By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:00 – 3:09
Defining Network Effects and Why They Matter
Sonal introduces the guests and the internal debate behind the deck. Anu defines a network effect as a platform becoming more valuable to existing users as more users join, and Jeff explains why network effects equal moats.
- •Anu's definition: as more users join, the platform is more valuable to existing users
- •The telephone as a hardware origin of network effects
- •Jeff says network effects equal moats and make a business defensible and less exposed to price pressure
- •The firm debates internally whether a given company has a network effect
- 3:09 – 5:48
Viral Growth vs. Network Effect: Facebook's Retention
Anu separates viral growth (speed of adoption, often zero acquisition cost) from a true network effect. Facebook's daily-to-monthly active ratio is the example, backed by Jeff's story of Mark Zuckerberg's 53% daily metric.
- •Viral growth is speed of adoption and usually implies near-zero acquisition cost
- •Facebook focused on retention, measured as daily actives divided by monthly actives
- •The ratio rose from 52% to 55% to 57% in the early period
- •Jeff recalls Zuckerberg telling him in 2006 that 53% of all historic users were active the day before
- 5:48 – 9:04
Bootstrapping: Facebook's Clustered Launch and Growth Hacks
Sonal asks how companies build a network before an effect exists. Anu describes Facebook's disciplined launch at Harvard, its 10-friends-in-14-days insight, and Jeff adds the directory hack and the chicken-and-egg problem.
- •Facebook wanted about 80% of Harvard signed up and over 50% daily engagement before launching at Stanford
- •Relationship status drove engagement, and 10 friends in 14 days predicted return
- •Zuckerberg pre-populated Harvard directories so new users weren't alone
- •Nearly every marketplace had a hack to solve the chicken-and-egg problem
- 9:04 – 11:10
OpenTable: Come for the Tool, Stay for the Network
Jeff explains how OpenTable started on the restaurant side with a $200 suite of tools, sold one restaurant at a time. Once diners began booking, the flywheel took off, but it had to be rebuilt city by city.
- •Entrepreneurs should have a theory on how to bootstrap and which side to start with
- •OpenTable began with restaurants because consumers had no use for it without selection
- •Early salespeople sold three or four restaurants a month; later reps sold up to 20
- •Each city had to be built separately, unlike eBay, which was nearly global from the start
- 11:10 – 13:47
Local vs. Global Launches: WhatsApp's Small Focus Group
The discussion turns to whether to go global fast. Anu argues you need to measure product-market fit close to users, using WhatsApp's start in a small Russian community in West San Jose as an example.
- •Local marketplaces are slow but very hard to displace once built
- •The hunch to go global fast must be weighed against measuring fit
- •WhatsApp began as a status feature tested with about 50 friends
- •Users turned it into a way to message family in Russia, and features followed that behavior
- 13:47 – 16:54
Airbnb's Slow Start and the Myth of Overnight Success
Sonal asks about Airbnb, which had no existing behavior to build on. Anu describes its non-viral start, event-targeted marketing and cereal boxes, and Jeff notes that overnight successes are rare, citing Airbnb and Pinterest.
- •Airbnb had no viral growth because letting strangers into your home is a hard sell
- •It targeted events when hotels sold out and sold themed cereal boxes to fund the company
- •Anu says it took almost 36 months to build critical mass
- •Jeff says Pinterest saw nothing for two to three years before product-market fit
- 16:54 – 20:28
Case Studies in the Data: Facebook, Angry Birds, Medium
Anu walks through charts that separate growth from network effect. Facebook's rising retention counts as a network effect, Angry Birds shows viral growth without one, and Medium's direct traffic to non-viral posts is an early sign.
- •Rising user counts show speed of adoption, not a network effect
- •Rising usage as users grow, such as daily-to-monthly actives, is the sign
- •Angry Birds went viral but players gain no extra value from other players
- •Medium's non-viral posts get most of their traffic directly, suggesting matching of readers to writers
- 20:28 – 23:52
Reading Early Micro Signals and Small Experiments
Jeff says early calls rest on micro metrics from a small market. He describes his OpenTable diligence on San Francisco and the eBay collectibles story, and advises running smaller experiments to validate.
- •Macro reads are hard early, so look at micro behavior in the first market
- •OpenTable's San Francisco data showed rising sales productivity, reservations per restaurant, and falling attrition
- •That gave the board confidence to open many new markets despite slow profitability
- •The eBay collectibles and decoy duck story illustrates small-scale behavior signaling larger potential
- 23:52 – 25:36
Replicating Beyond the Silicon Valley Bubble
Sonal raises the worry that what works in San Francisco may not work elsewhere. Jeff says that for local businesses they look for replication outside the home market, using Cherry car washing as an example.
- •Local businesses should replicate outside their initial market to show scalability
- •Cherry worked near Sand Hill Road, but Jeff questions Minneapolis in winter
- •The test is whether it works outside an affluent or convenience-focused environment
- •Sonal notes a city is the original network effect
- 25:36 – 28:44
On-Demand, Ride Sharing and the Hacks That Jumpstart Marketplaces
Anu distinguishes supply-side economies of scale in ride sharing from network effects, noting shared rides may have network effects. Jeff describes Uber's subsidy hack and Instacart's markup hack.
- •Ride sharing is point to point and mainly shows supply-side economies of scale
- •Shared-ride products like Lyft Line and Uber Pool could have network effects
- •Uber subsidized cars until demand made subsidies unnecessary
- •Instacart marked up prices early, then used volume to cut deals with grocers
- •In a two-sided network, the key question is who shows up first
- 28:44 – 32:26
Paid vs. Organic Acquisition and ROI
Sonal asks whether paid marketing is bad for a network business. Anu says it depends on ROI (CAC vs. LTV) and whether growth persists when spend is reduced. Jeff notes many top networks spent almost nothing on acquisition.
- •Airbnb found demand cheap to bootstrap but paid to acquire hosts
- •CAC versus LTV determines whether paid acquisition makes sense
- •The red flag is growth that slows when marketing is dialed back
- •Facebook, WhatsApp, Medium and OpenTable spent little on acquisition
- •eBay spent heavily, but the extra spend produced users who showed network behavior
- 32:26 – 35:20
When to Monetize: Dispelling the Growth vs. Revenue Dichotomy
Anu and Jeff discuss monetization timing. Marketplaces tend to monetize earlier since transactions already flow through them, and Jeff calls growth versus monetization a false dichotomy.
- •Anu says investors don't push monetization but companies should plan for it
- •Facebook monetized late, while marketplaces tend to monetize earlier
- •Payments can improve the marketplace experience, as with OfferUp
- •Jeff says some companies shifted to smart growth with improved profitability and similar growth
- •Jeff thinks there are more reasons to monetize early than not
- 35:20 – 41:02
Audience Q&A: OpenTable's Competition and Branded Marketplaces
Jeff explains how OpenTable survived while two direct competitors failed during the funding winter, and how Grubhub and HomeAway were roll-ups. He then answers a question on private labeling, saying they branded everything and value branded marketplaces above white label.
- •OpenTable ran unopposed for five years because its competitors ran out of money
- •Grubhub and HomeAway became roll-ups in local markets
- •OpenTable's growth hack was a forced redirect from restaurant pages to its own page
- •Fragmented suppliers lack power, unlike Fandango's few large theater chains
- •Branded marketplaces sell customers, while white label sells a utility
- 41:02 – 48:09
Brands, Data Network Effects, and External Shocks
Answering audience questions, the guests discuss whether brands like Nike have network effects, which Jeff doubts. They also cover data and database network effects, and how shocks like 9/11 and the 2008 financial crisis affected eBay and OpenTable.
- •Anu places brands closer to Sarnoff's law than Metcalfe's, and Jeff would not call brand a network effect
- •Jeff says brand popularity brings diseconomies, citing Disney and Nike's hot-and-cold cycles
- •More startups rely on data or database network effects, such as Pinterest's signals
- •Anu offers a B2B shipping example where aggregated data produces better recommendations
- •After 9/11 eBay volume fell 40-50% but recovered in weeks; OpenTable's market dropped 15% for about 15 months
