The Network Effects Assessment Model
If each new user makes the product more valuable for all users, you win
Bill Gurley uses a deceptively simple mental model for evaluating technology investments: plot customer value (Y-axis) against market penetration (X-axis). If the trend line goes up and to the right — meaning each additional user makes the product more valuable for all existing users — you have a network effect business. This single characteristic determines whether a business will create winner-take-all dynamics or compete in a commoditized market. Network effects are the strongest competitive moat because they are self-reinforcing: more users create more value, which attracts more users, which creates more value. Unlike other moats (brand, economies of scale, switching costs), network effects actually strengthen over time rather than erode. Gurley emphasizes this as his 'main stage mental model' for investment decisions — the first and most important question he asks about any technology business. The model forces clarity because most businesses that claim network effects do not actually have them. True network effects require that each marginal user makes the experience better for every existing user, not just that the company benefits from more customers.
- If the trend of customer value versus market penetration goes up and to the right, you have a network effect business.
- Network effects are the strongest competitive moat because they strengthen over time rather than erode.
- Most businesses that claim network effects do not actually have them — true network effects require each new user to improve the experience for all existing users.
- The key investment question is: can a well-funded competitor replicate your advantage? If yes, it is not a real moat.
- Map the Customer Value vs. Penetration CurveFor any business you are evaluating (your own or an investment), plot a conceptual graph with customer value on the Y-axis and market penetration on the X-axis. Ask: as more people use this product, does it become more valuable for existing users? If you are a social network and more friends join, value increases (network effect). If you are a marketplace and more sellers join, buyers get more selection and better prices (network effect). If you are a SaaS tool and more people subscribe, existing users get nothing additional (no network effect). Be rigorous — do not confuse 'more customers means more revenue for the company' with 'more customers means more value for existing customers.'Pro tipAsk specifically: what does the 100th customer get that the 10th customer did not, purely because there are now 100 customers? If the answer is nothing, there is no network effect.WarningDo not confuse supply-side economies of scale with demand-side network effects. Economies of scale reduce the company's costs; network effects increase the customer's value. They are different moats.
- Evaluate the Competitive Moat StrengthAssess whether the network effect creates a genuine barrier to competition. Gurley identifies four structural competitive advantages: network effects (strongest), economies of scale, high switching costs, and brand (weakest). Apply his litmus test: can a well-funded competitor replicate your advantage? For network effects, replication requires the competitor to rebuild the entire user base from scratch, which is extraordinarily difficult once the incumbent has critical mass. For brand alone, a competitor can outspend you. The test reveals whether your moat is structural (embedded in the product architecture) or tactical (dependent on execution that others can match).Pro tipLayer multiple moats when possible. The strongest businesses combine network effects with high switching costs and economies of scale, making them virtually unassailable.
- Assess Marketplace Health CharacteristicsFor marketplace businesses specifically, Gurley identifies several characteristics that predict success: high fragmentation of supply (many small sellers rather than a few large ones), superior user experience versus the status quo, and low friction for supplier onboarding. He also distinguishes between 'promiscuous' relationships (restaurants, hotels — customers use many different suppliers) and 'monogamous' relationships (babysitters, hairstylists — customers stick with one). Promiscuous relationships are better for marketplaces because they generate more transactions and more switching. Monogamous relationships create a marketplace that eventually disintermediates itself.Pro tipThe best marketplace opportunities have fragmented supply, bad incumbent user experience, and promiscuous customer-supplier relationships.WarningBeware of marketplaces where suppliers and buyers build direct relationships that bypass the platform after the initial match. This 'leakage' undermines the marketplace's value over time.
Uber demonstrates Gurley's network effects model perfectly. More riders attract more drivers (who want more customers). More drivers reduce wait times for riders. Shorter wait times attract more riders. The cycle is self-reinforcing, and each new participant (driver or rider) makes the platform more valuable for all existing participants. Gurley was an early investor at Benchmark and recognized this dynamic before it was widely understood.
Gurley describes how Glassdoor's founders did 'tons of unscalable things' to build initial content — including personally interviewing Cisco employees at Starbucks to get company reviews. This grassroots effort created the initial content that attracted more users, who contributed more reviews, which attracted even more users. The unscalable beginning was necessary to reach the critical mass where network effects took over.
Bill Gurley developed this assessment model through decades of technology investing at Benchmark Capital, one of Silicon Valley's most respected venture capital firms. Benchmark's portfolio includes Uber, eBay, Twitter, Snapchat, and other companies where network effects were the primary competitive advantage. Gurley's particular expertise is marketplace businesses — platforms connecting buyers and sellers — where network effects are both the most powerful and the most commonly misunderstood dynamic. He crystallized this model by studying why some marketplace businesses (eBay, Uber, Airbnb) achieved dominant positions while others with seemingly similar structures failed. The differentiator was almost always the presence or absence of genuine network effects versus simply having a lot of customers.