Notes — Hamilton Helmer on 7 Powers

Four questions [Adler frame]

Q1 — What is it about? A framework for evaluating the structural determinants of lasting business value. Hamilton Helmer’s 7 Powers model defines exactly what makes a business durable against competition — answering when and how a company can earn superior returns on capital. Covers when to think about strategy, what power means (benefit + barrier), the full taxonomy of seven power types, how powers accumulate over a business lifecycle, and a brief framework for thinking about AI.

Q2 — How is it argued? Through precise definition (contrast with loose competitor use of “moat”), illustrative company cases (Netflix, Uber/Lyft, Toyota, TSMC, Intel, Microsoft), and a “to be or not to be test” (single-firm vs. competitive industry). Hamilton trained Netflix’s top 100 people on this framework; the cases are drawn from that teaching context. The AI section applies the same three-class lens from the electricity analogy.

Q3 — Is it true? The benefit + barrier definition is more precise than common uses of “moat” (which conflates the two sides). The network economies vs. network effects distinction is important and often violated by practitioners. The critique of data as a spurious scale economy is well-grounded. The AI taxonomy (technology / enabled companies / reconfigured incumbents) seems plausible but is predictive and untested as of the source date.

Q4 — What of it? For product builders: the framework changes the question from “do we have a moat?” to “what specifically is the mechanism of benefit and barrier?” Counter-positioning is the most relevant for early-stage; branding and process power are effectively unavailable to startups. The power progression gives a sequencing heuristic. The AI reconfiguration point implies that established product companies adopting AI well may compound existing advantages rather than be displaced.

Glossary

Power: a structural attribute of a business that produces both a benefit (cost, price, or quality advantage over the competition) and a barrier (a reason competitors cannot easily replicate it). Both are required; either alone is insufficient.

Barrier: the mechanism that prevents competitors from replicating a benefit. Hamilton’s term for what Buffett calls a “moat” — but moat = barrier only; it excludes the benefit side.

Benefit: the above-average return that the barrier permits — lower cost, higher price, better quality.

Counter positioning: a new business model that incumbents cannot adopt without damaging their existing business. The barrier is the incumbent’s own disincentive [?].

Scale economies: unit costs fall as output rises. True scale economy = material effect sustained at scale (curve does not flatten). Data scale economies are often false (curve flattens early).

Switching costs: costs a customer would incur by moving to a competitor. Includes financial, procedural, and relational components.

Network economies: network effects that produce a material, durable cost or price advantage relative to a competitor serving a smaller network. Weaker than often assumed: Uber/Lyft show network effects without material network economies.

Branding: durable price premium from accumulated customer trust and associations — requires years of consistent experience, therefore unavailable to startups.

Process power: operational complexity that is so embedded it is effectively inimitable, even though nothing proprietary prevents replication in principle. Toyota Production System as canonical case. Distinct from operational excellence (which is imitable and therefore treadmill, not power).

Resource power: unique asset — physical, intellectual, or relational — that cannot be replicated at any cost. Different class from the other six: governed by access, not by strategic action.

Operational excellence: continuous improvement in execution — necessary to compete but imitable. A treadmill: everyone runs it. Not power. [Contrast: process power]

Power progression: the sequence in which powers become accessible across a business’s lifecycle (origination → take-off → stability). Counter positioning earliest; branding and resource power latest (or never).

When to think about strategy

Hamilton’s answer: always — even pre-PMF. Strategy thinking does not require a finished product or established market position. The value is probabilistic: it tilts the odds, not determines outcomes. Founders often defer strategy as a post-PMF activity; Hamilton argues this loses the early counter-positioning window.

Power definition: benefit + barrier

The “to be or not to be test”: a power exists if its removal would collapse a firm from a durable competitor to an ordinary one. Both conditions must hold:

• Benefit: the firm captures above-average returns from the structural attribute.
• Barrier: competitors cannot replicate the attribute without structural disadvantage.

Buffett’s “castle with a moat” only names the barrier. Power requires the castle and the moat.

7 Powers taxonomy

Power progression

Startups should focus primarily on counter positioning in the origination phase — it is the one power directly available from the beginning. As the company grows:

• Take-off: scale economies, switching costs, network economies become accessible.
• Stability: branding and process power may accumulate if groundwork was laid early.
• Effectively removed from startup consideration: branding (requires years), process power (requires embedded operational complexity), resource power (externally governed — luck or acquisition, not strategy).

False powers founders claim

Three common errors:

1. Data scale economies: founders claim that proprietary data creates a durable cost advantage. Hamilton’s critique: the learning curve (marginal improvement from additional data) flattens quickly. After a threshold, more data adds little. Very few companies have demonstrated durable data-based power.

2. Flywheels without material effect: “flywheel” as a rhetorical term has been decoupled from the original Bezos meaning (compounding scale economy + switching cost + network effect combo). Most so-called flywheels are operational virtuous cycles — beneficial but not structurally defensive.

3. Early-stage branding claims: branding requires years of consistent customer experience to build trust. A pre-PMF company claiming brand power is conflating awareness or enthusiasm with the structural attribute.

Network economies vs. network effects

Network effects (more users = more value for each user) are common. Network economies (material, durable cost/price advantage from serving a larger network than a competitor) are rare.

Uber/Lyft case: both have network effects. Neither has network economies — driver-to-rider matching is geographically local; a competitor in the same city can replicate the density. Uber won through: (1) geographically specific, modest scale economies; (2) a protracted war of attrition that depleted Lyft’s capital. Not a network economy story.

Process power vs. operational excellence

Porter’s insight (extended by Hamilton): operational excellence is a treadmill, not a power. Every company must run it; competitors can copy it. It is necessary to compete but insufficient to produce above-average returns.

Process power is the narrow exception: when operational complexity becomes so embedded, path-dependent, and opaque that competitors cannot replicate it even in principle (not just in practice). Toyota Production System; TSMC’s yield process. Rare. Unavailable to startups.

Three drivers of company value

Hamilton’s exhaustive set:

1. Power — structural competitive advantage (benefit + barrier)
2. Market size — total addressable opportunity
3. Operational excellence — execution quality

No other variable matters at the fundamental level. “You can always add more variables, but they reduce to these three.”

AI framework

Hamilton’s three-class model (analogy: electricity adoption):

1. Technology itself: companies whose primary product is the AI technology (Intel for AI = GPU/chip vendors, model providers). Enormous power possible if the technology is proprietary; otherwise commodity risk.

2. Companies that couldn’t exist without AI: entirely new business categories enabled by AI capabilities (analogous to Microsoft enabling personal computing). These may be the next generation of dominant platform companies.

3. Companies that existed before/after AI but use it to reconfigure: Hamilton’s view is that this class — established companies deeply integrating AI into operations — will produce the largest aggregate economic impact, analogous to electricity’s reconfiguration of factory floors or the 1990s business process re-engineering wave.

Hamilton’s bet: the biggest impact of AI will be tertiary-class reconfiguration of existing industries, not the primary or secondary classes.

See also

• Hamilton Helmer on 7 Powers
• 7 Powers
• DHM Model
• Gibson Biddle on Product Strategy