TheShortlist^™

Founders pitch their moat by leading with one thing. Network effects. Data flywheel. Brand. Regulatory licensing. The single defensible advantage they think will hold.

We tested the moat-stack hypothesis across 1,022 venture-backed companies. 55 hand-researched anchor companies (42 winners, 13 named failures including Theranos, WeWork, Magic Leap, Quibi). 967 Y Combinator alumni with resolved outcomes, scored against the Hamilton Helmer 7 Powers framework. Hand-coding for the canonical winners and failures. Engine-coded for the YC validation set.

What collapsed in each failure

Failures fell into three groups. Three companies had no defensible moat at funding. Four claimed one moat that eroded. Six claimed two or more moats that did not survive contact with the market.

No moat at funding

• Jawbone. Industrial design as positioning. Apple Watch and Fitbit erased the design premium in eighteen months. No software, network, or data behind it.
• Magic Leap. Custom optics IP that turned out to be standard waveguide tech. Two and a half billion dollars spent with no other defensible advantage.
• Webvan. Logistics scale only. Asset-light competitors proved the same market could be served without the capital-intensive infrastructure.

One moat that eroded

• Theranos. Proprietary microfluidic technology. The technology never worked.
• Juicero. Hardware lock-in via subscription packets. The packets were squeezable by hand. One Bloomberg article ended it.
• Quibi. Hollywood content relationships. Free competitors (TikTok, YouTube Shorts) had network effects and habit Quibi did not.
• Fast. Network effects between merchants. Never reached the scale required for the effect to bite, and Shopify's Shop Pay had the install base.

Two or more moats that did not hold

• WeWork. Brand, community network effects, and scale of locations. Real estate sublease arbitrage was always commoditizable, and the brand could not protect against competitors copying the offering. Three claimed moats; none durable.
• Bird. First-mover scale and brand among urban riders. Both eroded once Lime, Spin, and city regulators arrived. Two claimed moats, both eroded fast.
• Convoy. Two-sided trucker-shipper network and lane-pricing data. Network never reached liquidity to defend against Uber Freight. Two moats, both shallow.
• Olive AI. AI/automation IP and hospital relationships. Both replicable; AI commoditized fast and the relationships were not exclusive.
• Pets.com. Brand (sock puppet) and first-mover scale. Brand without unit economics; scale without margin.
• Better.com. Software automation, customer brand, and refi market scale. None of the three survived rate hikes.

The lesson: claiming two moats is necessary but not sufficient. The question that separates the cohorts is durability — whether each claimed advantage actually compounds and resists imitation.

Winners always layered moats

The pattern on the winning side is the inverse. Every winner could name two or more defensible advantages at funding, and the moats compounded with each other rather than substituting.

• Airbnb. Two-sided network effects plus the trust system (verified profiles, reviews, secure payments) plus brand. Three moats.
• Stripe. Developer brand plus API simplicity plus integration breadth plus regulatory licenses globally. Four moats.
• Snowflake. Architecture moat (separation of compute and storage) plus customer data lock-in plus data marketplace network effect. Three moats.
• Coinbase. Regulatory licensing plus brand trust plus liquidity network effects. Three moats.
• NVIDIA. Architecture IP plus manufacturing relationships plus CUDA software ecosystem plus brand among gamers and developers. Four moats.

Why does this happen?

A single moat is exposed to a single failure mode. When the moat erodes, the company has nothing left. Two or more moats means each one can compensate when the others come under pressure. WeWork's brand could not save it when the unit economics broke. Airbnb's trust system kept hosts on the platform when listings briefly slowed.

The pattern is not that winning companies got lucky and stumbled into multiple moats. It is that they architected the company around layered defensibility from the start. That is a planning choice, not an outcome.

The framework: building a moat stack that actually holds

The strategy literature has been clear on this for forty years. Michael Porter's Competitive Strategy (Harvard Business School, 1980) introduced the modern concept of structural advantage. Hamilton Helmer's 7 Powers: The Foundations of Business Strategy (2016, taught at Stanford GSB) refined Porter into seven categories of durable defensibility:

1. Scale Economies. Cost advantage that grows with size. AWS, Walmart.
2. Network Economies. Each new user makes the product more valuable to others. Airbnb, Visa.
3. Counter-Positioning. A business model the incumbent cannot copy without destroying their own. Netflix vs. Blockbuster.
4. Switching Costs. Customer pain to leave. Salesforce, Snowflake.
5. Branding. Affective association that lets you charge a premium. Apple, Hermès.
6. Cornered Resource. Exclusive access to an input. Pixar's animation team. ASML's lithography monopoly.
7. Process Power. Embedded operational excellence that compounds over time. Toyota, Costco.

The pattern in our dataset is not that winners had random combinations of these. It is that winners stacked complementary types. Stripe's developer brand (#5) compounded with API simplicity creating switching costs (#4) compounded with regulatory licenses globally (a Cornered Resource, #6). Each one made the others harder to attack. Failures had a single power that, once eroded, left the company exposed.

What this means if you are applying to a fund

When the partner reading your deck asks about your moat, they are asking which one of two failure modes you are. Naming one moat puts you in the failure cohort by default. Naming two or more, drawn from different rows of Helmer's framework, puts you in the winning cohort, conditional on the moats being credible.

1. Identify two moats from different categories. Two from the same row do not stack — two flavors of brand are still one defense. Combine across rows: network economies plus switching costs. Process power plus a cornered resource. Counter-positioning plus brand.
2. Stress-test each moat against twelve months of well-funded competition. If a competitor with infinite resources tried to copy you in twelve months, which moat would they fail to replicate? If the answer is none of them, the moats are weak.
3. Claim only what you can defend. Inflated moat claims read worse than honest ones. If you can credibly claim one moat right now, name it as one and articulate the second moat you are building toward, with the milestone that will prove it.

The single-moat trap is not a question of cleverness. It is a question of structural awareness. Founders who survive in venture pay attention to the second layer of defensibility before the first one is even tested.

Further reading. Hamilton Helmer, 7 Powers: The Foundations of Business Strategy (Deep Strategy, 2016). Michael E. Porter, Competitive Strategy (Free Press, 1980). Aswath Damodaran (NYU Stern), The Little Book of Valuation, ch. 4 on competitive advantage. Pat Dorsey, The Little Book That Builds Wealth (Wiley, 2008).