Capability, Field, Moat

1Introduction

New technologies rarely create durable enterprise value on their own. Whoever masters a capability first often loses the returns to firms that control the necessary complementary assets (Teece 1986), or sees the innovation absorbed by incumbents when it improves their existing business rather than challenging it (Christensen 1997). For a holding company with a long investment horizon, the decisive question is therefore less which technology prevails than under which conditions its value stays with a firm.

We separate two questions that are often conflated in practice. The first concerns location: in which configuration of technology, market and competition can a venture retain value at all? The second concerns mechanism: how is value created for the individual customer? Part I answers both generically, without reference to particular technologies or industries (Sections 2 to 4), and sets out the foundations on which the thesis rests (Section 5). Part II applies the model to three firms and compares them (Sections 6 to 11). Methodologically, the paper follows theory-building from case studies: the cases were chosen to differ in capability, field and moat, so that the model is sharpened by contrasts rather than only by confirmations (Eisenhardt 1989; Yin 2018). The cases are anonymized. Throughout, the thesis is treated as an established premise, and the paper argues from it rather than towards it. Section 12 defines the scope of the thesis, Section 13 derives principles for the holding, and Section 14 concludes.

2The triangle

2.1 Three conditions

The capability (C) is a technological possibility that has only recently become economically available; it answers why a venture is possible now and not earlier. The field (F) is an area of application in which a buyer has urgent demand and the willingness to pay. The moat (M) is a structure that prevents platform providers or incumbents from capturing the value created, in the sense of a durable competitive advantage (Porter 1985; Helmer 2016).

2.2 Formalization

Each condition is rated on the interval [0, 1]. The value of a venture is their product:

The multiplicative form is the central modeling assumption. An additive form would let the strength of one corner compensate for the weakness of another; in the multiplicative form, a missing corner sets the value to zero. The decision rule is that every corner must reach a threshold θ:

Operationally, (2) corresponds to three questions, each requiring a clear yes: Which capability makes the venture possible only now? Who pays urgently, and why? What prevents platforms or incumbents from replicating it? The three two-corner configurations describe typical failure modes (Table 1).

3Value compression

3.1 Concept

The triangle states where value can be retained durably, not how it is created for the customer. For this we introduce the concept of value compression. A customer process typically contains many frictions: tasks, hand-offs, sources of error and consequential costs, dispersed across people, places and points in time. A venture compresses value when it takes over as many of these frictions as possible and hands the customer a single, reliable outcome instead.

Value compression is defined from the customer’s side. What counts is not what a supplier delivers, but the compression the customer receives – the frictions actually removed from their operations – and the compression the customer perceives – the frictions they notice as resolved and attribute to the venture. This follows the view that value is determined by the customer in use rather than embedded in the offering (Vargo & Lusch 2004), and that purchasing decisions rest on perceived rather than objective value (Zeithaml 1988). The two can diverge: a customer may perceive more compression than they receive, or receive more than they perceive.

The idea builds on the call not to automate work but to obliterate it (Hammer 1990), and on the view that customers do not buy products but hire them to get a job done (Christensen et al. 2016). It differs from uses of the term value compression that denote falling prices or margins; here it denotes the compression of many frictions into one outcome.

3.2 Friction profile

To make frictions comparable, we assign them to seven dimensions (Table 2). The friction profile of a field indicates in which dimensions frictions accumulate.

3.3 Measure

Let a customer process contain n frictions with costs c₁, …, c_n per period. A venture takes over the subset K and leaves the customer t ≥ 1 touchpoints that the customer still has to handle. Received compression is the cost actually removed from the customer’s process:

Perceived compression weights each removed friction by the share π_i ∈ [0, 1] to which the customer notices it as resolved and attributes it to the venture:

The degree of compression, as experienced by the customer, is:

V_r measures how much cost the venture removes; V_p how much of it the customer sees; κ how strongly it is bundled. The limiting case is a venture that takes over all frictions and leaves the customer exactly one touchpoint: the outcome. The difference V_r − V_p is value received but not perceived. It tends to grow as compression succeeds: a friction that no longer occurs is no longer felt. Perceived compression above received compression, in turn, is a promise the customer will eventually check.

3.4 Standardization as a boundary

Value compression is not possible everywhere. Retail and aviation have already removed many of their frictions by other means: standardization. In retail, the bar code captures every sale at the checkout; since the introduction of the Universal Product Code in the 1970s, every outflow of goods has been an enforced, machine-readable event from which stock levels and replenishment can be derived (Brown 1997). In aviation, regulation, standard operating procedures and checklists have unified processes to the point where deviations are rare and responsibilities unambiguous (Gawande 2009).

Standardization is thus an earlier, organizational form of compression. Where it has taken place, the residual potential is small. One piece of evidence in this direction is the case of Walmart: in 2020 the company ended its use of camera-equipped shelf-scanning robots in about 500 stores; according to reports, it had found that employees achieved similar results (Associated Press 2020). Shopper reactions were reported as a further reason; the core finding matches the thesis. Hospital storerooms differ from retail in exactly the decisive respect: the withdrawal of supplies is not an enforced capture event. Material leaves the shelf without a checkout, and capture falls back on people.

With the degree of standardization σ ∈ [0, 1] of a field, the residual potential for compression can be written as:

Compression pays where friction costs are high and standardization is low – in dispersed, changing processes carried by people.

3.5 Compression and the triangle

Value compression connects the three corners. The capability determines which frictions can be taken over at all. In the field, compression shows up as willingness to pay, which follows perceived compression V_p. And received compression creates a moat: a competitor must take over all compressed frictions, not just one, and a customer who receives them has little reason to switch. In this view, verticalization is the most far-reaching form of compression – the firm takes responsibility for the outcome itself, and the customer is left with a single contract. Bureaucracy can be read as the sum of uncompressed coordination costs (Coase 1937; Williamson 1985).

4Five predictions

The thesis makes five predictions about how successful ventures differ from unsuccessful ones. Each is stated with the measure by which it shows up in practice.

5Foundations

Each component of the thesis rests on established research or on well-documented practice (Table 3). The thesis combines them; its contribution is the combination and the concept of value compression.

6Assignment

We populate the corners of the triangle with specific trends (Table 4, Fig. 2). For the physical-AI ventures the capability is the same; field and moat vary. Case C is deliberately chosen as a contrast outside this hub: a software firm without physical AI, which shows what the capability contributes to compression. All three cases are anonymized; the descriptions draw on the firms’ publicly available materials, which are not cited in order to preserve anonymity.

7Combinations

With one capability, three fields and two moats there are 1 × 3 × 2 = 6 combinations (Table 5). The assessment is the author’s.

In T1, sovereignty is a reason to buy, not a label: critical infrastructure is not guarded by systems whose data or updates are controlled outside Europe. In T2, T3 and T5, the firm delivers the service itself and compresses all the way to the outcome. T3 has the largest market and the highest operational demands in staffing, licensing and liability. In T4 and T6, sovereignty is more a mandatory requirement than a reason to buy; these combinations hold only if public buyers explicitly demand it.

7.1 Friction profiles by field

Table 6 sketches typical frictions for each field and the outcome into which they can be compressed. The profiles are preliminary; for hospital logistics the profile is elaborated in Section 8.

7.2 The hub

The three profiles share three dimensions from Table 2: dispersion, duration and change. These frictions cannot be taken over by software alone; they require systems that stay on site permanently, operate without constant connectivity and decide locally. A vehicle on the seabed, a sensor in a bedroom and a camera in a hospital storeroom perform the same task: watching over something valuable. Under H4, this is why a holding company can build the capability once and deploy it across several fields.

8Case A: hospital supply logistics

8.1 Friction profile

In hospital storerooms, supplies are counted, scanned and reordered by hand. The firm in Case A mounts cameras on the shelves that continuously capture stock levels, warn of stockouts and expiring items, and trigger replenishment through the hospital’s ERP system. Table 7 assigns the frictions of the process to the dimensions in Table 2.

The profile covers all seven dimensions and is densest for people. This explains why training and process instructions do not solve the problem durably: they address individual frictions and themselves depend on engagement and skills. Compression instead addresses all twelve at once:

“The shelf is always correctly stocked.” Twelve frictions, one outcome: |K| = 12, t = 1, hence κ = 12 by (5). The value is illustrative and depends on the granularity with which frictions are counted.

According to the firm, the system counts with high accuracy; in a case study at an academic medical center it paid for itself within a year and released a substantial share of tied-up working capital. These are the firm’s own figures. They concern received compression and match H2: payback within a year corresponds to a large V_r. Perceived compression is a separate question. Once the shelf is reliably stocked, the counting, scanning and walking no longer happen and are no longer felt; alerts on imminent stockouts and expiring items are what keep the removed frictions visible to the customer.

8.2 Assessment in the triangle

8.3 The path to the third corner

The value of 0.5 shows where the model sees the lever. The firm compresses twelve frictions into one outcome but leaves responsibility for that outcome with the customer. Part I suggests two paths. First, T5, compression all the way to the contract: the firm takes responsibility for a hospital’s stock and replenishment and shares in the working capital released instead of being paid per camera; the customer is left with one touchpoint. Second, for European hospitals, T6: sovereign operation with data that does not leave Europe.

9Case B: subsea infrastructure

9.1 Context

Subsea infrastructure – cables, pipelines, quay walls, wind-farm foundations, dams – is hard to monitor. Below the surface there is no GPS, no radio and no light, and the water column cannot be observed from orbit; according to the firm, assets are effectively operated blind, with years between two inspections. The urgency is publicly documented: in October 2023 the Balticconnector gas pipeline was damaged, according to Finnish investigators probably by the anchor of a cargo ship (Al Jazeera 2023); after further cable damage in the Baltic Sea, NATO launched the Baltic Sentry mission in January 2025 to protect critical undersea infrastructure (Belga News Agency 2025).

The firm in Case B, a European start-up founded in 2026, is developing resident autonomous underwater vehicles: they remain at the asset, dive without a vessel, tether or pilot, survey cables, foundations and walls with camera and sonar, surface for a position fix and transmit status and alerts to shore by radio. First deployment areas are ports, harbor basins, inland waterways and cable landfalls. The system is under development; the following describes the venture, not demonstrated performance.

9.2 Friction profile

“We know at all times how our asset is doing under water.” Nine frictions, one outcome: |K| = 9, t = 1, hence κ = 9 by (5), assuming the system performs as described.

The degree of standardization is low: under water there is no event that captures an asset’s condition by itself. By (6), the residual potential is correspondingly high.

Case B also shows the widest possible gap between received and perceived compression. Much of what continuous monitoring delivers is prevention: damage detected early or deterred altogether. A prevented incident leaves no trace, so the customer perceives the condition picture, but not the outage that did not happen. For the field corner this means that willingness to pay rests on the perceived picture and on public events such as those above, not on the received value alone.

9.3 Assessment in the triangle

Case B is the only one of the three that occupies all corners – two of them, however, only as a claim. Under the thesis, its success turns on two things: demonstrating the capability in its first waters, and buyers paying for sovereignty. Because the vehicles operate without vessel and pilot, the operator model of T2 also suggests itself: monitoring as a service instead of selling vehicles.

10Case C: social-care documentation

10.1 Context

Social-care providers – youth welfare, disability support, assisted living, addiction services – are subject to extensive documentation, reporting and billing obligations toward funding bodies and statistics. In German addiction services, for example, providers record a standardized national core data set whose versions change over time.

The firm in Case C offers cloud software for these providers with a digital client record, service recording and documentation, document management, staff management, service billing with e-invoicing, and scheduling. In an anonymized example, the vendor reports that staff gained about two hours per week for client work – a measure of perceived compression, as reported by users, rather than an observed one.

10.2 Friction profile

As in Case A, the profile covers all seven dimensions. The difference lies in the depth of compression. A camera on a shelf replaces capture; documentation software structures it. In Case C, the system bundles the frictions in one application, but the input itself remains with the staff: in documentation, in service recording and in time recording.

“Service delivered, documented and billed.” Eight frictions, three remaining touchpoints: |K| = 8, t = 3, hence κ ≈ 2.7 by (5). Illustrative; the number of touchpoints is the author’s estimate.

The case also shows a variant of standardization from Section 3.4. The core data set standardizes the data, not the event: unlike at a checkout, the record does not arise by itself but must be produced by people. Standardized data without a standardized event therefore increases friction instead of reducing it.

Unlike in Case B, the gap between perceived and received compression is small: the remaining input stays with the staff, so they feel directly how much of it has gone.

10.3 Assessment in the triangle

According to the model, the levers lie in two corners. In capability: technology that replaces input rather than structuring it – for example, generating documentation from what is spoken or captured anyway – would reduce t and increase κ. In the moat: taking over documentation and billing as a service for the provider, that is, verticalization in the sense of Section 3.5. Both are derivations from the model, not statements about the firm’s plans.

11Cross-case comparison

* intended, not yet demonstrated.

The comparison adds four lessons to the thesis.

12Scope

The thesis applies to ventures that take over frictions in a customer’s process. It is strongest where those frictions are physical, dispersed and changing, and where no standardization has removed them already. It says little about consumer markets driven by brand or about fields that are already standardized. The ratings in Tables 5, 8, 10, 12 and 13 are the author’s assessments; figures about the firms are as the firms report them, and for Case B they describe a venture under development.

13Principles for the holding

Taken as established, the thesis yields six principles for allocating capital and developing portfolio companies.

1. Three yeses or no deal. A venture qualifies only if every corner clears the threshold in (2). Two strong corners do not compensate for a missing third.
2. Go where standardization is absent. Prefer fields with low σ by (6): dispersed, changing processes carried by people, with no enforced capture event.
3. Replace capture, do not structure it. Prefer capabilities that perceive and act over software that organizes human input (L1); they reach a single touchpoint.
4. Compress to the contract. Move portfolio companies from selling tools towards owning the outcome; verticalization closes the moat corner (L3).
5. Measure both compressions. Track received compression V_r and perceived compression V_p for every portfolio company, and manage the gap; make removed frictions visible to the customer (L4).
6. Build the capability once. Develop physical AI as a shared hub and deploy it in fields with a similar friction profile (Section 7.2).

14Conclusion

Durable value arises in the conjunction of capability, field and moat, and it is created for the customer through value compression: many dispersed frictions taken over and one reliable outcome delivered, received and perceived. Compression succeeds where standardization is absent – not in retail and not in aviation, but in hospital storerooms, under water and in social care. Its depth is decided by the capability: systems that perceive and remain on site replace capture; software alone structures it. And because successful compression makes removed frictions invisible, received value must be made perceivable to be paid for. The name of the company sums up its subject: curatio is Latin for care and stewardship.

Suggested citation. Mohr, J. (2026). Capability, Field, Moat: A Thesis of Durable Enterprise Value from Conjunction and Value Compression, Developed from Three Case Studies. Curatio Formatio Working Paper No. 1, version 6. Hamburg.

References

1. Al Jazeera (2023, October 25). Finland says Chinese vessel’s broken anchor caused Balticconnector damage. aljazeera.com.
2. Associated Press (2020, November 3). Walmart abandons shelf-scanning robots, lets humans do work. techxplore.com.
3. Belga News Agency (2025). NATO launches Baltic Sentry programme to protect undersea infrastructure. belganewsagency.eu.
4. Brooks, R. A. (1991). Intelligence without representation. Artificial Intelligence, 47(1–3), 139–159.
5. Brown, S. A. (1997). Revolution at the Checkout Counter: The Explosion of the Bar Code. Cambridge, MA: Harvard University Press.
6. Christensen, C. M. (1997). The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Boston: Harvard Business School Press.
7. Christensen, C. M., Hall, T., Dillon, K. & Duncan, D. S. (2016). Competing Against Luck: The Story of Innovation and Customer Choice. New York: HarperBusiness.
8. Coase, R. H. (1937). The nature of the firm. Economica, 4(16), 386–405.
9. Draghi, M. (2024). The Future of European Competitiveness. Brussels: European Commission.
10. Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532–550.
11. Gawande, A. (2009). The Checklist Manifesto: How to Get Things Right. New York: Metropolitan Books.
12. Hammer, M. (1990). Reengineering work: Don’t automate, obliterate. Harvard Business Review, 68(4), 104–112.
13. Helmer, H. (2016). 7 Powers: The Foundations of Business Strategy. Deep Strategy.
14. OECD (2016). The Ocean Economy in 2030. Paris: OECD Publishing.
15. Porter, M. E. (1985). Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press.
16. Shelanski, H. A. & Klein, P. G. (1995). Empirical research in transaction cost economics: A review and assessment. Journal of Law, Economics, & Organization, 11(2), 335–361.
17. Teece, D. J. (1986). Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research Policy, 15(6), 285–305.
18. United Nations, Department of Economic and Social Affairs, Population Division (2024). World Population Prospects 2024. New York: United Nations.
19. Vargo, S. L. & Lusch, R. F. (2004). Evolving to a new dominant logic for marketing. Journal of Marketing, 68(1), 1–17.
20. Williamson, O. E. (1985). The Economic Institutions of Capitalism. New York: Free Press.
21. Yin, R. K. (2018). Case Study Research and Applications: Design and Methods (6th ed.). Thousand Oaks, CA: SAGE.
22. Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value: A means-end model and synthesis of evidence. Journal of Marketing, 52(3), 2–22.