7 Powers: The Foundations of Business Strategy

The bottom line

A business is worth its market scale times its Power, the conditions that let it keep above-normal returns despite competition. Builders must invent their way to one of seven Powers in the window when it can form, because operational excellence alone gets copied.

Key takeaways

• A business is worth its market scale multiplied by its Power. Competitors pull returns down to normal unless something stops them. Introduction
  • Intel brought the same management and first-mover status to memories and microprocessors. Memories ended worth $0 and Intel exited. Microprocessors grew to some $50B in revenue and around a $150B market cap.
• Power needs a Barrier, and operational excellence alone does not supply one. Lower costs and better quality are common, and rivals copy them unless the gains take years to reach. Ch. 7 · Process Power
  • Toyota held 0.1% of the US market in 1969 against GM's 48.5%. By 2014 Toyota had nearly matched GM and Ford, and GM had fallen under 20%.
  • GM's 1984 NUMMI venture got defect rates close to Toyota's Japanese plants. Toyota shared everything, yet GM could not spread the results to its other plants.
• A firm can reshape its industry's costs to create Scale Economies. Turning a variable cost into a fixed one lets the biggest player spread it thinner than anyone else. Part I, Strategy Statics
  • Original content made programming a fixed cost for Netflix. A $100M show spread over 30M subscribers costs a bit over $3 each, while a rival with 1M subscribers pays $100 each.
• A newcomer can beat a strong incumbent with a better model the incumbent will not copy. Copying would damage the incumbent's profitable core, so staying put is the rational choice. Ch. 3 · Counter-Positioning
  • Vanguard's index fund launched in August 1976 with only $11M and took over a decade to catch on. Assets passed $3T by the end of 2015. Fidelity had the skills to respond, but cheap passive funds would have cannibalized its high-fee active business.
• Every Power starts with invention, which must then be steered toward a Barrier. Copying creates no Power, and most inventions are easy to copy too. Part II, Strategy Dynamics
  • Netflix launched streaming on January 16, 2007 with about 1000 titles. Its interface, recommendations and cloud setup could all be copied. The Epix exclusive on August 10, 2010 made some content a fixed cost, so scale finally mattered.
• Each Power can only be first established in its own stage of growth. Share leads and customer lock-in come cheap only during takeoff, and the chance does not return. Ch. 9 · The Power Progression
  • Intel's Operation Crush targeted 2000 design wins in a year. It landed the IBM PC, which shipped August 12, 1981 with the 8088 and sold 750,000 units in its first year.
  • Apple's buggy Apple III cost over $4000, against $1600 for the IBM PC a year later. Apple lost share steadily and later nearly died.

Always look to the Barrier first.

Hamilton Helmer, Introduction

A business is worth market scale times Power, and great execution alone does not create Power

Lasting value comes from Power that survives competition, not from smart management, first-mover status or flawless execution.

• Competition drives returns down to normal, so common strengths cannot explain lasting value. Introduction
  • Smart, informed managers and first movers are everywhere. Cleverness is necessary but far from sufficient.
  • Intel had the same leaders and the same head start in memories and microprocessors, yet only one business escaped arbitrage.
  • Noyce and Moore left Fairchild in 1968 and first called Intel "The Memory Company." Microprocessors began as a job for Busicom, a Japanese calculator maker, taken on to raise cash for memories.
  • Microprocessors faced at least as much competition as memories did: IBM, Motorola, AMD, Zilog, National Semiconductor, ARM, NEC, TI and others. Escaping arbitrage despite that is what Power means.
  • Game Theory's idea of strategy is too broad to guide value creation. Exiting memories was an optimal move, yet it created no Power.
  • A Google Scholar search for "strategy" returns 5,150,000 hits. So a company's strategy is defined narrowly: a route to continuing Power in significant markets.
• Value equals market scale times Power, and Power has to last. Introduction
  • The Fundamental Equation writes value as NPV = M0 x g x s x m.
  • M0 x g is market scale. Holding share s at a positive differential margin m is Power.
  • A firm growing 10% a year gets only about 15% of its value from the next three years.
  • M0 is current market size and g is the discounted growth factor. s is long-term share, and m is net margin above what covers the cost of capital.
  • The equation gives potential value only, and operational excellence realizes it. Under Andy Grove that excellence ran both Intel businesses. Still, memories' margin went negative while microprocessors' stayed high.
  • A few good years of margin that then fade are worth little. Faster-growing firms hold even more of their value in the out years.
  • Investors' long-term models reflect this. Quarterly results still matter because they signal future performance.
• Power needs a Benefit plus a Barrier, and the Barrier is the scarce part. Introduction
  • The Benefit materially raises cash flow through higher prices, lower costs or lower investment.
  • The Barrier stops competitors from arbitraging the Benefit away.
  • Every major cost-cutting effort counts as a Benefit. That is why a Benefit alone proves nothing.
  • Check Power against each competitor one by one: potential, existing, direct and functional. A single arbitrageur can erase differential margins.
  • Judge each strategically separate business on its own, even inside one corporation. Issues across several businesses under one roof belong to Corporate Strategy, which lies outside the framework.
  • Keep industry economics apart from the firm's own competitive position.
• Flawless execution cannot save a strategy that gives away the sources of advantage. Foreword by Reed Hastings
  • Execution must take most of everyone's time, since failing at it guarantees stumbling.
  • The IBM PC was executed brilliantly. Yet IBM outsourced the OS and the chip, and with them its lasting advantage.
  • The PC had 40,000 customers on announcement and over 100,000 in its first year. It was backed by an inspired Charlie Chaplin campaign.
  • IBM let Microsoft sell the OS to others. That gave up the network economy that had powered its System 360 mainframes. IBM also handed the microprocessor to Intel while promoting applications tied to that chip.
  • The PC became an unattractive box-assembly business. IBM sold it to Lenovo in a 2005 fire sale.
• Strategy has to be simple enough to teach and complete enough to miss nothing. Foreword by Reed Hastings
  • Netflix managed through context, not control, so leaders could not impose a strategy from the top.
  • Helmer taught strategy to many of Netflix's key people. The standard was Feynman's test: what you cannot teach, you do not understand.
  • Helmer and Larry Tint of Strategy Capital, a hedge fund invested in Netflix, met Hastings on Sept. 29, 2004. Helmer presented his Power Dynamics framework instead of probing management.
  • By 2009 Netflix was on track for almost $1.7 billion in sales. But DVD by mail was transitional, and Google, Amazon, Time Warner and Apple loomed.
  • Feynman agreed to lecture on a hard area of Quantum Mechanics, then backed out days later: You know I couldn't do it…That means we really don't understand it.
  • Existing frameworks failed the test, because the simple ones were simplistic and the full ones were not simple. The goal is a prepared mind for rare, high-flux decisions, after Pasteur.

Make share too expensive for rivals to buy: Scale and Network Economies

Scale and network leaders keep their share because any rival trying to buy it would lose more than it gains.

• A scale leader turns every share grab into a losing trade for smaller rivals. Part I, Strategy Statics
  • Unit costs fall with volume, so the leader can match any price cut from a stronger cost position.
  • After a few bouts, followers expect retaliation and stop attacking. That expectation is the Barrier.
  • Intel built Scale Economies in microprocessors. AMD challenged it doggedly for a very long time. Intel answered every move with its scale-based economics, stayed a great business, and left AMD in persistent pain.
  • The follower has no good move. Matching the leader's offer hurts its profits, and offering less loses customers. That dead end is the mark of Power.
  • Fixed costs are the common source, especially in tech. Volume/area relationships, distribution network density, learning and purchasing economies also count.
  • The Barrier does not maintain itself. The leader has to keep retaliating visibly so followers keep learning that attacks do not pay.
• Scale Power needs both a scale-intensive industry and a real scale lead. Part I, Strategy Statics
  • Surplus Leader Margin, the leader's margin when prices leave the follower at zero profit, splits into scale intensity and scale advantage.
  • Both must be clearly positive. A big fixed cost with no scale gap gives no Power.
  • The steps: model cost as c times Q plus C, and profit as (P minus c) times Q minus C. Set P so the weak firm earns zero. Then SLM = [C / Leader Sales] times [(Leader Sales / Follower Sales) minus 1].
  • The first term reflects industry structure and the second reflects relative position. Both can be targets of strategy. The formula covers only the fixed-cost case.
  • Streaming first looked weak for Power, since falling IT and cloud costs lowered entry barriers. In 2011 Chief Content Officer Ted Sarandos pushed for exclusive rights. In 2012 Netflix committed heavily to originals, starting with House of Cards.
  • The move looked risky. Originals cost more, and Netflix's earlier Red Envelope Entertainment had gone badly. The key insight took four years in streaming to emerge.
• Power sets potential value, and execution decides how much of it is realized. Part I, Strategy Statics
  • After Netflix built streaming Power, its stock rose six-fold in six years while the market doubled.
  • Its 2011 plunge came from operating errors, yet the Power held.
  • The climb was uneven. 2010 to 2013 was a roller coaster, because in high-flux markets cash flow can lag Power.
  • The missteps were not fatal only because the strategy stayed intact. Without operational excellence, Power stays a potential.
• Once a network tips, even a rich, competent challenger cannot catch up. Ch. 2 · Network Economies
  • Each user gains as the installed base grows, so the leader can charge more. A follower would need an unthinkable discount to close the gap.
  • Google+ could not unseat Facebook. Facebook beat MySpace by getting the product right early.
  • The network version of SLM assumes all costs are variable and prices the follower at cost. The leader then adds δ times the base gap. SLM = 1 - 1/[1 + δ(SN - WN)], which nears 100% as the gap grows.
  • δ is the benefit each existing member gets when one more joins, divided by variable cost per unit. It measures network strength against industry costs.
  • Indirect effects run through exclusive complements. A new phone OS starts with few apps, and developers will not spend scarce resources on a small market.
  • IBM mainframes, Microsoft's operating systems, Steinway Pianos and Exchange Traded Funds were built on this Power. LinkedIn's HR Solutions Suite charges more because its value comes from LinkedIn's user count.
• Network Power ends at the network's edge, and some networks can never pay. Ch. 2 · Network Economies
  • BranchOut tried to carry Facebook's base into professional networking and collapsed. Facebook at Work failed too.
  • If network size N times δ is below unit cost, no firm can profit.
  • Rick Marini launched BranchOut in 2010 to catch LinkedIn's 70M members from Facebook's base, almost 10x larger. Funding reached $49 million and users peaked near 14 million in spring 2012. Engagement was thin. Facebook banned its spammy wall posts, usage collapsed, and Hearst bought the assets in September 2014.
  • Facebook and LinkedIn both thrive because users wall off personal life from work. The boundaries of the network effects determine the boundaries of the business.
  • Twitter shows the viability risk: heavy upfront investment, then uncertain monetization. Management often takes the blame when the real problem is the economics.
  • Real network value is not linear. A classmate on the network is worth far more to a US college student than a stranger in Ulan Bator.

Customers you already have will pay more rather than face the cost of leaving: Switching Costs

When customers would lose value by moving to another supplier, the firm that holds them can charge more on their later purchases, and a challenger must pay them to leave.

• Lock-in lets a firm charge more than rivals for equivalent follow-on products. Ch. 4 · Switching Costs
  • Switching Costs arise when a customer values compatibility across repeat or complementary purchases from one supplier.
  • SAP customers often dislike the product yet keep paying, because staying is the safer choice.
  • A challenger must compensate customers for what they lose by moving, so share gains cost it more than they return.
  • Compuware studied 588 SAP customers in Europe and the US. 43% were unhappy with response times across all components, 50% could not predict SAP performance, and nearly all saw financial risk in performance problems.
  • A separate survey of more than 1,000 customers found 89% expected to keep paying annual SAP maintenance fees. Low satisfaction paired with high retention is the signature of Switching Costs.
  • A simple model makes the leverage concrete. Firm S holds the customers and firm W does not. With equal utility and no fixed costs, S can price at W's price plus the switching cost ∆. S's margin when W prices to earn zero profit measures its edge.
• The cost of moving comes in three kinds, and it can far exceed the product's price. Ch. 4 · Switching Costs
  • Financial: plain money, such as a new database and all its complementary applications.
  • Procedural: lost familiarity, retraining, and the risk of errors during the transition.
  • Relational: broken bonds with sales and service teams, with the product and user identity, and with a user community.
  • In May 2004 Christina Hanger, HP's Senior Vice President of American Operations, moved the North American server sales divisions, $7.5 billion in revenue, onto SAP. She had already run five SAP migrations after the Compaq acquisition.
  • She budgeted three weeks for the order-entry changeover, three weeks of extra server inventory and extra factory capacity in Omaha. After go-live in early June, as many as 20 percent of server orders stalled between the legacy system and SAP for the rest of the month. Customers went to Dell or IBM, and CEO Carly Fiorina reported a $160M hit.
  • Apple's iTunes downloads used a proprietary format. A customer who moved to another program lost past purchases, which kept many buyers in place.
  • The risk of a move is itself a reason to stay: No one ever got fired for sticking with SAP.
• The premium pays only if there are follow-on products to sell, so holders keep widening the line. Ch. 4 · Switching Costs
  • The Benefit covers later sales to current customers. It gives no edge with prospects.
  • SAP builds a long list of add-ons and runs an ambitious acquisition program.
  • Acquisitions act as outsourced development and extend the product line faster than building alone.
  • A wider portfolio raises all three kinds at once. Deeper integration into customer operations makes untangling harder. Extensive training adds retraining costs and bonds with the supplier. More revenue falls under the lock-in.
  • Locked-in users can seed other Powers. Linking users and complements can create Network Effects, and preference that spreads to new customers can become Branding.
• Every firm with customers enjoys this Power, so the gains go to whoever signs customers before rivals bid the price of winning them up. Ch. 4 · Switching Costs
  • The Power is non-exclusive, and its intensity comes from industry economics that all players face equally.
  • Position is binary: a firm either has the customer or it does not.
  • A large technology shift can erase the lock-in entirely.
  • In mature markets every player can value an acquired customer. Firms then compete hard for new ones, and that competition arbitrages away the Benefit of acquiring them.
  • ERP vendors fear being leapfrogged by cloud applications, which could let customers leave without paying the old costs of moving.
  • A challenger should budget for customers' full switching costs. Even transition help it offers tends to be costly and imperfect.

Beat a strong incumbent with a model it cannot copy without damaging its own core

A challenger with a better business model is safe from a strong incumbent when copying that model would cost the incumbent more in its old business than it would gain.

• An incumbent that ignores a better model is often making a sound calculation, not a mistake. Ch. 3 · Counter-Positioning
  • The Benefit is a model with lower costs, higher prices, or both.
  • The Barrier is collateral damage. Matching the model would eat into the profits the incumbent already earns.
  • On May 1, 1975, John C. Bogle won over a reluctant Wellington Management board. Vanguard was owned by its funds and ran at cost. A year later it dropped sales commissions.
  • The fund held just $17M by mid-1977. Brokers disliked a product that implied they added no value.
  • The economics favored the challenger. Active funds as a group must earn the market's gross return, and their higher fees push average net returns below passive funds. From 2007 to 2013, active mutual funds lost $600B while ETFs and domestic equity mutual funds gained more than $700B.
• Incumbent inaction has three causes, and they often work together. Ch. 3 · Counter-Positioning
  • Milk: the new model pays on its own, but once cannibalization is counted, its combined NPV is negative.
  • History's Slave: routines built on past success make an unproven model look worse than it is. Job Security: the leaders' pay or standing suffers if the core is upended.
  • The test runs in order. Is the new approach attractive to the incumbent as a stand-alone bet? If not, this is not Counter-Positioning. If it is and the combined NPV is negative, it is Milk. If the NPV is positive but the CEO sees deeper losses, it is History's Slave. If incentives diverge from firm value, it is Job Security.
  • Fidelity under Ned Johnson faced a sharp revenue drop from losing high fees and loads. Passive funds would also have undercut its own case for active management. Johnson's question, Why would anyone settle for average returns?, shows bias on top of the math.
  • Kodak fails the first test. Its Power came from film. Digital meant semiconductor storage, where Kodak brought nothing and prices were heading toward commodity. Kodak was well managed. Calling its business "image storage" would not have given it semiconductor skills.
• The incumbent eventually responds, usually too late, because the Barrier wears down. Ch. 3 · Counter-Positioning
  • As the core shrinks and the new model proves itself, the expected collateral damage falls.
  • Dell vs Compaq, Amazon vs Borders, Schwab vs Merrill Lynch and Netflix vs Blockbuster all followed this path.
  • The common script runs through Denial, Ridicule, Fear, Anger and Capitulation. Nokia led smartphones in 2008 and was gone by 2014. Stephen Elop's 2011 Burning Platform memo admitted that the iPhone and Android had overtaken it. It reads as the Anger stage.
  • Disney launched DisneyLife streaming in the UK in 2015 and planned to expand across Europe. It kept the service out of the US, its biggest market, because it would overlap with Disney's cable and satellite deals.
  • Incumbents often dabble, testing the new model without enough commitment to answer the threat. They should either commit or milk the core on purpose. Watch for the signal: the challenger gains share fast with strong or promised profits, while the incumbent loses share, has no counter, and shakes up management.
• Challengers should pick fights the incumbent loses even when it plays its best game. Ch. 3 · Counter-Positioning
  • High-margin incumbents have the most to lose, so they are the best targets.
  • Price low at first and avoid boasting, which delays the incumbent's clear-eyed reassessment.
  • Muhammad Ali beat George Foreman with the Rope-A-Dope, turning Foreman's straight-ahead style against him. By contrast, AMD's long head-on fight with Intel shows that attacking a competent incumbent with Power rarely pays.
  • Start the new model's margin far below the old model's, even below cost. The incumbent sees the price but not the true cannibalization rate. Later, raise prices as the price leader.
  • The strength of the position depends on δ, which links the incumbent's possible gains in the new model to its losses in the old one. If entering costs the incumbent no old volume, δ is 0 and there is no Counter-Positioning. The model is a stylized single-period sketch.
• Counter-Positioning protects only against the incumbent, so it needs a second Power behind it. Ch. 3 · Counter-Positioning
  • In-N-Out holds it against McDonald's but gains nothing against Five Guys.
  • Many challengers can hold it against the same incumbent at once.
  • It differs from Disruptive Technology. In-N-Out counter-positions with no new technology. Kodak faced a disruptive technology without Counter-Positioning. Netflix streaming against HBO via cable is both.
  • Disruptive Technology says nothing about Power, so it says nothing about value.

Own a coveted resource on terms that leave the value with you: Cornered Resource

A coveted asset is Power only when a firm holds it by fiat, on terms that leave most of its value with the firm, and when it passes five strict tests.

• A Cornered Resource is preferential access, on attractive terms, to an asset that raises value on its own. Ch. 6 · Cornered Resource
  • The Benefit can be hit products, a valuable patent, a required input or a cost-saving process.
  • The Barrier is fiat. It comes from personal choice, patent law or property rights, not from ongoing market dealings.
  • Benefit examples range widely: Pixar's hit films, a limestone deposit next to the plant, and a cost-saving process like spin casting.
  • Power intensity can be measured. The strong firm's profit equals the weak firm's price plus the resource's per-unit gain ∆, minus unit cost, times its quantity. Then subtract k, the yearly extra cost of holding the resource.
  • Pay above what a similar replacement would cost counts in k. A resource whose k eats all of ∆ gives no Power.
• Pixar's run of hits came from a cornered group, the Brain Trust. Ch. 6 · Cornered Resource
  • Its first ten films averaged 94% on Rotten Tomatoes.
  • They earned about four times the gross profitability of the average film.
  • The group's trust was forged by surviving the near-failure years together.
  • Toy Story premiered November 22, 1995. It cost $30M, earned over $350M worldwide and enabled Pixar's IPO that month. The first ten films grossed $5.3B. Only Cars scored below 90%, and all but WALL-E beat industry margins.
  • In February 1986 George Lucas, strained by his 1983 divorce, sold the Graphics Group to Steve Jobs for $5M. The company sold hardware to survive and fell below 50 employees by 1990. A 1991 three-picture Disney deal saved it.
  • After Tin Toy won an Academy Award in 1988, Eisner and Katzenberg tried to hire Lasseter back. He stayed: I can go to Disney and be a director, or I can stay here and make history.
  • His commitment later helped persuade Katzenberg to sign the 1991 deal. This is the Barrier working as personal choice.
• Five tests screen out resources that only look like Power. Ch. 6 · Cornered Resource
  • Idiosyncratic: protect the ability behind the assets, not the assets.
  • Non-arbitraged: Brad Pitt's pay captures his value, so he fails.
  • Transferable, ongoing and sufficient: the resource must work elsewhere, still drive returns, and be enough alone.
  • The Brain Trust is well paid but far below its value. Helmer earned a strong return as a Pixar investor.
  • Transferable: Disney paid $7.4B in 2006 because Iger saw no other way to get Pixar's talent. Catmull and Lasseter then revived Disney Animation. Most of Jobs's net worth came from this deal, not Apple.
  • Ongoing: Spenser Silver kept hunting a use for his weak glue. Once Post-it was established, its returns rested on U.S. Patents 3,691,140 and 5,194,299 and the trademark. Jobs was similarly formative at Pixar but not needed later.
  • Sufficient: George Fisher led Motorola well, but Kodak was tied to chemical film in a solid-state age. His success had relied on complements Kodak lacked.
• A people-based resource may not extend to new members, so renewal is the main risk. Ch. 6 · Cornered Resource
  • Brain Trust members directed all of Pixar's first eleven films except one.
  • Newcomer directors were often replaced.
  • Replaced newcomers included Ash and Colin Brady on Toy Story 2, Jan Pinkava on Ratatouille and Brenda Chapman on Brave.
  • Brad Bird made The Incredibles and rescued Ratatouille. He still fits the pattern: he was Lasseter's CalArts classmate and had already directed The Iron Giant.
  • Some credit Lasseter and Catmull alone, citing Disney Animation's revival. The newcomers' failures point to the group's shared history as the true resource.

Some advantages take so long to build that rivals cannot catch up: Branding and Process Power

Branding and Process Power both rest on hysteresis: a rival can match the result only by spending years rebuilding it.

• A brand is Power only when it earns a lasting premium on an offering that is objectively the same. Ch. 5 · Branding
  • Tiffany charged $16,600 for a ring appraised much like a $6,600 Costco ring.
  • Tiffany's Blue Box comes free with a purchase, yet it resells on its own on eBay.
  • In 2005 Good Morning America bought both rings and had gemologist Martin Fuller appraise them. He put the Costco ring at $8,000 plus setting and the Tiffany ring at $10,500 plus setting at a non-brand retailer.
  • Fuller said buyers pay for an earned reputation for quality control: You can go there [and] you don't have to think twice about your purchase.
  • The history runs from 1837: silver awards at the 1867 Paris World's Fair, the Tiffany Diamond in 1878, the six-prong Tiffany Setting in 1886. Its prices run nearly double Blue Nile's for generic online offerings.
  • Unlike Switching Costs, the premium does not require that the buyer owned the product before.
• The premium comes from good feelings and peace of mind, and only time builds either. Ch. 5 · Branding
  • Affective valence: tasters still pay more for Coke after learning Safeway's cola matches it blind.
  • Uncertainty reduction: some buyers pay Bayer a 117% per-tablet premium over Kirkland.
  • Copycats face a long build with no assured payoff, plus trademark risk.
  • On Amazon, 200 Bayer 325 mg tablets cost $9.47 and 500 Kirkland tablets cost $10.93.
  • Only some goods qualify: identity-linked consumer goods and products where failure is costly, such as safety, medicine, food and transport. B2B goods rarely do, because buyers judge objective deliverables.
  • The brand price multiple B(t) starts at 1 and rises along a logistic curve toward a ceiling Z. The leader's margin is 1 minus 1/B(t), and dilution and underinvestment shrink it.
• Brands can be diluted, faked, stranded or fenced in, and big awareness is often scale, not Branding. Ch. 5 · Branding
  • Halston's J.C. Penney deal diluted it. Tiffany sued Costco and eBay over false association and fakes.
  • Nintendo's family brand did not follow gamers into adulthood. Sony's TV premium stopped at Japan's border.
  • Coke can afford Super Bowl ads and RC Cola cannot. That gap is scale.
  • Halston took $1 billion from J.C. Penney for mass-market lines. Bergdorf Goodman dropped the label, the Penney line failed, and Halston never regained its Branding. A down-market push can reset the hysteresis clock.
  • Tiffany told investors in 2013 that it had never let a discount warehouse store like Costco sell its fine jewelry. It said it never would.
  • Branding is non-exclusive. Prada, Louis Vuitton and Hermès chase the same buyers, and all out-earn unbranded rivals.
• Process Power is the rarest Power, because the barrier to copying process gains seldom exists. Ch. 7 · Process Power
  • The barrier has two sources: complexity across a long production chain, and opacity of tacit, uncodified knowledge.
  • Just-in-time, kaizen, kanban and andon cords are only the visible surface of TPS.
  • Toyota itself needed fifteen years to transfer TPS to its suppliers.
  • TPS began after Eiji Toyoda spent three months at Ford's River Rouge plant in 1950. He judged its deep inventories wasteful and took his model from supermarkets that restock only empty shelves.
  • GM ran NUMMI at a Fremont, California plant and sent workers to Japan to train. Van Nuys manager Ernie Shaefer said GM questioned the assembly line and missed the supporting system across the organization.
  • The test: do the gains persist as staff turn over, and can rivals match them only after long effort, whatever they spend? By 2014 Toyota was worth nearly $200B.
• Falling costs from experience are shared by everyone, so they signal arbitrage, not Power. Ch. 7 · Process Power
  • The Experience Curve tracks costs over time, not gaps between rivals at one moment.
  • Each doubling of cumulative units cuts deflated unit cost to 70% to 85% of its earlier level. In a 1990 Science sample, about 60% of 108 examples fell in that range.
  • If the curve created gaps, a firm twice its largest rival's size would commonly hold operating margins 15% to 30% higher. Such gaps are rare.
  • Nelson and Winter's routines, the tacit processes behind cases like TPS, are mostly a Benefit with no Barrier.

All Power starts with invention, which must then be steered toward a Barrier

Power is reached by inventing something new and steering it toward a Barrier, not by planning, pricing for share or copying.

• Getting to Power is a separate problem from holding it, and invention is always the first cause. Part II, Strategy Dynamics
  • Early consultants saw that high share went with high returns, so they bought share with low prices. Rivals fought back until the gains cost more than they were worth.
  • Each Power traces back to something new: a scale model, a developed resource, a superior business model, a complex process, years of consistent creative choices.
  • Planning rarely creates Power. At most it reinforces Power that already exists. Invention runs on action, risk, passion, monomania and domain mastery, which is why founders keep mattering.
  • The static map still guides the journey. A builder must know which destinations pay before judging which routes are worth taking, and only seven destinations do.
  • Some Powers need a prior step. Switching Costs need a customer base first. Cornered Resource usually means developing a resource and then owning it, often through a patent.
• Invention lifts value twice: it opens a path to Power and it grows the market. Part II, Strategy Dynamics
  • Value equals Market Size times Power, and invention moves both factors.
  • Market size comes from compelling value: a product so much better that customers say 'gotta have' and adopt fast.
  • The bar depends on the field. Andy Grove's 10x fits semiconductors. A 50% gain in photovoltaic efficiency or double the battery storage density would likely be enough.
  • The value moment arrives during the high-flux period of invention. Doubts about market size and Power drop sharply, and future cash flow suddenly comes into view.
• Power emerges from crafting under flux, with a constant watch for an opening. Part II, Strategy Dynamics
  • Start from resources, from a person's taste to a firm's data and relationships. Match them to flux in technology, competition or law. Invent by adaptive experiment, not up-front design.
  • Helmer invested in Netflix in 2003 for its DVD-by-mail Power. Streaming showed none yet.
  • The DVD Power was Counter-Positioning against Blockbuster, Process Power, and modest Scale Economies over mail rivals. Blockbuster filed for Chapter 11 on September 23, 2010. Hastings wrote in 2005 that the name was Netflix, not DVD-By-Mail, because internet delivery would come.
  • Netflix tested streaming at 1 to 2% of revenue and signed CBS, Disney, Starz and MTV in 2008 to 2009. Owners charged by usage, so growth built no Power. Exclusives also proved fragile: Epix left for Amazon on Sept 4, 2012.
  • On March 16, 2011 Netflix announced $100 million for 26 episodes of House of Cards, beating HBO, CBS and Showtime. Originals also shifted its bargaining with content owners. By 2015 the share price had risen nearly 100x, to about $50B in market cap.
  • Betting everything on streaming too early would have sunk Netflix. The Rudder Only Works When the Ship Is Moving
• Each path to compelling value carries its own risk and needs its own tactics. Part II, Strategy Dynamics
  • Capabilities-led turns a skill into a product for an unknown need. Go small, stay for years, distrust stated wants.
  • Customer-led targets a known need no one can meet. The risk is technical, so lean on skills rivals lack.
  • Competitor-led must beat a winning product as a whole. Expect big early bets, fast responses, and locking in complements ahead.
  • Adobe Acrobat 1.0 launched June 15, 1993 amid hype and sold barely $2 million in year one. Release 2.0 flopped, and general managers came and went while technical lead Bob Wulff stayed. The Internet created the need, since HTML reflows documents. Sales hit $25 million by end of 1996 and nearly a billion ten years later.
  • Corning was small and new to telecom beside AT&T. Rivals tweaked short-distance glass formulas, but Frank Maurer started from pure silica, which Corning knew best. In September 1970 Don Keck saw laser light pass back and forth through 200 meters of the clearest glass ever made.
  • Sony's Ken Kutaragi bet on real-time 3D graphics. In May 1993 the Executive Committee approved $50 million for the core chip, and 1.3 million chips were ordered. The Japan launch on Dec. 3, 1994 sold 300,000 consoles in a month. By 2000 the base was 90 million, against 18 million N64 and 9 million Sega Saturn.

Each Power can only be first built in its own stage of growth, and a missed window does not reopen

Each Barrier can only be set up during one stage of a business's own growth, so a firm must chase the Powers its stage allows.

• The stages follow the business's own growth, and each Barrier belongs to one stage. Ch. 9 · The Power Progression
  • Origination comes before the compelling value threshold. Takeoff runs above roughly 30 to 40% unit growth a year. Stability follows.
  • These are not product life cycle stages. Origination can come before any sales.
  • Each generic Barrier has its stage. Hysteresis needs time, so it fits stability. Collateral Damage and Fiat fit origination. Cost of Gaining Share is only cheap in takeoff.
  • At the cutoff rate, a market doubles in two years and is fluid enough for leaders to swap places. The right line depends on flux and uncertainty.
  • About seven years of Stanford student research papers back the mapping, with some exceptions. Powers built early carry into later stages. Only first establishment is time-bound.
• Resources and new models must be locked in before takeoff. Ch. 9 · The Power Progression
  • A Cornered Resource must be underpriced, and takeoff makes its value widely known.
  • A Counter-Positioning model is what launches takeoff, so it has to exist first.
  • Intel invented the microprocessor under a contract for Busicom, a Japanese calculator maker, and had to buy back the rights. Sales head Bill Graham fought it and the board worried about cost. Noyce and Arthur Rock prevailed, three years before takeoff.
  • The Noyce, Moore and Grove trio is another plausible Cornered Resource: vision, production problem solving, execution. Rock said Intel needed them in that order.
  • Drug patents secure Power from the start. That is why branded pharmaceuticals pour billions into risky research, and the industry has created hundreds of billions in shareholder value.
• Takeoff is a race for relative scale with one winner. Ch. 9 · The Power Progression
  • High flux slows arbitrage. Rivals are still resolving uncertainty, building capacity and setting up channels, so customers come cheap.
  • Scale needs a share lead by the end of growth. Networks tip, and Switching Costs reward whoever reaches customers first.
  • In late 1978 Intel was losing design wins, and insiders admitted Motorola's 68000 was better. IBM had its own stronger 801 RISC chip. But Don Estridge's Project Chess in Boca Raton was dropping usual practices to build a PC in one year.
  • PC takeoff probably ran from 1975 to 1983. One or two more years without Intel and the chance for Scale Economies would have passed. From March 17, 1980 to the end of 2015, Intel stock rose over 8500% against about 2000% for the S&P 500 TR.
  • Apple greeted the IBM PC with a Wall Street Journal ad: Welcome, IBM. Seriously. Its Apple III shipped May 19, 1980 with boards that shorted. Apple told buyers to drop it three inches to reseat chips.
• Process Power and Branding only form in stability. Ch. 9 · The Power Progression
  • Only a firm that has scaled and run long enough has processes too complex to copy fast.
  • Brand associations take time, so a leap into a new field rarely carries pricing Power at once.
  • Hermès Cognac and Porsche sunglasses failed as extensions. Disney's move into theme parks is the rare exception.
  • Network Economies can still fade later. Software abstracted away from Intel chips, but by then Intel's scale lead was huge.
• Name your stage, then test each candidate Power hard. Ch. 9 · The Power Progression
  • Measure your business's unit growth, not the industry's. Pursue at most three new Powers.
  • Apply the 3 S test: Superior, Significant and Sustainable gains in free cash flow.
  • Strong financials during takeoff are a common false positive. Once growth slows, arbitrage catches up. A well funded entrant is not validation either.
  • Check materiality on its own, since claimed network effects often turn out not to matter. If no route to any of the seven appears, the strategy problem is still unsolved.
  • Good managers rarely rescue a business without Power, yet leaders at Yahoo, Twitter and Zynga were blamed for near impossible situations. Building Power in the first place still depends heavily on leadership.

Put it to work

Stop counting good execution as strategy, and start asking which Barrier your business can build in the stage it is in right now.

1. Test what you have for a Barrier before a Benefit For each business line and each rival, ask what stops a capable competitor from copying your edge. A cost cut or better feature that nothing protects from imitation is not Power. Introduction
   • Include potential and functional rivals, not only direct ones. A single competitor that arbitrages the edge can erase the margin.
   • Analyze each strategically separate business on its own. Intel's memories and microprocessors shared management but needed different answers.
2. Measure your own unit growth and name your stage Origination comes before you cross the compelling value threshold. Takeoff runs while unit growth stays above roughly 30 to 40% a year, and stability follows. Each stage is the only window for first building certain Powers. Ch. 9 · The Power Progression
   • The trap is using the industry's product life cycle instead of your business's growth. The breakpoints differ, so the Powers on offer get misjudged.
3. Invent, then steer the invention toward a Barrier Copying never creates Power, so build a new product, process, model or brand that customers feel they must have. Spend meaningfully but not at bet-the-company scale. Look for moves that make the invention hard to copy. Part II, Strategy Dynamics
   • Netflix's streaming interface, recommendations and cloud setup could all be copied. Exclusives and then originals turned content into a fixed cost, so its scale finally counted.
4. In origination, pick a model the leader cannot copy and lock up key resources Choose a business model that the incumbent could only match by damaging its own profitable core. Target incumbents with high margins. Secure underpriced rights, patents or teams before takeoff makes their value obvious. Ch. 3 · Counter-Positioning
   • Not every incumbent that holds back is counter-positioned. Kodak passed on digital because semiconductor storage was an unattractive business for it on its own.
   • Price the new model low at first, even below cost, so the incumbent sees heavy collateral damage from matching it.
5. In takeoff, race for share and do not stumble Scale Economies, Network Economies and Switching Costs come cheap only while growth is explosive. Win customers and design-ins first, ship quality, and treat a well funded entrant as a rival in a race with one winner. Ch. 9 · The Power Progression
   • Intel's Operation Crush aimed for 2000 design wins in a year. The IBM PC win then carried the company for decades.
   • Measure engagement and churn, not signups. BranchOut peaked near 14 million users, then collapsed when Facebook banned its wall post method.
6. In stability, build brand and process slowly, and guard them Branding and Process Power come only from long runs of reinforcing work that rivals cannot shortcut. Refuse down-market extensions and fight counterfeits and false claims of association. Ch. 5 · Branding
   • Halston took $1 billion from J.C. Penney for mass-market lines. Bergdorf Goodman dropped the label, and Halston never regained its earlier standing.
7. Keep operational excellence, but do not call it strategy Execution still takes most of management's time, and execution errors can erase value even when Power holds. Count a process gain as an edge only if it is complex, tacit and slow to copy. Ch. 7 · Process Power
   • GM's Van Nuys plant manager said GM asked about the assembly line and missed the larger system across Toyota's organization.

The book's words

Power, its Benefit and Barrier, and the stage when each Barrier can form are enough to discuss all of this.

• Power The conditions that let a business keep earning above-normal returns even against capable, committed rivals.
  • Potential value equals market scale times Power. A company's strategy is a route to continuing Power in significant markets.
• Benefit A material boost to cash flow through higher prices, lower costs or lower investment.
  • Benefits are common. Every major cost-cutting effort produces one, so a Benefit alone is not Power.
• Barrier Whatever leaves rivals unable or unwilling to compete the Benefit away.
  • Four kinds map to stages: collateral damage and fiat in origination, cost of gaining share in takeoff, hysteresis in stability.
• Surplus Leader Margin The margin the Power holder earns when prices are set so its rival makes zero profit.
  • It marks leverage, not a predicted price. For Scale Economies it splits into scale economy intensity times scale advantage.
• Counter-Positioning A newcomer's superior model that the incumbent rationally refuses to copy because copying would hurt its core.
  • It comes in three forms that often combine: Milk, History's Slave and Job Security.
• Switching Costs The value a customer expects to lose by moving to another supplier for further purchases.
  • They are financial, procedural or relational. HP's 2004 move to SAP stalled orders and cost $160M.
• Cornered Resource Preferential access on attractive terms to a coveted asset that can raise value on its own.
  • It must be idiosyncratic, non-arbitraged, transferable, ongoing and sufficient. Pixar's Brain Trust passes. A star paid his full value fails.
• Hysteresis The long series of reinforcing steps needed to reach a position, which caps how fast anyone can copy it.
  • It protects Branding and Process Power. Toyota itself took fifteen years to transfer TPS to its suppliers.
• Compelling value An improvement so large that customers feel they have to have it, which drives fast adoption and market size.
  • The bar depends on the industry. Andy Grove's 10x fits semiconductors, while double the battery storage density would likely be enough.
• Takeoff The stage of explosive business growth, above roughly 30 to 40% unit growth a year.
  • PC market takeoff probably ran from 1975 through 1983. Intel broke away from the pack just in time.

Chapter by chapter

Where each idea is in the book.

• Foreword by Reed Hastings Great execution is necessary but not sufficient. A company also needs the right strategy, and its people need a clear, teachable understanding of strategy so they can apply it themselves.
• Introduction A business's potential value equals market scale times Power, the conditions that let it earn persistent above-normal returns despite competition. Strategy should be a simple but not simplistic framework that prepares people to spot and seize Power in rare, decisive moments.
• Part I, Strategy Statics Scale Economies are a source of Power when unit costs fall as a business grows and smaller rivals find that buying share costs more than it returns. Power needs both a scale-intensive industry structure and a real scale lead, and a firm can create either one.
• Ch. 2 · Network Economies Network Economies exist when each customer gets more value as the installed base grows. A leader that passes the tipping point can charge more while a challenger could not match its value even at a loss, but this Power stops at the edges of the network.
• Ch. 3 · Counter-Positioning A newcomer can beat a powerful incumbent by adopting a superior business model that the incumbent will not copy, because copying it would damage the incumbent's existing business. The incumbent's own strength becomes the barrier that protects the challenger.
• Ch. 4 · Switching Costs Switching Costs give a firm Power when customers would lose value by moving to another supplier for later purchases. The holder can charge more than rivals on follow-on sales to customers it already has, and rivals must pay customers to cover the cost of leaving.
• Ch. 5 · Branding Branding is a Power when a firm can durably charge more for an objectively identical offering because of its history with customers. The barrier is the long, uncertain time it takes to build that history, so the brand must be guarded from dilution, counterfeits and misuse.
• Ch. 6 · Cornered Resource A Cornered Resource is preferential access, on attractive terms, to a coveted asset that can raise value on its own. It counts as Power only if it passes five tests: idiosyncratic, non-arbitraged, transferable, ongoing and sufficient.
• Ch. 7 · Process Power Process Power comes from improvements in cost or product quality that are embedded in a company's organization and can only be matched after a long, sustained effort. It is rare because the Benefit (operational excellence) is common, while the Barrier (hysteresis from complexity and opacity) is seldom present.
• Part II, Strategy Dynamics Every type of Power begins with invention, whether of a product, process, business model or brand, so copying others never creates Power. Invention also creates market size through compelling value, but it must be steered toward one of the seven Powers or it ends up as an easily copied commodity.
• Ch. 9 · The Power Progression Each type of Power can only be first established in a specific window of a business's growth: Counter-Positioning and Cornered Resource in origination, Scale Economies, Network Economies and Switching Costs in takeoff, Process Power and Branding in stability. If a window passes unused, that Power is gone for good.