The Manufacturing Moat Framework
Understand the Business. Direct the Technology.
For private equity operators and manufacturing executives
In this series: Building Moats: The Manufacturing Technology Playbook and The Four Pathways Value Creation Framework.
TL;DR
Most mid-market manufacturers are defended by exactly two things: the capital cost of their equipment and the freight cost of shipping around them. That is a real defense, but it is a thin one, and it is the profile that gets attacked by imports, customer insourcing, and any competitor willing to buy the same machines.
Who this is for: private equity deal teams and operating partners underwriting or improving industrial businesses, and the executives who run them.
What it answers: three questions. What actually defends this business, and where is the evidence in the numbers? Which defenses can realistically be built inside a hold period, and what does the market make them worth? And how does defensibility convert into EBITDA, multiple, and return? The fourth question, how to sequence and govern the digital and AI build itself, is the subject of the companion piece, Building Moats: The Manufacturing Technology Playbook.
How it works: the framework scores a manufacturing business across nine structural moats, validates each claimed moat against financial evidence, and identifies which moats can realistically be deepened inside a private equity hold period. The punchline: the moats manufacturers typically lack (data, workflow embedment, ecosystem, network, and brand beyond a regional reputation) are precisely the ones that can be built without heavy capex, at a fraction of the cost of the physical moats the business already paid for, and cheaper to build than they have ever been, because generative AI coding agents have collapsed the cost of software construction.
Score honestly, demand evidence, and prioritize by moat-per-dollar. That is the entire method. The rest of this document is the detail.
Origins and Attribution
This framework did not originate with me, and it did not originate recently.
Warren Buffett is the progenitor, and the anchor of everything that follows. The castle-and-moat metaphor runs through his shareholder letters and annual meetings from the 1980s onward: wonderful businesses are economic castles, capitalism guarantees that rivals will assault any castle earning high returns, and management's job is to widen the moat every year. By his 2007 letter he had formalized the idea that a truly great business requires an enduring moat protecting excellent returns on invested capital, precisely because high returns attract the capital that competes them away. Buffett's version came with two operational tests that this framework preserves: the capital test (could a well-funded competitor take meaningful share at any price?) and the pricing power test (can you raise prices ahead of inflation without losing volume?).
Academia validated the concept from several independent directions over four decades. Michael Porter's entry barriers (1980), Jay Barney's resource-based view (1991), and Hamilton Helmer's 7 Powers (2016) each arrive at the same conclusion by different routes: sustained returns require structural barriers that competitors cannot arbitrage away. Helmer contributes a discipline this framework preserves throughout: a real moat must provide both a benefit and a barrier. Possession is not defensibility.
Modern investors have adapted the concept for the software era, and this framework borrows its taxonomy from one of them: venture investor Gokul Rajaram, who in 2026 laid out eight moats for enduring software companies (data, workflow, regulatory, distribution, ecosystem, network, physical infrastructure, and scale), with a rule of thumb that four or more indicates durable defensibility, including against AI-driven disruption.
Private equity applies it in practice. I have personally seen this approach at work inside private equity diligence and value creation across mid-market manufacturing: deal teams scoring targets on structural defensibility, and operators insisting that every claimed moat show up somewhere in the numbers. That insistence on evidence shaped both the manufacturing translation of each moat and the validation metrics in the appendix.
One deliberate choice deserves a note. Some modern software adaptations, Rajaram's included, drop Brand from the list as too hard to measure. This framework keeps it, because the progenitor demands it: Buffett's formative moats were brand moats. See's Candies and Coca-Cola taught him that share of mind confers pricing power for decades. And the validation layer answers the measurability objection directly: brand either shows up as a price premium and a win rate when you are not the low bid, or it scores zero. Measured that way, brand stops being a narrative and becomes a number.
The software adaptations were built for SaaS. The irony is that half of those moats were borrowed from the physical economy in the first place. Translating them back to manufacturing is not a stretch. It is a homecoming, with one twist: the strong and weak categories flip. A software company struggles to build physical infrastructure. A manufacturer usually has nothing but physical infrastructure, and struggles with everything digital. That asymmetry is where the strategy lives.
The Two Validation Tests
Before scoring anything, internalize Buffett's two tests. They convert the framework from a self-assessment exercise into a diligence tool.
The capital test. Give a competent competitor serious capital, scaled to the market. Could they take meaningful share from this business? Where exactly would they fail? If the honest answer is "they'd replicate us in eighteen months with the same equipment vendors," there is no moat, regardless of what the checklist says.
The pricing power test. Can the business raise prices ahead of input inflation without losing volume? A manufacturer that pushed through material surcharges in an inflation cycle and kept the business demonstrated a moat. One that claims workflow embedment and regulatory qualification but bled volume the moment it raised prices has a paper moat. Pricing power is the moat showing up in the P&L.
A moat that appears on the scorecard but never appears in the numbers is a narrative, not an asset. Score it down.
Products and Markets: Where Moats Live
Before scoring a single moat, read the terrain, because two things outside the scorecard determine what any score is worth.
The market is a multiplier on the score. Moats are company-level; their value is set by the market around them. Newspapers had magnificent moats, and the moats held right up until the market underneath them evaporated. A high score in a declining, import-exposed category is worth less than a modest score in a growing, consolidating one. Five reads matter for manufacturers: growth or decline of the end market, cyclicality (housing-linked, defense budgets, ag cycles), fragmentation (a fragmented market is roll-up runway), import exposure (freight economics and labor content decide whether the product can be attacked from offshore), and substitute pressure (material substitution, steel to aluminum to composite, is the manufacturing version of disruption).
The product shapes the feasible moats, but it does not finish the argument. Where a product sits on the spectrum from commodity to engineered sets the default menu. A commodity converter's natural moats are scale, physical infrastructure, and distribution. Engineered, configured, and specified products unlock the rest by default: data, brand, workflow, regulatory. But commodity does not mean locked out; it means the knowledge moat moves from the product to the application. I lived this pattern as VP of Marketing at American Gypsum: we pursued proprietary fire-rated assembly designs, UL-tested and approved, that architects and specifiers incorporated into their building designs. The wallboard was a commodity; the tested assembly knowledge around it was not. The strategic consequence holds across the spectrum: adding engineering content, configurability, application knowledge, or attached services expands the set of moats the business is allowed to build. Product strategy is moat strategy.
Niche depth beats broad shallowness. The reliable mid-market path to a high moat score is dominating a defensible niche with deep application expertise, the pattern Hermann Simon documented in his hidden champions research. A niche adds its own protection: a market too small to justify a large competitor's capital passes the capital test by default.
Read the market, place the product, pick the niche. Then score the nine moats against that terrain.
The Nine Moats, Translated for Manufacturing
1. Data
In SaaS: proprietary data accumulated from a critical mass of customers or secured through exclusive access agreements.
In manufacturing: engineering and process knowledge. Product configurators, BOM (bill of materials) logic, and rules engines. Process recipes and machine parameters. Tolerances learned through twenty years of scrap. Quality and warranty history. Quoting and estimating history. The long half-life data (product configuration rules, process recipes) is the most protective; commodity pricing intelligence decays in weeks.
The hard truth in the mid-market: this moat usually exists but is not captured. It lives in a retiring plant manager's head, in paper travelers, and in spreadsheets. A moat you possess but have not codified walks out the door at retirement. Trade secrets and design know-how live here rather than under Regulatory: unlike patents they never expire, but they are only a moat while captured and protected.
Digital and AI lever: capture tribal knowledge into configurators, rules engines, and agents that codify expert judgment; foundation models are commodities available to every competitor, and proprietary captured knowledge is the only durable AI advantage.
Full Data build program is in the companion playbook, Building Moats.
2. Workflow
In SaaS: mission-critical software embedded in daily operations.
In manufacturing: embedment in the customer's operations. VMI (vendor-managed inventory) and consignment programs. EDI wired into the customer's MRP (material requirements planning system). Kanban replenishment. Your part numbers drawn into their engineering documents. Their tooling in your building. Engineering change collaboration.
The test is simple: does switching you out require the customer to run a project, or just cut a PO to the next name on the approved vendor list? Approved-vendor status is table stakes that gets you into the game. The workflow moat is what happens when leaving you becomes a project instead of a purchasing decision.
One caveat every deal team will apply: workflow depth and customer concentration travel together. Deep embedment with a customer representing forty percent of revenue is not a moat; it is a dependency wearing a moat's clothes. Score workflow strength against the breadth of customers it spans, and treat concentration as the discount it is.
Digital and AI lever: EDI, API, and VMI integration deepen embedment, and AI agents at order intake extend it; embedment is also the defense against agentic procurement, which will rebid transactional suppliers annually at zero effort.
Full Workflow build program is in the companion playbook, Building Moats.
3. Regulatory
In SaaS: licenses and compliance barriers.
In manufacturing: arguably stronger here than in software. Certifications and qualifications gate entire markets: AS9100 and NADCAP in aerospace (the industry's quality and special-process accreditations), ITAR and CMMC in defense (export control and cybersecurity requirements), PPAP in automotive (the production part approval package), UL listings, FDA clearance in medical. Requalifying a new source on an aerospace or defense program can take one to two years, which is why qualified legacy suppliers hold sole-source positions on aging platforms for decades.
The sleeper asset: environmental permits. A permitted plating line, foundry, or chemical process is nearly irreplaceable, because those permits are effectively no longer granted new.
Patents belong here as well. A patent is state-granted exclusivity, the purest regulatory moat there is, with two caveats: it expires on a known date, and it is only as strong as the willingness to enforce it.
Digital and AI lever: a digital QMS (quality management system) turns compliance speed into a weapon, and AI collapses the cost of FAI (first article inspection) and PPAP preparation and audit readiness; the fastest qualified supplier wins programs.
Full Regulatory build program is in the companion playbook, Building Moats.
4. Distribution
In SaaS: channels and platform access.
In manufacturing: dealer networks, two-step distribution, retail shelf space, rep agencies, approved-vendor positions at major accounts, private-label relationships. And for freight-sensitive products (anything bulky and low value-per-pound), geography itself is distribution: a plant network positioned near customers is a channel competitors cannot ship their way around.
Digital and AI lever: dealer portals, e-commerce, and spec, CAD, and BIM (building information modeling) tools get the product designed in before the bid, and machine-readable product data is the new shelf space as buyers begin asking AI assistants what to buy.
Full Distribution build program is in the companion playbook, Building Moats.
5. Ecosystem
In SaaS: third parties building businesses on your platform.
In manufacturing: thinner, but real. Installers and contractors trained on your product. Architects and specification engineers who default to your detail drawings. Accessory and aftermarket players building around your platform. And the installed base itself, which compounds: more units in the field means more technicians who know the product, more parts demand, and more reasons the next job specifies the same brand. The fire-rated assembly story from earlier is this moat forming in real time: tested knowledge drawn into architects' construction documents begins life as a Data moat and matures into ecosystem gravity, one specified assembly at a time.
Digital and AI lever: installer apps, partner APIs, and an AI troubleshooting copilot trained on decades of field data make your product the easiest to install and service; competitors can copy the app, not the corpus.
Full Ecosystem build program is in the companion playbook, Building Moats.
6. Network
In SaaS: two-sided network effects, the crown jewel.
In manufacturing: rare in its pure form, and that is worth being honest about. True network effects mostly do not exist in manufacturing. The interesting move is that this is the one moat digital can create where the physical business has none: a dealer portal or configurator that aggregates demand signal across hundreds of channel partners produces forecasting and product insight that no single competitor or customer can match. Call it a manufactured data network effect.
Digital and AI lever: AI-driven forecasting and pricing convert aggregated channel demand signal into a compounding advantage no single competitor or customer can reconstruct.
Full Network build program is in the companion playbook, Building Moats.
7. Physical Infrastructure
In SaaS: the hardest moat to build.
In manufacturing: the native strength. Plants, specialized equipment, tooling libraries (thousands of dies and molds, and it matters enormously who holds title to them), permitted facilities, capacity that takes years and eight figures to replicate, and a freight-logical footprint. Competing with an entrenched manufacturer means out-capitalizing installed capacity, not underbidding one job.
Digital and AI lever: MES (manufacturing execution systems), OEE (overall equipment effectiveness), predictive maintenance, and vision inspection convert paid-for assets into throughput without capex; the prerequisite is capturing machine data now, because uncaptured process data is training signal lost forever.
Full Physical Infrastructure build program is in the companion playbook, Building Moats.
8. Scale
In SaaS: infrastructure economics and market coverage.
In manufacturing: purchasing leverage on steel, aluminum, and resin. Fixed-cost absorption. Multi-plant redundancy, which is itself a sales weapon as major customers increasingly demand dual-site supply continuity. National footprint for national accounts. And the capacity to fund automation that single-plant competitors cannot, which converts scale into cost curve.
Digital and AI lever: one data platform and one AI investment amortize across every plant; scale defended by data leverage is strengthening while scale defended by headcount leverage erodes.
Full Scale build program is in the companion playbook, Building Moats.
9. Brand
In Buffett's canon: the original moat. Share of mind, not share of market, is what let See's raise prices every year for decades.
In manufacturing: brand operates along a spectrum defined by who makes the buying decision.
At the pure B2B end, brand takes the shape of channel reputation: the installer who defaults to your product because it never generates callbacks, the specifier who draws your details because technical support answers the phone, the purchasing manager for whom you are the safe choice nobody gets fired for selecting. The underlying asset is reliability compounded over decades: warranty performance, on-time delivery history, field failure rates. Slow to build, fast to destroy, which is precisely what makes it a moat.
For consumer goods manufacturers, no translation is needed. Brand is the classic consumer moat Buffett underwrote at See's and Coca-Cola: awareness, loyalty, and share of mind that command a premium at the shelf or in the cart.
Between the two sits a growing middle. Building products and consumer-adjacent categories have always had pull-through, where end-customer recognition creates demand the channel cannot ignore. And a rising number of traditionally channel-bound manufacturers are now selling direct to consumer. The moment they do, they compete on consumer brand terms, which demands consumer brand capabilities (demand generation, e-commerce experience, post-sale service) and disciplined management of the channel conflict that comes with them.
The measurability discipline: brand must be validated as pricing power. If the business commands a premium over comparable competitors, wins a meaningful share of bids where it is not the low price, or gets specified by name, brand is doing real work. If none of that is true, the "great reputation" in the CIM (the confidential information memorandum that markets the business) is a narrative, and it scores zero.
And one of the best ways to test that reputation is also the oldest: talk to customers. Reference calls, win-loss interviews, and two plain questions, why do you buy from them and what would make you leave, will reveal within a week whether the brand promises are real or narrative. Customers do not repeat marketing copy; they describe what they actually experience, and the gap between the two is the diligence finding.
Digital and AI lever: machine-legible specs and citable technical content decide whether AI assistants recommend you, while field data and named case studies supply what AI-generated content cannot fake.
Full Brand build program is in the companion playbook, Building Moats.
Scoring, and the Commodity Converter Problem
Score each moat 0 to 3 using the rubric in the appendix, for a total out of 27. A useful threshold: a business scoring 2 or higher on at least four dimensions (roughly 14 or higher in total) has durable, multi-dimensional structural defense.
Score a typical mid-market manufacturer honestly and the profile is predictable: strong on physical infrastructure, partial credit on regulatory and distribution, brand rarely extending beyond a regional reputation, and weak to zero on data, workflow, ecosystem, and network. That profile has a name: the commodity converter. Defended by capital intensity and freight. Attacked by imports, customer insourcing, and anyone willing to buy the same equipment.
The strategic question for a platform or add-on then becomes concrete: which one or two moats can be deepened inside a five-year hold, and does that move the exit multiple?
Be honest about build horizons, because the nine moats are not equally movable inside a hold. Data, workflow, network, and decision-support advantages can be built in two to three years. Distribution and channel reputation deepen on a five-year horizon with sustained investment. Regulatory qualifications, ecosystems, and consumer brands are decade-scale builds that are usually acquired rather than constructed, which is one reason add-on M&A is often the fastest moat construction available. And note that AI compresses construction, not the clock: coding agents can ship the tool in a quarter, but adoption, trust, and data accumulation set these horizons, and none of them respond to compute. Match the roadmap to the clock.
Two cautions, both inherited from Buffett:
Trajectory beats inventory. Buffett judged managers on whether the moat widened this year, not on how many moats the business held. A four-moat company with all four eroding is worse than a one-moat company deepening it. For every dimension, ask: wider or narrower than twelve months ago?
Count is descriptive, not predictive. A high score describes structural defense. It does not guarantee market performance, and it does not survive contact with a business model shift. Moats are stress-tested by transitions, not by steady state.
How Moats Become Returns
A moat score is not the point. The point is what it does to the numbers every sponsor underwrites: Enterprise Value = EBITDA x Multiple, and Equity Value = Enterprise Value minus Net Debt. Value creation travels four pathways: grow the EBITDA base, earn the multiple premium, kill the discount, and delever. The moat framework plugs into all four.
Moats earn the premium. The multiple a buyer pays is a price on durability, and the nine moats are the structure underneath every premium story a banker tells: revenue durability, market position, sticky relationships, management-independent process. A moat score backed by validation evidence is the premium narrative a buyer's diligence cannot puncture. And the premium is not hypothetical: GF Data, the ACG-owned database of middle-market private equity transactions, has documented for two decades a measurable quality premium in purchase multiples for businesses with above-average financial performance, defined as trailing-twelve-month EBITDA margin and revenue growth both above 10%. The same data carries a warning worth underwriting around: below that line, premium stories rarely get paid, which makes margin and growth work the gate that opens this pathway.
Foundations kill the discount. The Foundation Check in the appendix maps directly to the discount side: cyber and IT risk, data integrity, key-person dependency, operational discipline, concentration. Discounts do not average against premiums; a single severe one (a breach, a concentration cliff, numbers that fail diligence) can zero the deal on its own. That asymmetry is why foundations never add points but always cap them.
Moat construction grows the base. The best moat initiatives are double-entry, provided the base-growth claim is stated as a measurable mechanism rather than a hope: a configurator deepens the Data moat while raising quote capacity, speed-to-quote, and option attach; MES deepens Physical Infrastructure and releases throughput; a dealer portal deepens Distribution and extends selling capacity without adding headcount. Score effects, not functions, and let initiatives that move multiple pathways jump the queue.
Data funds deleveraging. Free cash flow that retires acquisition debt builds equity value even when enterprise value stands still, and in manufacturing the largest cash reservoir is usually working capital. Releasing it runs on data: inventory turns require inventory accuracy, and working-capital discipline requires a trusted system of record. Debt paydown is the quietest of the return sources, and the one most directly funded by getting the data layer right.
This works in both directions. Buy-side, score the target's moats and discount exposure to underwrite the multiple. Sell-side, spend the final eighteen months of the hold building the evidence: the moat narrative a buyer will pay for is constructed, not discovered. That work pays measurably too: GF Data's recent transaction analysis found that sellers who invested in sell-side quality of earnings preparation averaged higher purchase multiples than those who did not.
The complete treatment of the four pathways, the gates that control when each one pays, and the exit execution overlay is in the third article of this series, The Four Pathways Value Creation Framework.
From Score to Roadmap
The score's first use is diligence. Its second is direction, because the moats manufacturers typically lack are the ones built with software, services, content, and process rather than capital equipment. A mid-market manufacturer cannot buy another physical moat for less than eight or nine figures, but it can build a configurator, a dealer portal, a digital QMS, and a captured knowledge base for well under seven. Measured in moat-per-dollar, digital is among the highest-leverage moves available inside a hold period.
Two facts sharpen that math. First, the reason digital moats got cheap has a name: generative AI coding agents have collapsed the cost and time of building custom software, flipping the buy-versus-build calculus that kept the mid-market renting compromised off-the-shelf tools for twenty years. The caveat that keeps the thesis honest: construction is cheap for competitors too, so the tool is never the moat. The moat is what the tool encodes and accumulates: the captured rules, the process data, the customer embedment. Second, cheap construction moved the bottleneck rather than removing it. Coding agents compress build time and do nothing for adoption, and change management is a human problem solved by humans on human timelines. For the investor this cuts both ways: build horizons hold even as construction compresses, and a management team with demonstrated organizational change capability just became an underwritable asset, because when everyone can build the tools, moving people through change is the differentiating capacity.
One discipline governs that spend. Most mid-market digital transformation is modernization wearing the wrong name: replace the ERP, migrate to the cloud, refresh the licenses, declare victory. Modernization makes the business current. Transformation makes it defensible. The test is to name the moat a program deepens, and if the room goes quiet, it is an IT refresh with a transformation budget.
The execution layer, the three types of IT projects, the foundational fixes that gate everything, the ERP trap, and the per-moat build playbook with its AI angles, is the subject of the companion article, Building Moats: The Manufacturing Technology Playbook. This article's job is the lens and the score. The playbook's job is the build.
Putting It to Work
The method compresses to five moves:
- Read the terrain: market trajectory, product position, niche.
- Score the nine moats honestly, 0 to 3, recording trajectory alongside every score.
- Run the Foundation Check, letting any at-risk foundation cap what it touches.
- Validate every claimed moat against the numbers: pricing power, rebid resilience, switching evidence, the capital test.
- Set direction by moat-per-dollar, front-loading initiatives that deepen a moat and move more than one value pathway, with foundational work scheduled first; the companion playbook governs that sequencing.
Done that way, the same exercise serves both audiences. The investor gets a diligence lens that converts "great business" into a scored, evidenced claim. The operator gets a technology agenda in the language a sponsor-controlled board actually rewards, because in a portfolio company the board mostly is the sponsor: EBITDA, cash, and the exit multiple, with defensibility as the mechanism that connects the spend to all three.
Buffett's instruction to his managers was not to count their moats but to widen them, every year. That is the real output of this framework: not the score, but the next twelve months of trajectory.
Appendix: Sample Diligence Rubric
Score each dimension 0 to 3. Total out of 27.
Scoring Anchors
| Moat | 0 | 1 | 2 | 3 |
|---|---|---|---|---|
| Data | Tribal knowledge only; nothing structured | Core data in ERP but incomplete; key knowledge held by individuals | Structured, maintained product/process/quality data actively used in decisions | Proprietary, compounding data asset (constraint rules, decades of process history) driving quoting, pricing, or automation |
| Workflow | Transactional PO supplier; one of several on the approved list | Preferred supplier; basic portal or EDI receipt | Embedded: EDI/API into customer MRP, customer tooling held, part numbers in customer drawings | Customer operations depend on you: VMI/consignment, sequenced delivery, engineering-change collaboration; displacement requires a customer-side project |
| Regulatory | No certifications beyond baseline | Table-stakes certs (ISO 9001 or equivalent) | Industry gate certs (AS9100, NADCAP, PPAP, UL, FDA, ITAR/CMMC) with qualified program status | Sole or dual-source qualified positions; permits or licenses effectively unobtainable new |
| Distribution | Spot and bid work; broker-dependent | Repeat direct accounts; no structural channel | Established channel: dealer network, distribution agreements, major-account approved status, freight-advantaged footprint | Proprietary or exclusive access: locked channel, specification position, private label at majors, footprint competitors cannot ship around |
| Ecosystem | None | Informal installer or user familiarity | Trained installer base, spec/CAD/BIM presence, accessory compatibility | Third parties build on your platform: certified installer programs, default-spec position, aftermarket ecosystem, large installed base |
| Network | None (the manufacturing default) | Informal multi-customer demand visibility | Portal or configurator aggregating demand signal across the channel | Aggregated signal materially improves forecasting, pricing, or product decisions in ways no competitor can match |
| Physical Infrastructure | Generic equipment; capacity easily replicated | Some specialized equipment or tooling | Significant specialized capacity; tooling libraries; multi-year, eight-figure replication cost | Effectively irreplaceable: permitted processes, unique capacity, freight-logical multi-plant footprint |
| Scale | Single plant; subscale purchasing | Regional scale; modest purchasing leverage | Multi-plant; meaningful purchasing leverage and fixed-cost absorption | Category-leading: national footprint, dual-site continuity for majors, automation funding rivals cannot match |
| Brand | No recognition beyond existing accounts; competes purely on price and spec | Respected within a region or niche; reputation limited to current customers | Recognized name in the category; channel or consumer preference; the safe choice that earns the benefit of the doubt | Category-defining name: measurable price premium, specified or requested by name, direct consumer demand |
Foundation Check
Foundations are the basic IT and operating systems the moats are built on: systems of record, plant systems, decision support, cybersecurity, and the talent, process, and change leadership that run them. They never add points. Rate each adequate or at-risk; any at-risk rating caps the related moat scores at 1 and flags a diligence issue. The full treatment of foundations belongs to the companion playbook.
| Foundation | At-risk looks like | Caps these moats |
|---|---|---|
| Systems of record (ERP, CRM) | Fragmented instances, dirty item masters, customer knowledge held by individuals | Data, Workflow, Scale |
| MES and plant systems | Paper travelers, no machine data capture, disconnected plant floor | Data, Physical Infrastructure |
| Decision support (BI) | Competing spreadsheets, no single source of truth, KPIs argued rather than read | Data, and the validation metrics themselves |
| Cybersecurity | No MFA, flat or unsegmented IT/OT network, no incident response plan, compliance gaps in regulated work | Regulatory, Brand, Workflow |
| Talent, process, and change leadership | No change control, hero-dependent IT, stalled project history, no business owner named for change initiatives | Every moat on the roadmap |
Validation Metrics (the Buffett layer)
Every claimed score of 2 or 3 must be supported by evidence. Four places to look:
- Pricing power. Gross margin stability through input-cost cycles. Surcharge pass-through history. Price increases taken without volume loss. Price premium relative to comparable competitors and win rate when not the low bid (this is where a claimed brand moat must prove itself).
- Rebid resilience. Win rate and margin retention on annual rebids and contract renewals.
- Switching evidence. Actual instances where a customer attempted to resource the work. What happened, how long it took, and whether they came back.
- The capital test. What would a well-funded entrant replicate in two years, and where specifically would they still fail?
A claimed moat that leaves no trace in these four places is a paper moat. Score the dimension down.
Interpretation
| Total Score | Profile | Read |
|---|---|---|
| 0 to 8 | Commodity converter | Defended only by capital and freight; price-taker; vulnerable to imports and insourcing |
| 9 to 13 | Defensible but attackable | One or two real moats; strategy should concentrate on deepening them |
| 14 to 17 | Durable | Four or more meaningful moats; structural defense across multiple dimensions |
| 18 to 27 | Franchise | Rare; verify with the validation metrics before believing it |
For each dimension, record trajectory alongside the score: wider, stable, or narrower than twelve months ago. Two identical scores with opposite trajectories are two different investments.
FAQ
What is a moat in a manufacturing business?
A structural barrier that lets the business sustain returns competitors cannot arbitrage away. The concept is Warren Buffett's: wonderful businesses are economic castles, rivals will assault any castle earning high returns, and management's job is to widen the moat every year. A real moat provides both a benefit and a barrier, and it must show up in the numbers - chiefly as pricing power. A moat that appears on the scorecard but never appears in the P&L is a narrative, not an asset.
What are the nine moats?
Data, workflow, regulatory, distribution, ecosystem, network, physical infrastructure, scale, and brand. The taxonomy comes from software investing, but half of those moats were borrowed from the physical economy in the first place. The twist in manufacturing: the strong and weak categories flip. A manufacturer usually has nothing but physical infrastructure, and struggles with everything digital.
What is a good moat score?
Score each moat 0 to 3, for a total out of 27. A business scoring 2 or higher on at least four dimensions (roughly 14 or higher in total) has durable, multi-dimensional structural defense. Below 9, the business is a commodity converter: defended only by capital intensity and freight. Trajectory matters more than the number - two identical scores with opposite trajectories are two different investments.
What is a commodity converter?
The profile most mid-market manufacturers score as: strong on physical infrastructure, partial credit on regulatory and distribution, brand rarely extending beyond a regional reputation, and weak to zero on data, workflow, ecosystem, and network. Defended by the capital cost of equipment and the freight cost of shipping around it. Attacked by imports, customer insourcing, and anyone willing to buy the same machines.
Which moats can be built inside a private equity hold period?
Data, workflow, network, and decision-support advantages can be built in two to three years. Distribution and channel reputation deepen on a five-year horizon. Regulatory qualifications, ecosystems, and consumer brands are decade-scale builds that are usually acquired rather than constructed. AI compresses construction, not the clock: coding agents can ship the tool in a quarter, but adoption, trust, and data accumulation set these horizons.
How do you validate a claimed moat?
Four places to look: pricing power (margin stability through input-cost cycles, price increases taken without volume loss), rebid resilience (win rate and margin retention on renewals), switching evidence (what actually happened when a customer tried to resource the work), and the capital test (what would a well-funded entrant replicate in two years, and where would they still fail). A claimed moat that leaves no trace in those four places is a paper moat.
This framework is grounded in Warren Buffett's concept of the economic moat, validated by four decades of strategy research from Michael Porter, Jay Barney, and Hamilton Helmer, and adapted for the modern era by investors including Gokul Rajaram, whose software-moat taxonomy informs the categories used here. The manufacturing translation and validation discipline draw on the author's firsthand experience seeing this approach applied within private equity diligence and value creation across mid-market manufacturing.
About the author. Mike Franklin has spent more than two decades operating inside manufacturing businesses, most of them private equity sponsored, on both sides of the technology relationship: the seats that consume it (VP of Marketing, operations leadership, commercial P&L) and the seats that deliver it (technology roles in startups, four tours as CIO/CTO). He has built product configurators and BOM rules engines, and run IT under a hold-period clock. He is the Managing Partner of Cold Iron Labs, a technology value-creation practice for private equity backed mid-market manufacturers, from diligence through fractional CIO and Chief AI Officer leadership.