Most PM artifacts mention "defensibility" once and move on. A real moat conversation forces you to name which moat, with which evidence, and which still-undefended flank a competitor would attack first.
This file is the working-PM reference for that conversation. Useful when writing a strategy memo, defending a roadmap to leadership, or stress-testing a "we'll be the leader because we're better" claim.
A moat is a structural reason a competitor cannot copy your product into commodity. Quality, polish, and clever features are not moats — they're table stakes that any well-funded incumbent will catch up on. Moats are the things that get harder for a competitor to overcome the longer you run.
Two common misuses:
- Aspirational moats. "We'll have network effects once we hit critical mass." Every two-sided product says this. It's not a moat until users are visibly choosing you because the network is bigger; until then it's a hypothesis. Rate it emerging, not strong.
- Brand confused with reputation. A reputation for shipping good product is not a moat. Brand becomes a moat when users default to you without comparing alternatives — and that takes years and visible category leadership.
A useful test: if a well-resourced competitor launches a 1.5× better version next quarter, how much of our user base do we lose in 12 months? If the honest answer is "most," the moat isn't real.
| Moat | What it looks like | Common failure mode |
|---|---|---|
| Habit / behavioral lock-in | Users have made the product part of their daily flow; switching means rebuilding muscle memory | Confusing high engagement (vitamin) with habit (painkiller in a daily ritual) |
| Network effects | The product gets better for each user as more users / counterparties join | Overstating one-sided usage as a network effect; ignoring local network density |
| Switching costs | Data lock-in, workflow integration, training investment, social ties | Mistaking inconvenience for switching cost; switching cost only counts if it changes user behavior |
| Proprietary data advantage | Datasets that compound with use and improve the product (recommendations, models, fraud detection) | Assuming generic public data is proprietary; assuming model quality is data quality |
| Ecosystem / platform | Multiple products reinforce each other; integrations make the bundle costlier to leave than any single piece | Calling a product suite an ecosystem before integrations are real |
| Scale economics | Unit costs fall structurally as volume grows; competitors can't match price without losing money | Confusing fixed-cost amortization with structural scale advantage |
| Brand and trust | Users default to you in a category where trust matters (financial, health, security) | Treating awareness as trust; trust requires demonstrated reliability over time |
| Regulatory / IP | Licenses, certifications, patents, or compliance moats that take competitors years to acquire | Patents that don't actually block the workaround; certifications that competitors can also obtain |
Most products have one real moat, one or two emerging, and several that are aspirational at best. Naming the categories honestly is more useful than claiming all eight.
Habit is the most-claimed and least-tested moat. The bar is high:
- Daily or near-daily use, anchored to a recurring user trigger (morning email, end-of-shift report, every commit). Weekly or monthly use is engagement, not habit.
- Returns reinforced by user investment. The user has built up something inside the product (data, history, tags, contacts, settings) that they'd lose by switching.
- Low cognitive cost at the moment of use. Habits run on autopilot. If users still consciously decide whether to open the product, it's not a habit yet.
The Fogg behavior model framing: habit forms when motivation × ability × prompt all line up consistently. Most products fail one of the three — usually ability (too much friction at the moment of use) or prompt (no reliable trigger).
The most-claimed moat after habit, and even more often overclaimed. The empirical bar is tight.
A true network effect means the product becomes more valuable to existing users as new users join. Not "we have lots of users." Not "users invite other users." Those are growth dynamics, not network effects.
The honest test:
- Adding one more user creates dramatic new value for existing users. A new buyer on a marketplace gives existing sellers more demand. A new contact on Slack gives existing teammates more reach. A new node in a payment network unlocks settlement to a new geography.
- The value scales with the relevant network, not the total network. A two-sided marketplace needs local density, not global. A workplace tool needs density inside a single company, not across all companies. Counting global users when only local density matters is the most common overclaim.
- The value is not replicable by a competitor with a smaller network. If a smaller competitor can deliver the same value to a user, the network isn't the moat; the features are.
The failure mode is calling any product where users invite other users a "network effects business." Most of them are not — they're products with referral mechanics. Referrals are a growth lever, not a moat. They don't compound. When the referral chain stops, the value to existing users doesn't change.
A useful filter: if a competitor with 10 percent of our user base launched tomorrow, would their users get 10 percent of the value our users get, or much less? If 10 percent, no real network effect. If much less, the network is doing work.
Pre-PMF, moat questions are mostly distracting. The first job is to find a desperate audience that buys (decision-making/value-hypothesis.md). Worrying about defensibility before you've validated the value hypothesis is a way of avoiding the harder question.
That said, two pre-PMF cases where moats matter:
- The "right and consensus" check. If anyone could build what you're building, you don't have a moat and you don't have a non-consensus insight. That's the lowest-return position. (See
decision-making/value-hypothesis.md§ "The 'what'.") - B2B distribution. In enterprise, the relationship and the integrations are the moat for many years. PMs from consumer backgrounds underweight this. The product can be technically inferior and still win because the buyer already has procurement, security, and integrations sorted.
- Strategy memo to leadership — claims about defensibility need evidence, not adjectives.
- Pricing decisions — products with real moats can hold price; products without one cannot.
- Build-vs-buy — buying often makes more sense for table-stakes capabilities and less sense for moat-relevant ones.
- Acquisition integration — the moat usually lives in one of the acquirer's or acquiree's stacks. Pick the one with the moat; rebuild the rest if needed.
- Competitive response — when a credible competitor enters, the defensible flanks tell you where to invest and where to concede.
The most underrated category of moat among PMs from product backgrounds. The pattern: a "good enough" product wins because it's already inside the chooser's existing stack — integrated, paid for, IT-approved, and procurement-cleared. Microsoft Teams beat Slack in many enterprises this way. Not by being better. By being bundled into Microsoft 365.
Distribution is a moat when:
- The chooser is consolidating vendors. Fewer contracts, fewer integration risks, fewer security reviews. A pre-existing relationship outweighs marginal product quality.
- The integration cost is real. SSO, identity, audit logs, data residency, network controls — already done with the bundled vendor, not yet done with the challenger.
- The chooser's incentive is operational efficiency, not user delight. CIOs are measured on uptime, cost, and risk. None of those reward a product that the users prefer but adds vendor count.
For a challenger, the implication is structural. Beating an incumbent on product alone, when the incumbent has bundled distribution, requires significantly better — usually 5x to 10x better on a measurable outcome, not just nicer to use. The challenger's strategy has to answer:
What stops the chooser from defaulting to the bundled alternative?
Three honest answers, in rough order of how often they hold:
- A new category the incumbent's bundle doesn't cover. The bundle is irrelevant if there's no equivalent inside it. Many AI-native products are in this position relative to Microsoft today.
- Regulatory or compliance pressure that forces a category-specific tool. Healthcare, finance, and security buyers sometimes can't accept a generalist bundled tool.
- User revolt strong enough to force the chooser's hand. Rare. Requires both excellence and a chooser who listens to users — and even then, often loses in renewal cycles.
If the strategy doesn't have a credible answer to that question, distribution is the moat and the challenger doesn't have one yet. See decision-making/user-vs-chooser.md for the broader buyer-vs-user framing.
Best-of-breed products eventually reach a strategic fork that most PMs underweight. Three viable end states:
- Stay best-of-breed. Keep depth, accept the ceiling. The product remains the leader in a narrow category, sells alongside the customer's other tools, never becomes the system of record. Viable but exposes the product to long-run consolidation: bundled competitors eventually approximate the depth, and the chooser switches for vendor reduction.
- Become a platform. Expand outward into adjacent categories so the product itself becomes a multi-category bundle. Hard, slow, and dilutes the original depth. Slack tried this with the workspace expansion (video, dashboards, tabs); Notion tried it with the all-in-one workspace pitch. The risk is becoming bloated without becoming bundled.
- Get acquired into a platform. Sell into a bundle that already exists. Trade independence for distribution. Often the rational outcome when distribution beats product quality at the chooser's level.
The fork is not free to delay. Sitting in best-of-breed without choosing means the bundled competitor approximates the depth on a five-to-seven-year timeline. By the time the company chooses, the options have narrowed.
The PM-level implication: a strategy memo for a best-of-breed product needs to take a position on which path it's on, not just "we're the leader in our category." Each path implies different roadmap, different acquisitions, different hiring, different metrics. Hedging across all three usually means executing none of them.
Engagement and habit-formation tactics can be designed in ways that exploit users (variable rewards, dark patterns, friction asymmetries between getting in and getting out). Building a habit moat by making the product worse for users — harder to leave, harder to ignore, harder to pause — is short-term defensibility at the cost of long-term trust.
The clean version of a habit moat: users keep using the product because it's the best way to get a recurring job done. The exploitative version: users keep using it because the team made leaving artificially painful.
Trust is a moat too. Burning it to build engagement is a bad trade.
Before claiming defensibility in a memo or deck:
- Which moat, specifically? Pick from the eight. Generic "we have a moat" doesn't survive a review.
- Where's the evidence? Behavioral data, retention curves, switching events, integration depth — not adjectives.
- Aspirational vs. emerging vs. real? Be honest. Most teams over-rate.
- What's the still-undefended flank? Every product has one. Naming it is more useful than pretending it doesn't exist.
- Would this survive a 1.5× better competitor? If not, the moat isn't real yet.
- Are we building the moat ethically? Habits built on dark patterns are short-term defensibility for long-term trust loss.
If two of these can't be answered, the strategy doesn't yet rest on a moat. It rests on hope, which is fine to admit and dangerous to disguise.