Moats and Defensibility: How Tech Companies Build Durable Competitive Advantages

Companies Using This Strategy

The seven canonical moats (Helmer framework)

Hamilton Helmer's '7 Powers' framework identifies seven structural advantages: (1) **Counter-positioning**: incumbent has business model that can't be replicated without cannibalizing existing business. Example: Netflix vs. Blockbuster — DVD-by-mail/streaming would have cannibalized late fees. (2) **Scale economies**: per-unit costs decrease with scale. Example: AWS, Amazon retail. (3) **Switching costs**: customers face high costs to switch providers. Example: Salesforce CRM data and integrations. (4) **Network effects**: product becomes more valuable as more users adopt. Example: Visa/Mastercard, Facebook. (5) **Branding**: customers pay premium for branded products independent of functional differentiation. Example: Apple, Hermès. (6) **Cornered resource**: exclusive access to scarce resource. Example: rare earth mineral deposits, exclusive content rights. (7) **Process power**: company has process knowledge that competitors can't replicate. Example: Toyota manufacturing, SpaceX rocket design. The framework is structurally complete — most claimed moats fit into one of these categories. Companies often combine multiple powers; e.g., Stripe combines scale economies, branding, and process power.

Network effects: the canonical modern moat

Network effects are the most-cited modern moat. The mechanism: each new user makes the product more valuable for existing users. Types of network effects: (1) **Direct network effects**: users interact directly with other users. Example: Facebook, WhatsApp. (2) **Indirect network effects**: users on one side benefit from users on other side. Example: marketplaces (Airbnb, Uber), platforms (iOS app store). (3) **Data network effects**: more users produce more data that improves product. Example: Google Search, Netflix recommendations. (4) **Social network effects**: users want to be where their network is. Example: LinkedIn, Twitter. Andrew Chen's 'Cold Start Problem' systematizes network effect dynamics. Key insight: network effects don't activate until critical mass; the 'cold start' phase requires substantial investment without proportional value. Network effects are not all equal. Some are stronger than others (payments networks > social networks > messaging networks). Some are more defensible (multi-product networks) than others (single-product networks).

Switching costs: the B2B SaaS moat

Switching costs are the dominant moat in B2B SaaS. The mechanism: customers face real costs (data migration, training, integration rework) to switch providers. Switching cost sources: (1) **Data accumulation**: customer data exists in incumbent system. Migration requires technical work and may lose historical analytics. (2) **Integration footprint**: incumbent product integrates with customer's other systems. Switching requires rebuilding integrations. (3) **Training and skill investment**: users have learned incumbent product. Switching requires retraining. (4) **Process integration**: customer processes are designed around incumbent product. Switching requires process change. (5) **Multi-product expansion**: incumbent product family extends beyond original purchase. Switching requires evaluating alternatives for multiple products. Switching costs accumulate with customer tenure and product breadth. Salesforce customers using multiple Salesforce products with customizations face high switching costs. Notion customers using Notion for company wiki face moderate switching costs. For BD operators evaluating SaaS partnerships, switching cost analysis reveals which incumbents have durable positions vs which are commodity-vulnerable.

AI-era moat dynamics

AI dynamics may produce new moat types and invalidate existing ones: (1) **Data accumulation moats**: companies with proprietary training data may have durable advantages. Example: TikTok recommendation system, Bloomberg financial data. (2) **Model quality moats**: leading model providers (OpenAI, Anthropic) may have temporary moats from model quality. Whether moats persist depends on whether model quality plateaus or continues to compound. (3) **Compute scale moats**: training frontier models requires substantial compute investment. Companies with compute access have advantages. (4) **Distribution moats amplified**: incumbents with distribution (Microsoft, Google) can deploy AI features broadly faster than startups. Distribution moats may matter more in AI era. (5) **Switching cost moats threatened**: AI may reduce some switching costs by making data migration and integration easier. SaaS companies relying purely on switching costs may face displacement. (6) **Brand moats threatened in some categories**: AI tools may commoditize categories where brand matters. Companies relying on brand without underlying differentiation may struggle. The AI-era moat dynamics are uncertain. Frameworks for evaluating remain provisional. Honest analysis acknowledges substantial uncertainty in which moats will persist.

Evaluating moat quality

For evaluating a company's moat quality, analytical questions: (1) **Which of the 7 powers does this company have?** Most companies claim moats they don't actually have. Specific identification helps separate real moats from marketing. (2) **How strong is the moat?** Network effects vary in strength; switching costs vary in size; scale economies vary in cost curve steepness. (3) **Is the moat widening or narrowing?** Some moats compound (network effects accumulate); others erode (technology changes reset switching costs). (4) **What could invalidate the moat?** Specific technology changes, regulatory changes, or competitor strategies that could erode the moat. AI is current canonical threat to many moats. (5) **How long can the moat sustain pricing power?** Moats have effective time horizons. Patent moats expire; brand moats erode without reinvestment; network effects can fragment. For BD operators evaluating partnerships, moat analysis reveals which partners have durable positions worth long-term commitments and which face structural displacement risk.

How to Implement

1. 01Identify which of the 7 powers your company has (specific, not claimed).
2. 02Evaluate moat strength and durability honestly.
3. 03Invest to widen the moat (scale, network growth, switching cost accumulation).
4. 04Combine multiple moats for reinforcement.
5. 05Monitor for invalidation risks (technology, regulation, competitor strategies).
6. 06Compete in categories where your moats matter; avoid categories where they don't.
7. 07Use moat analysis in partnership evaluation and investment decisions.

Common Pitfalls

• 01Claiming moats that don't structurally exist.
• 02Relying on single moat type without reinforcement.
• 03Inadequate reinvestment causing moat decay.
• 04Ignoring AI-era threats to traditional moats.
• 05Competing in categories where your specific moats don't matter.

Sources

• Hamilton Helmer '7 Powers'
• Andrew Chen 'Cold Start Problem'