Restaurants were first sold a tool suite that had utility even before network effects kicked in. OpenTable sales deployment: City-by-city, including San Francisco, Miami, Tokyo, Munich - The marketplace had to be built market by market because local density mattered. Airbnb bootstrapping timeline: ~36 months - Anu said it took roughly three years to build critical mass and observe network effects. Airbnb initial demand pattern: Big conferences/events - Airbnb targeted sold-out travel periods to create early demand. Airbnb cereal-box fundraising hack: Obama and McCain cereal boxes - The founders reportedly used novelty cereal boxes to raise money during the election cycle. Pinterest early product-market-fit period: 2-3 years - Jeff cited Pinterest as taking years before it started working and then accelerating quickly. Medium traffic analysis: Non-viral posts had Medium as biggest source - Used to argue that Medium was building direct network value rather than only social-driven traffic. eBay early collectible demand: 5 duck decoys found online - Jeff used this anecdote to show how niche demand on eBay proved the value of aggregation. Uber/Lifeline subsidy strategy: Subsidize enough cars to ensure service quality within 5 minutes - Used to explain an early marketplace hack that jump-started rider adoption. Instacart markup strategy: Marked up initially; later more than half of deliveries not marked up - The company used markup to bootstrap demand and later converted grocers into direct partners. 9/11 impact on eBay volume: 40-50% overnight drop - Example of an external shock temporarily suppressing network activity and new users. Financial crisis impact on restaurant consumption: 15% drop - OpenTable saw restaurant consumption decline sharply during the 2008 crisis. OpenTable capital survival window: ~5 years unopposed - Competitors failed during the funding drought, allowing OpenTable to build its network effect.

Pivotal Quotes: "as more users join the platform, it's more valuable to existing users" — Jeff: Direct definition of network effects at the start of the discussion. "we actually told it, put a slide up to our please last year, network effects equals moats" — Jeff: Explains why the firm values network effects as a source of defensibility. "come for the tools, stay for the network" — Jeff: Summarizes OpenTable’s strategy of using utility products to seed the network.

Implications: Founders should prove retention and value compounding early, not just fast signups. Investors should test small markets, watch for organic pull, and separate network effects from paid growth, branding, or temporary virality.