Moat / defensibility lens

Status: Documented, not shipped · Evidence: P · Family: Strategy and opportunity · Verdict: reject (2026-06-09)

What it is

The moat or defensibility lens asks one question about a position, a product, or a plan: what stops a competent competitor from copying you and competing the advantage away? Instead of asking “is this good?” it asks “is this holdable?” - and answers by walking a fixed taxonomy of structural barriers and checking, for each, whether you genuinely hold it and how long it would take a rival to neutralise it. The metaphor is Warren Buffett’s: a great business is a castle with a wide, deep moat that keeps attackers out for decades.

The durable cognitive operation underneath the metaphor is a directed adversarial durability test against a known barrier menu. Three things make it specific:

1. The object is the durability of an advantage, not the advantage itself. The lens presumes you already have (or plan) an edge and interrogates whether it survives contact with imitators, fast-followers, and incumbents.
2. The test is run through a canned taxonomy of barrier types. Practitioners do not generate the barriers from scratch; they look them up. The two dominant menus are Morningstar’s four sources of structural advantage (intangible assets such as brands and patents, customer switching costs, network effects, cost advantages), codified by Pat Dorsey, and Hamilton Helmer’s Seven Powers (scale economies, network economies, counter-positioning, switching costs, branding, cornered resource, process power), where each Power is explicitly a Benefit paired with a Barrier to imitation.
3. The emitted artifact is a defensibility audit: a table of candidate barriers, a held / partial / absent verdict on each, an estimated time-to-erode, and the one or two on which the position actually rests. In venture practice this is the “why can’t a competitor copy this?” page of a diligence memo.

So the honest description separates the famous metaphor from the working move. The metaphor (“build a moat”) is a goal. The move (“audit which of these known barrier types you hold, and how durable each is, against an imagined imitator”) is a domain-specific adversarial checklist - and that, not the castle imagery, is what has to clear the selection bar.

When it helps / when it misleads

The lens helps when a plan’s appeal rests on an advantage that competitors can see, and the real question is whether the advantage is fenced. It is a good corrective to the optimism of a strong-on-day-one product, it forces a named reason a position will still be valuable in three to five years, and the barrier taxonomy is a genuine retrieval aid - it reminds a thinker to check network effects or switching costs they would otherwise skip. It is sharpest in competitive-strategy, investing, and startup-diligence contexts, which is exactly the domain it was built for.

It misleads or wastes effort when:

• The taxonomy is mistaken for the analysis. Ticking “we have a brand” or “there are network effects” is a label, not a finding. The same NFX and a16z practitioners who evangelise network effects warn that named network-effect businesses routinely split markets and lose to copycats; the barrier word is not the barrier. A checklist run as a lookup produces a confident list of moats that do not hold.
• It is pointed at a problem that is not about defensibility. For an internal process choice, a reversible call, or a decision with no competitor, “what stops someone copying us?” is the wrong question and burns effort manufacturing a competitive frame that is not there.
• Durability is assumed rather than the very thing in doubt. The empirical record is that durable advantage is rare and getting rarer (Wiggins and Ruefli). A lens whose first move is “which moat do we have?” can smuggle in the premise that a moat exists; the honest version starts from “does any barrier survive a determined imitator?” and is willing to answer no.
• It substitutes for valuing the bet. Defensibility is one input. A perfectly moated business in a tiny or shrinking market is still a poor decision; the lens does not price the opportunity, only its fences.

What the evidence says

The honest grade for the candidate’s stated move - “test what stops a competitor copying you” - is P (practitioner), and the grading has to be unusually careful here because moat analysis is a case where a large, genuinely strong economics literature sits next to a method that literature never actually tested. Borrowing the former to grade the latter is precisely the transferred-evidence laundering this library exists to prevent.

What the record supports - about the phenomenon. That durable competitive advantage exists and pays off is well evidenced. The resource-based view (Barney 1991) holds that resources which are valuable, rare, costly to imitate, and organisationally exploited (the VRIN / VRIO tests, where “costly to imitate” is the moat question restated) yield sustained advantage; a meta-analysis of 125 studies spanning more than 29,000 organisations (Crook, Ketchen, Combs and Todd 2008) found the strategic-resources-to-performance relationship at r = 0.22 overall, rising to r = 0.29 for resources that meet the RBT criteria. Wiggins and Ruefli (2002) document empirically, across many industries and decades, that persistent superior performance occurs but is rare and that “sustained” advantage is often a chain of temporary advantages; their 2005 follow-up (“Schumpeter’s Ghost”) finds it has become harder to sustain over time. McGahan and Porter (1999) measure how profitability shocks persist. This is a respectable M-to-S body - on whether moats are real and matter.

What the record does NOT support - about the method, and the laundering trap. None of that work tests the moat / defensibility lens as a thinking exercise: whether a decider who runs a barrier audit makes better calls than one who does not. The studies measure the outcome (firms with durable advantages out-earn firms without) and the correlates of that outcome (which resource properties predict persistence), not the act of analysing defensibility. Crook et al.’s r = 0.29 says VRIO-qualifying resources pay; it says nothing about the value of running a VRIO or moat checklist over a plan. Attaching that effect size to “moat analysis works as a method” would be laundering the economics literature’s robustness onto an analytical ritual it never evaluated - the same error the catalog flagged when it capped TRIZ and reference-class-style methods at their move’s own evidence rather than their parent theory’s. The practitioner taxonomies themselves - Dorsey’s four sources, Helmer’s Seven Powers - are codifications of investing and strategy experience, not tested instruments; they are P-grade authority, no controlled comparison.

Transfer caveat (required). All of the adjacent evidence is from human firms, investors, and strategists in financial and organisational settings; none of it studies a moat or defensibility audit performed by or with an AI agent. The evidence is transferred from human contexts and is not validated for AI-augmented use. The conservative governing grade is therefore P: a recognised practitioner lens with strong evidence for its premise (moats are real and valuable) and no controlled evidence for its move (running the lens improves decisions), with the M-to-S resource-based-view results explicitly not counted toward the grade because they measure the phenomenon, not the method.

Excluded figures. Morningstar’s Wide Moat Focus Index reports roughly 14.6% average annual return since its 2002 inception versus about 10.1% for its benchmark; this is a vendor index figure, selection-laden, and - by Morningstar’s own statement that the moat rating is “designed to replicate analyst-assigned moat ratings rather than predict excess returns” and is “not meant to predict excess returns” - it does not measure the moat lens as a decision method and may not move the grade. The widely repeated diligence rule of thumb that “a true data advantage takes competitors over three years to replicate” traces to no primary source measuring replication time; it is recorded as practitioner lore and excluded as fact.

Why it is / is not a skill here

Verdict: Reject; status pm (the irreducible remainder belongs to the sibling pm-skills product). This overturns the catalog’s prior cand / build / P tag, and the reason is concrete rather than a re-weighting. The prior tag was the un-researched placeholder boilerplate (“Clears the bar but lower priority (P-tier coverage)”) shared by roughly twenty unexamined P candidates; on actual research the move does not clear the overlap ceiling as a thinking skill, and what is left over after the overlap is removed is product- and business-strategy domain content, which this library routes to pm-skills.

The Build burden is to name one distinct, durable, general cognitive move that no shipped skill (or short chain) already produces, and to show its working mechanism shares less than about a fifth with every shipped skill. The moat lens fails that burden because it decomposes cleanly into one general operation that is already shipped and one remainder that is out of scope:

• The general operation is red-team-light. Strip the business vocabulary and the move is: take an asset you believe in, then construct the strongest case that it does not survive a determined adversary - here the adversary is an imitator and the asset is your advantage’s durability. That is red-team-light (“construct the strongest adversarial case against a thesis”) pointed at a specific thesis (“this advantage is durable”). The generative mechanism - adopt the attacker’s seat, marshal the strongest erosion case, report what holds - is shared well above the ~20% ceiling. Red-team-light is the home for the operation, exactly as it is for steelmanning.
• The barrier taxonomy is a prompt, not a separable move - and it is the cognitive-bias-checklist pattern. What makes the lens specifically moat analysis is the canned menu of barrier types (scale, network effects, switching costs, brand, intangibles, counter-positioning, process). Running a subject against a fixed taxonomy and rating each entry is structurally the move the catalog already rejected in cognitive-bias-checklist (“the taxonomy is a prompt, not a move”). A barrier menu can ride inside red-team-light as an optional reference; it is not a new mechanism.
• The irreducible remainder is competitive-strategy domain content, which the library defers to pm-skills. The candidate sits in the strategy-and-opportunity family, whose registry header reads “weakest cross-domain fit; many defer to pm-skills.” Every neighbour that is a competitive-analysis frame is already routed away: porters-five-forces (flag / reject, pm-domain, “competitive-structure scan”), value-proposition-contrast (pm, “sharpen vs the next-best alternative”), opportunity-solution-tree (pm), blue-ocean-tools and jobs-to-be-done (flag / reject), swot (excluded). The moat lens is the same kind of object: a business-strategy competitive-economics method whose value is the domain taxonomy. Its closest in-family build candidate, white-space-adjacent-possible, is the inverse question (where is there reachable opportunity?) and does not subsume it; but neither does it rescue the moat lens as a thinking skill, because the thing that is uniquely “moat” about it is the competitive remainder, and that remainder is pm-skills’ charter.

Why pm / reject rather than fold or a clean build. A fold into red-team-light would over-claim: red-team-light carries the adversarial operation but not the barrier taxonomy, so something real is left unaccounted, and that something is product-strategy domain knowledge, not a thinking move - which is exactly what pm is for (see the decision-brief-pr-faq precedent, rejected to pm because its general move was an existing skill and its remainder was product-domain). A build would ship a domain-specific competitive-analysis checklist into a general thinking library that has deliberately rejected every sibling of that kind. So the honest service is to document the lens, point the operation to red-team-light (run the strongest “a competitor catches up” case), point the premortem-flavoured variant (“imagine the advantage was copied, work back to how”) to premortem, and route the defensibility content itself - the moat taxonomy and the diligence question - to pm-skills, where competitive strategy lives. The learning value of the NO: a famous and genuinely useful strategy frame, backed by strong evidence for its premise, is still not a general thinking skill when its distinctive content is domain knowledge and its general operation is a skill the library already ships.

Lineage and who to read

The “economic moat” metaphor is Warren Buffett’s, used for decades in Berkshire Hathaway letters and talks to describe a durable structural advantage that protects a business “castle.” It was codified into an analytical framework at Morningstar, chiefly by Pat Dorsey, who named four sources of structural advantage (intangible assets, switching costs, network effects, cost advantages) and built Morningstar’s wide / narrow / no-moat rating; read his The Little Book That Builds Wealth (2008). The most systematic modern treatment is Hamilton Helmer’s 7 Powers: The Foundations of Business Strategy (2016), which defines Power as a Benefit paired with a Barrier to imitation and enumerates seven of them; counter-positioning is his distinctive contribution. The academic backbone is the resource-based view - Jay Barney, “Firm Resources and Sustained Competitive Advantage” (1991), and the VRIO test - with the empirical record carried by Crook, Ketchen, Combs and Todd’s 2008 meta-analysis and Wiggins and Ruefli’s 2002 and 2005 studies of how rare and how shrinking sustained advantage actually is, alongside McGahan and Porter (1999) on profit persistence. For the startup-diligence dialect of the same lens (“why can’t a competitor copy this?”), read the network-effects writing of NfX (the Network Effects Manual / Bible) and Andreessen Horowitz (Andrew Chen), noting their own warnings that named moats often fail to hold. None of these is a trademark the library must annotate: “economic moat” and “defensibility” are generic descriptive terms, and while 7 Powers and VRIO are named frameworks attributable to their authors, the move itself is documented descriptively and is not flagged as branded.

Named sources

• Jay B. Barney, “Firm Resources and Sustained Competitive Advantage,” Journal of Management 17(1) (1991): 99-120. The resource-based view; the VRIN / VRIO tests, where “costly to imitate” is the defensibility question stated formally. Foundational theory - it specifies the conditions for durable advantage, not the value of auditing for it. (M, for the phenomenon)
• T. Russell Crook, David J. Ketchen Jr., James G. Combs and Samuel Y. Todd, “Strategic Resources and Performance: A Meta-Analysis,” Strategic Management Journal 29(11) (2008): 1141-1154. Meta-analysis of 125 studies / 29,000+ organisations: strategic-resources-to-performance r = 0.22, rising to r = 0.29 for RBT-qualifying resources. Strong evidence the moat premise holds; measures resources-to-outcome, NOT the moat lens as a method - cited to mark the boundary the grade may not cross. (M/S, for the phenomenon, explicitly not counted toward the method grade)
• Robert R. Wiggins and Timothy W. Ruefli, “Sustained Competitive Advantage: Temporal Dynamics and the Incidence and Persistence of Superior Economic Performance,” Organization Science 13(1) (2002): 81-105. Large-sample longitudinal evidence that persistent superior performance is rare and that sustained advantage is often a chain of temporary advantages. Cited for the “durability is rare, do not assume it” caveat. (M, for the phenomenon)
• Robert R. Wiggins and Timothy W. Ruefli, “Schumpeter’s Ghost: Is Hypercompetition Making the Best of Times Shorter?,” Strategic Management Journal 26(10) (2005): 887-911. Finds competitive advantage has become harder to sustain over time. Supports the misleads-when-durability-assumed warning. (M, for the phenomenon)
• Hamilton Helmer, 7 Powers: The Foundations of Business Strategy (Deep Strategy LLC, 2016). The most systematic barrier taxonomy: seven Powers, each a Benefit plus a Barrier to imitation; source of counter-positioning. Practitioner / foundational for the method’s content; no controlled test. (P)
• Pat Dorsey, The Little Book That Builds Wealth: The Knockout Formula for Finding Great Investments (Wiley, 2008), and the Morningstar Economic Moat Rating methodology. Codifies the four sources of structural advantage and the wide / narrow / no-moat scheme. Practitioner authority; Morningstar states the rating is “not meant to predict excess returns.” (P)
• Anita M. McGahan and Michael E. Porter, “The Persistence of Shocks to Profitability,” Review of Economics and Statistics 81(1) (1999): 143-153. Decomposes and measures how profitability shocks persist across industries and firms. Background evidence on persistence; not a test of the lens. (M, for the phenomenon)

Excluded under the evidence rule: Morningstar’s Wide Moat Focus Index return figure (~14.6% vs ~10.1% benchmark since 2002) is a selection-laden vendor index that Morningstar itself says is not designed to predict excess returns, so it does not measure the lens as a method and is not counted toward the grade; and the “a real data moat takes 3+ years to replicate” diligence rule of thumb has no traceable primary source and is excluded as fact.