Strategy Book Summaries

Key Concepts & Applications for AI Transformation Made Simple

1. Good Strategy Bad Strategy — Richard Rumelt (2011)

2. The Crux — Richard Rumelt (2022)

3. The Balanced Scorecard — Robert Kaplan & David Norton (1996)

4. Strategy Maps — Robert Kaplan & David Norton (2004)

5. The Three Horizons of Growth — Mehrdad Baghai, Steve Coley, David White (McKinsey, 1999/2000)

6. Strategy Beyond the Hockey Stick — Chris Bradley, Martin Hirt, Sven Smit (McKinsey, 2018)

7. Play Bigger — Al Ramadan, Dave Peterson, Christopher Lochhead, Kevin Maney (2016)

8. High Output Management — Andy Grove (1983/1995)

Good Strategy Bad Strategy

Richard Rumelt (2011)

Core Argument

Richard Rumelt's seminal work cuts through the noise of corporate strategy to reveal a fundamental truth: most of what passes for strategy in organizations is fundamentally flawed. Companies confuse vague aspirations with strategic direction, mistake financial goals for strategic objectives, and fail to honestly confront the core challenges they face. Rumelt identifies four hallmarks of bad strategy: fluff (statements that sound strategic but are empty), no diagnosis (failing to clearly identify the problem), treating goals as strategy (confusing objectives with the means to achieve them), and lack of coherent action (initiatives that don't reinforce each other).

Good strategy, by contrast, has a clear structure that Rumelt calls 'the kernel.' This kernel consists of three integrated elements: a diagnosis that identifies the most important challenge or obstacle, a guiding policy that establishes the overall approach to addressing this challenge, and coherent actions—specific steps and initiatives that directly support the guiding policy. Rumelt argues that strategy works through focus and leverage, not through comprehensive planning. The best strategists ruthlessly prioritize, identifying the point of greatest leverage and concentrating resources there. The book provides numerous case studies showing how companies like Apple, Southwest Airlines, and Starbucks succeeded by having clear kernels, and how others failed by lacking them.

The power of Rumelt's framework lies in its simplicity and universality. Whether a company is entering a new market, facing disruption, or trying to improve operational performance, the kernel structure provides a guide. Strategy is not about doing more—it's about doing the right thing, the right way, in the right order. This insight has become foundational to strategic thinking across industries.

Key Concepts

The Kernel

The core of good strategy consisting of three parts: (1) a diagnosis that identifies the critical challenge, (2) a guiding policy that determines the overall approach, and (3) coherent actions that specifically implement the policy. The kernel is not a list of goals or a vague mission—it is a focused statement of what matters most and how you will address it.

Bad Strategy's Four Hallmarks

Fluff (empty, meaningless language that sounds strategic), no diagnosis (failing to identify the actual problem), mistaking goals for strategy (confusing aspirations with methods), and bad execution (actions that don't reinforce each other). Rumelt shows that bad strategy is pervasive because it's easier than honest diagnosis and difficult choices.

Sources of Power

Rumelt identifies how strategy creates advantage: leverage (identifying a small set of actions that have outsized impact), proximate objectives (choosing goals that are achievable and create momentum for further progress), chain-link systems (where each element depends on the others working correctly), and dynamics (exploiting changes in the environment before competitors do).

Saying No

Strategy is as much about what you don't do as what you do. Rumelt emphasizes that focus requires choosing not to pursue attractive opportunities that don't align with the strategic kernel. This discipline of prioritization is what separates strategic from non-strategic organizations.

Proximate Objectives

Achievable goals that represent clear progress toward the ultimate vision, building momentum and providing stepping stones. Rather than trying to leap directly to a distant aspiration, proximate objectives create a realistic path forward while validating assumptions and building organizational capability.

Application to AI Transformation

Good Strategy Bad Strategy maps directly onto the framework in 'AI Transformation Made Simple.' The book's 5-step methodology—Diagnose, Envision, Prioritize, Execute, Measure—is fundamentally built on Rumelt's kernel structure. Step 1 (Diagnose) is about identifying the critical challenge your organization faces and the leverage points where AI can make a difference. Many organizations jump to AI initiatives without this diagnosis, creating what Rumelt would identify as bad strategy: vague aspirations ('we need to be AI-first'), confused goals ('our target is 15% cost reduction through automation'), and incoherent actions (implementing chatbots in customer service, RPA in finance, and generative AI in marketing with no connection between them).

The concept of 'sources of power' is particularly valuable for AI strategy. AI creates leverage when it multiplies the output of existing teams (a data analyst supported by AI tools can accomplish more), enables proximate objectives (starting with narrow use cases that build momentum), and functions as part of chain-link systems (where data quality, model accuracy, and organizational process redesign all depend on each other). The book emphasizes that AI transformation requires coherent action—initiatives that reinforce each other and support a clear guiding policy. For example, if your guiding policy is 'AI-enabled faster decision-making,' then your AI investments should target decision points, your data investments should ensure decision-relevant data quality, and your organizational changes should establish clear decision authorities and escalation paths.

Perhaps most importantly, Rumelt's emphasis on saying 'no' applies directly to AI strategy. Organizations often attempt too many AI initiatives simultaneously—the 'spray and pray' approach. This violates Rumelt's principle of focus and leverage. Effective AI transformation requires choosing a proximate objective (what can we achieve with AI in the next 6-12 months that builds organizational capability?), pursuing it with coherent action, and only then expanding to adjacent opportunities. The framework in 'AI Transformation Made Simple' incorporates this discipline through the Prioritize step, which explicitly forces choices about which initiatives align with the strategic kernel and which should be deprioritized or killed.

Key Ideas to Reference

The diagnosis should identify the most important challenge facing the organization and explain why that challenge is critical. Without clear diagnosis, the rest of the strategy is built on sand.

Strategy is not primarily about achieving financial targets—it is about identifying the leverage point where the organization's efforts will have the most impact relative to the effort expended.

Proximate objectives are not the ultimate goal but achievable targets that build momentum and capability. They demonstrate progress while validating assumptions about what's possible.

The coherence of strategy depends on whether the guiding policy and actions reinforce each other. If they don't, you don't have strategy—you have a collection of initiatives.

The Crux

Richard Rumelt (2022)

Core Argument

In this more recent work, Rumelt returns to the core insight that effective strategy requires identifying the crux—the critical challenge that, once solved, unlocks everything else. The crux is different from the long list of problems most organizations face; it is the one problem whose solution creates cascading progress on other fronts. Rumelt argues that most organizations are unable to identify their crux because they lack the intellectual honesty to distinguish between real problems and symptoms, and because identifying a specific crux requires the courage to say 'this matters more than everything else.'

The book distinguishes between 'gnarly' challenges (complex, interconnected, difficult to solve) and 'simple' ones (with clear cause and effect). Most strategic cruxes are gnarly, which is why leaders often avoid identifying them—gnarly problems don't have clean solutions, and admitting you're facing one feels uncomfortable. However, Rumelt shows that once a leader can articulate the crux with specificity and clarity, the path forward often becomes obvious. The crux then becomes the organizing principle for strategy: decisions are evaluated based on whether they address the crux, resources are allocated to solving it, and success is measured by progress on it.

Rumelt emphasizes that identifying the crux requires judgment, insight, and sometimes analogy—looking at how other organizations solved similar problems. The process is partly analytical (gathering data, testing hypotheses) and partly intuitive (pattern recognition, informed judgment). The book provides frameworks for structured crux identification, including adversarial collaboration (where different parts of an organization argue for different diagnoses) and pre-mortem analysis (imagining that the strategy failed and working backward to understand why). Throughout, Rumelt insists that strategy is not about being smart in the abstract—it's about the practical ability to identify what matters most and mobilize an organization around it.

Key Concepts

The Crux

The single most critical challenge whose resolution unlocks progress on multiple other fronts. The crux is not simply the biggest problem or the most complex—it is the one whose solution has cascading benefits. Identifying it requires brutal honesty about what is actually holding the organization back versus what leaders wish were the constraint.

Gnarly vs. Simple Challenges

Gnarly challenges involve multiple interconnected variables, unclear cause-and-effect relationships, and no obvious solution path. Simple challenges have clear causes and solutions. Most strategic cruxes are gnarly, which is why leaders often misidentify them (preferring to focus on simpler, more solvable problems) or avoid identifying a crux altogether.

Judgment and Insight in Strategy

Strategy requires judgment—the ability to evaluate incomplete information, recognize patterns, and make informed decisions about what matters most. This cannot be reduced to a formula or algorithm. Judgment develops through experience, study of history and analogies, and structured reflection on past decisions.

Analogy and Pattern Recognition

Much strategic insight comes from recognizing patterns across different contexts. If another organization faced a similar crux, how did they identify it? What was their diagnosis? How did they execute? Analogical reasoning accelerates strategic thinking and helps leaders avoid reinventing solutions that have already been discovered.

Action-Oriented Strategy

The crux is not a theoretical exercise—it is a practical diagnosis that leads directly to action. Once identified, the crux becomes the criterion for resource allocation, priority setting, and measurement. Rumelt emphasizes that identifying a crux without committing to address it is itself a form of bad strategy.

Application to AI Transformation

The Crux is the intellectual foundation for Chapter 3 of 'AI Transformation Made Simple,' titled 'Finding the Crux.' In AI transformation, organizations often face multiple problems that could potentially be solved with AI: customer churn, operational inefficiency, slow decision-making, innovation velocity, talent retention, regulatory compliance, and more. The crux framework forces leadership to ask: Which ONE of these problems, if solved, would have the most cascading impact on organizational performance and capability? An organization might identify that slow decision-making is the crux because it undermines all strategic initiatives—if leadership could make decisions faster and more accurately, it would accelerate innovation, improve customer responsiveness, reduce time-to-market, and enable the organization to adapt more quickly to market changes.

Rumelt's distinction between gnarly and simple problems is essential for realistic AI strategy. Many organizations approach AI as if it's a simple problem with a straightforward solution: 'Deploy a language model, get productivity gains.' In reality, most meaningful AI transformations address gnarly problems. If the crux is slow decision-making, the problem is gnarly because it involves data architecture, organizational process, decision rights, capability gaps, and cultural readiness—not just technology. A realistic AI strategy acknowledges this gnarl and designs a multi-faceted approach. The framework in the book explicitly addresses this by connecting the crux to multiple levers: technology (what AI capabilities), process (how decisions will be made), organization (who decides), and culture (what values guide decision-making).

The emphasis on judgment and insight is particularly important for AI strategy because many leaders try to mechanize strategic decisions about AI ('Let's digitize everything,' 'Deploy AI wherever it reduces cost'). Rumelt's framework insists that identifying the crux requires leadership judgment informed by deep knowledge of the business. This means that the Diagnose step in the book should not be delegated to data scientists or IT leaders alone—it requires business leaders who understand the strategic context. Similarly, recognizing analogies (how did other companies in our industry solve this? how did adjacent industries address similar cruxes?) accelerates strategic thinking and reduces the risk of pursuing solutions that won't work in your context.

Key Ideas to Reference

The crux is the one challenge whose resolution makes everything else more possible. Identifying it requires intellectual honesty and the courage to say this matters more than everything else.

Most organizations face gnarly problems but try to solve them with simple solutions. The path forward requires acknowledging the full complexity of the challenge.

Judgment in strategy is about recognizing patterns and applying relevant analogies from different contexts. It cannot be replaced by data alone, though data informs it.

Once you identify the crux, it becomes the organizing principle for all strategic decisions. Every resource, initiative, and change is evaluated based on whether it addresses the crux.

The Balanced Scorecard

Robert Kaplan & David Norton (1996)

Core Argument

Kaplan and Norton's groundbreaking framework addresses a fundamental problem in organizational management: financial metrics alone are insufficient for managing modern, knowledge-intensive organizations. Traditional management focuses almost exclusively on financial results (revenue, profit, ROI), which are lagging indicators—they reflect past performance but provide limited guidance for future strategy. By the time financial results are poor, it's often too late to correct course. The Balanced Scorecard adds three additional perspectives to financial metrics, creating a comprehensive system for strategy management and performance measurement.

The four perspectives of the Balanced Scorecard are (1) Financial—addressing shareholder value and profitability, (2) Customer—measuring customer satisfaction, retention, and value, (3) Internal Business Processes—tracking the operational excellence and process improvements that drive customer and financial results, and (4) Learning and Growth—assessing employee satisfaction, capability development, and innovation. Critically, these perspectives are not independent silos but are connected through cause-and-effect relationships: investments in learning and growth enable improvements in internal processes, which deliver better customer outcomes, which ultimately drive financial results. This cause-and-effect logic is visualized through strategy maps.

The Balanced Scorecard has become one of the most widely adopted management frameworks globally because it solves a practical problem: how do you operationalize strategy? How do you translate a strategic direction into specific metrics and management actions? The scorecard connects strategy to execution by identifying leading indicators (lagged performance that tells you if you're on track) in each perspective, cascading metrics throughout the organization, and creating a management system focused on strategic progress rather than just financial results. For knowledge-intensive and technology-driven organizations, this shift away from pure financial focus is essential because much of the value creation happens in learning, innovation, and capability development—areas that don't immediately show up on financial statements.

Key Concepts

Four Perspectives

Financial (profitability, revenue growth, shareholder value), Customer (satisfaction, retention, market share, value proposition), Internal Business Processes (operational efficiency, quality, innovation, safety), and Learning & Growth (employee capability, organizational culture, technology infrastructure). Together, these perspectives provide a balanced view of organizational health and strategic progress.

Leading vs. Lagging Indicators

Lagging indicators (like revenue or profit) measure historical results. Leading indicators (like employee satisfaction, process defect rates, or customer retention trends) predict future performance. A balanced scorecard combines both, allowing organizations to identify problems before they show up in financial results and to course-correct earlier in the strategy cycle.

Strategy Maps

Visual representations of the cause-and-effect logic that connects strategic initiatives across the four perspectives. Strategy maps show how investments in learning and growth enable process improvements, which drive customer value, which creates financial returns. They make strategy visible and help align the organization around common logic.

Cause-and-Effect Relationships

The hypothesis that improvement in one perspective creates improvement in others. For example, 'Improved employee training (Learning & Growth) leads to fewer errors in transaction processing (Internal Processes) leads to higher customer satisfaction (Customer) leads to increased revenue (Financial).' These relationships should be explicit and testable.

Cascading Scorecards

The scorecard structure is cascaded throughout the organization so that business unit scorecards align with corporate strategy, team scorecards align with business unit strategy, and individual goals align with team objectives. This creates alignment where every part of the organization understands how it contributes to overall strategy.

Application to AI Transformation

The Balanced Scorecard framework directly informs Chapter 11 of 'AI Transformation Made Simple,' 'Measuring What Matters.' Many organizations measure AI success using only financial metrics: cost reduction, revenue increase, or ROI. While these financial measures matter, they're lagging indicators that don't emerge until months or even years after AI implementation. A balanced approach to AI measurement includes: (1) Financial metrics (cost reduction from automation, revenue from new AI-enabled products, margin improvement), (2) Customer metrics (customer satisfaction with AI-assisted services, customer retention, net promoter score impact), (3) Internal Process metrics (cycle time reduction, decision accuracy, data quality, model performance metrics), and (4) Learning & Growth metrics (employee adoption rates, employee satisfaction with AI tools, capability development in AI literacy, cultural alignment with AI).

The cause-and-effect logic is particularly important for AI transformations. Organizations often invest heavily in AI (Learning & Growth and Internal Processes perspectives) expecting financial returns, but if they don't establish the intermediate cause-and-effect connections, they fail to see progress. For example: 'Investment in AI training for business teams (Learning & Growth)' should lead to 'Better prompt engineering and tool usage in daily work (Internal Processes)' which leads to 'Faster customer response time and fewer errors (Customer)' which drives 'Higher customer retention and increased revenue (Financial).' If any link in this chain breaks (e.g., employees are trained but don't adopt the tools, or they adopt tools but processes aren't redesigned to take advantage of them), the financial benefits won't materialize. The balanced scorecard helps diagnose where the chain is breaking.

A balanced AI measurement approach also addresses a key insight from Chapter 11: organizations should measure the 'shadow' outcomes of AI, not just the primary ones. If you deploy AI to accelerate decision-making, measure decision speed (lagging indicator) but also intermediate metrics like AI model performance, data quality, decision quality, and organizational adoption. This gives you real-time insight into whether the AI transformation is on track and where intervention might be needed. For example, if adoption rates are low even though the AI system is working correctly, the problem isn't the technology—it's the organizational or process dimension. The balanced scorecard's multi-perspective approach surfaces these insights.

Key Ideas to Reference

Financial metrics are lagging indicators—they tell you about past performance but provide limited guidance for future strategy. A balanced view includes leading indicators that predict future results.

Cause-and-effect relationships make strategy visible and testable. If you invest in learning but don't see process improvement, or invest in processes but don't see customer benefit, the chain is broken somewhere.

The scorecard translates strategy into actionable metrics and connects the organization around a common logic. It answers the question: how do we know if we're executing our strategy?

Measurement systems drive behavior. If you only measure financial results, organizations will pursue short-term financial gains. If you measure a balanced set of indicators, organizations will pursue sustainable, strategic value creation.

Strategy Maps

Robert Kaplan & David Norton (2004)

Core Argument

Strategy Maps extends the Balanced Scorecard framework by making the cause-and-effect relationships explicit and visual. Kaplan and Norton argue that the most critical asset in modern organizations is intangible: human capital (knowledge, skills, culture), information capital (data, systems, networks), and organizational capital (processes, governance, alignment). These intangible assets are largely invisible on traditional balance sheets, yet they are what enable competitive advantage. Strategy Maps visualize how these intangible assets create value through operational processes that deliver customer outcomes and ultimately financial results.

A strategy map is structured from bottom to top: it begins with learning and growth investments (building human, information, and organizational capital), shows how these capabilities enable improvements in internal processes (quality, efficiency, safety, innovation), which deliver customer value (meeting customer needs, building loyalty, creating solutions), which ultimately drives financial performance (revenue growth, profitability, enterprise value). The critical insight is that financial success is not independent—it flows from customer success, which flows from operational excellence, which flows from organizational capability. This structure is the opposite of traditional financial planning, which starts with a financial target and works backward to projects that might achieve it.

Strategy Maps have become essential in technology-intensive and innovation-driven organizations because they clarify the value creation logic. In traditional manufacturing, the cause-and-effect chain is straightforward: efficient operations lead to lower costs, which attract more customers, which increase revenue. In technology and AI-driven organizations, the logic is more subtle. For example, in an AI transformation, improving data quality (Learning & Growth) enables more accurate AI models (Internal Processes), which provides more reliable decision support (Customer), which drives better business outcomes (Financial). Without a clear strategy map showing this logic, organizations may invest in data quality without realizing it's foundational, or they may implement AI without ensuring their data is adequate. Strategy Maps make these dependencies visible.

Key Concepts

Intangible Assets

Human Capital (employee skills, knowledge, motivation, alignment), Information Capital (data quality, system architecture, intellectual property, networks), and Organizational Capital (culture, governance, processes, incentives). These assets are largely invisible but are the primary sources of value creation in modern organizations. They cannot be easily replicated and form the basis of sustainable competitive advantage.

Value Creation Process

The pathway from capability building to financial results: Develop and maintain intangible assets → improve internal processes → deliver customer value → achieve financial results. Each level depends on the levels below it. This structure clarifies what investments are necessary and in what sequence.

Strategic Themes

Clusters of related strategic initiatives that collectively address a strategic objective. For example, a 'Customer Intimacy' theme might include CRM system implementation, customer data platform development, customer-facing training, and incentive alignment. Strategic themes help organize multiple initiatives around a common strategic purpose.

Alignment of Assets to Strategy

The most important strategic asset is not the one you're investing the most in—it's the one that most directly enables your strategic objective. Strategy Maps force alignment between asset development and strategic priorities. If your strategy is innovation, you should be building information capital (data, systems) and organizational capital (collaboration culture, incentive alignment) more than optimizing existing operations.

The Perspective Structure

Learning & Growth → Internal Processes → Customer → Financial is not a generic structure but a strategic logic. Different industries and business models may need different structures, but the fundamental principle is that strategy is about building capabilities that enable processes that deliver customer value that drives financial results.

Application to AI Transformation

Strategy Maps are invaluable for communicating AI strategy to boards and leadership teams because they visualize the value creation logic. Many organizations struggle to explain to investors or boards why they should invest in AI—the connection between an AI project and financial results is often unclear. A strategy map makes it explicit. For example: 'Investment in machine learning engineering talent and data infrastructure (Learning & Growth) enables development of more accurate customer propensity models (Internal Processes), which enables personalized customer engagement (Customer), which increases customer lifetime value (Financial).' This chain is clear, testable, and sequenced. It clarifies why you're building capability even if the financial benefit isn't immediate.

The concept of intangible assets is particularly important for AI strategy. AI's value is almost entirely dependent on intangible assets: data (information capital), machine learning expertise (human capital), decision-making processes that can leverage AI predictions (organizational capital), and culture that embraces AI-assisted decision-making. Organizations often invest in AI technology (hiring ML engineers, buying platforms) but neglect the intangible assets that make AI valuable. A strategy map surfaces this. If your strategy map shows that Financial results depend on Customer outcomes, which depend on Internal Processes (decision-making), which depend on Information Capital (data quality) and Organizational Capital (process redesign), then your investment priorities become clear. You can't just buy AI technology—you must simultaneously invest in the data infrastructure, the process redesign, the talent development, and the culture change that make AI valuable.

Strategy Maps also help address a common AI implementation problem: strategic misalignment. Different parts of the organization may be investing in AI without a common logic. One group is building NLP models for customer service, another is implementing RPA for finance, another is exploring predictive analytics for supply chain. Without a common strategy map, these initiatives don't reinforce each other. With a map, you can ask: do these AI initiatives collectively build the intangible assets we need? Do they support the same customer value proposition? Do they enable the same internal processes? This alignment dramatically increases the probability that AI investments will compound into organizational capability rather than remaining isolated experiments.

Key Quotes & Ideas to Reference

Intangible assets—human capital, information capital, organizational capital—are the primary sources of competitive advantage in modern organizations. Yet they rarely appear on balance sheets, making them invisible to many leaders.

The value creation process moves from capability building through process improvement to customer value to financial results. If you invest in capabilities but don't redesign processes, you won't see customer benefits. If you improve processes but don't explain customer value clearly, adoption will be weak.

A strategy map makes strategy visible and executable. When everyone can see the cause-and-effect chain, alignment becomes possible.

The most powerful strategic investments are often those that develop intangible assets because they enable multiple initiatives. A single investment in data infrastructure can enable dozens of AI applications.

The Three Horizons of Growth

Mehrdad Baghai, Steve Coley, David White (McKinsey, 1999/2000)

Core Argument

This McKinsey framework, developed through research on growth patterns across multiple industries, identifies a fundamental problem in corporate portfolio management: organizations systematically over-invest in their core business (Horizon 1) while under-investing in emerging businesses (Horizon 2) and future growth options (Horizon 3). The result is a declining growth trajectory as mature businesses age and no new sources of revenue develop. The Three Horizons framework provides a structure for balancing the portfolio across different time horizons and growth mechanisms.

Horizon 1 consists of the core, mature businesses that generate current cash flow and profit. These businesses should be defended and extended through incremental innovation, efficiency gains, and market expansion within the existing business model. Horizon 2 comprises emerging businesses that are outside the core but represent significant growth opportunities over a 5-7 year horizon. These require investment in new business models, new customer segments, or new value propositions. Horizon 3 consists of exploratory investments in completely new businesses or entirely new market categories that may not generate meaningful revenue for 7-10+ years but position the organization for long-term growth. Most organizations spend 80-90% of resources on H1, 10-15% on H2, and nearly 0% on H3, creating a growth crisis.

The framework addresses a key strategic tension: businesses must generate cash today while building for tomorrow. Horizon 1 must be profitable and cash-generative. Horizon 2 investments should begin to move toward profitability but may still be cash-consumptive. Horizon 3 is purely exploratory and should be resourced as a small percentage of overall budget. The McKinsey research shows that companies that maintain a balanced 'growth staircase'—where H1 businesses are declining but H2 and H3 investments are scaling—maintain stronger overall growth trajectories than companies that concentrate exclusively on the core. The allocation challenge is not trivial: it requires discipline to invest in businesses that won't generate returns in the near term while defending cash-generative core businesses.

Key Concepts

Horizon 1: Defend and Extend

The core business that generates current revenue and profit. H1 strategies focus on market share within the existing category, operational efficiency, incremental innovation, and defending against competitors. Resource allocation: typically 70-80% of budget. Investment type: cash-generative.

Horizon 2: Build Emerging Businesses

Businesses outside the core that represent significant growth opportunity over 5-7 years. H2 might involve new customer segments, adjacent markets, new business models applied to existing capabilities, or new products for existing customers. Resource allocation: typically 15-25% of budget. Investment type: growth-focused, moving toward profitability.

Horizon 3: Create Options

Exploratory investments in completely new businesses, new market categories, or new technologies with uncertain payoff but potentially transformational upside. H3 investments should be small relative to overall portfolio but enable the organization to compete in future markets. Resource allocation: typically 5-15% of budget. Investment type: exploratory, high-risk/high-reward.

The Growth Staircase

A portfolio structure where declining revenue from mature H1 businesses is offset by emerging revenue from H2 and growing revenue from H3 options. This creates a staircase pattern of growth. Organizations without a growth staircase face a cliff as core businesses mature and no new sources of revenue develop.

Innovation Pipeline

The portfolio of initiatives across three horizons designed to maintain overall growth. Managing the pipeline requires clear criteria for which ideas move between horizons (when does an H3 experiment become an H2 bet? when does an H2 business move to H1 scale?), resource allocation processes, and governance that allows different performance metrics for each horizon.

Application to AI Transformation

The Three Horizons framework is directly adapted in Chapter 6 of 'AI Transformation Made Simple,' 'Proximate Objectives,' where the book structures AI investments across three horizons. Horizon 1 AI investments are quick wins within the existing business model—automation of repetitive tasks, basic chatbots for customer service, predictive maintenance in manufacturing. These generate near-term value and build organizational AI literacy. Horizon 2 AI investments are more significant: redesigning core business processes with AI (e.g., restructuring sales processes around AI-assisted lead scoring and prioritization), creating new revenue streams (e.g., AI-powered advisory services), or entering adjacent markets with AI capabilities. H2 typically requires 1-2 year development and significant organizational change. Horizon 3 AI investments are exploratory: could AI enable a fundamentally new business model? Could AI create an entirely new category in which we compete? These might involve emerging technologies like generative AI applications not yet proven in your industry, or business model experimentation that could take 3-5 years to mature.

The McKinsey research showing that organizations under-invest in H2 and H3 has profound implications for AI strategy. Many organizations attempt an 'all in' approach: they declare that everything they do will be 'AI-enabled' or they try to apply AI everywhere simultaneously. This is often just a repackaging of H1—they use AI to optimize the core business but don't make the harder strategic choices about what new capabilities, processes, and businesses AI enables. A disciplined approach allocates 50-60% of AI resources to H1 (quick wins and capability building), 30% to H2 (significant business process transformation and adjacent revenue), and 15-20% to H3 (exploratory bets on new business models or categories). This allocation forces choices: which quick wins are worth doing? which business processes are worth the disruption and cost of AI-enabled redesign? which new opportunities should we explore?

The framework also helps organizations sequence AI investments correctly. A common mistake is to attempt H2 and H3 investments before establishing H1 foundations. This is strategically unsound for two reasons: (1) H1 wins build organizational capability, data quality, and cultural readiness that are prerequisites for H2 and H3, and (2) H1 wins generate the cash that funds H2 and H3. A better sequencing invests in 2-3 carefully chosen H1 wins in the first 6-12 months, using that period to build technical capability and organizational understanding. Then, with confidence in H1 execution, the organization pursues H2 initiatives that require more extensive process redesign. H3 exploration is ongoing but modest in resource allocation. As H1 businesses mature and AI becomes more embedded, the organization should naturally shift resources toward H2 and H3. This sequencing is captured in Chapter 6's discussion of the 'AI Maturity Journey.'

Key Ideas to Reference

Organizations systematically over-invest in defending the core and under-invest in building emerging businesses. This creates a growth crisis when the core matures and no new revenue sources exist.

The growth staircase is not a nice-to-have—it is essential for long-term organizational health. Companies that maintain a disciplined portfolio across horizons outgrow competitors who concentrate entirely on the core.

Different horizons require different governance, metrics, and resource allocation approaches. Applying H1 metrics (immediate profitability) to H2 and H3 investments kills promising new businesses before they have time to mature.

The challenge is not identifying good ideas in H2 and H3—most organizations have plenty of them. The challenge is having the discipline to resource them adequately while still protecting H1 cash generation.

Strategy Beyond the Hockey Stick

Chris Bradley, Martin Hirt, Sven Smit (McKinsey, 2018)

Core Argument

Most strategic plans follow a predictable pattern: modest growth in the first few years followed by a dramatic uptick later—the 'hockey stick' shape. McKinsey research across thousands of companies shows that these projections almost never materialize. The hockey stick trap reflects three problems: (1) leaders anchor to historical growth rates and project them forward, (2) they assume success without specifying what must change to achieve it, and (3) they fail to make the big, difficult moves necessary to fundamentally change trajectory. The result is chronic strategic disappointment and organizations that consistently fall short of their own targets.

Bradley, Hirt, and Smit argue that meaningful change in company performance requires 'big moves'—significant shifts in resources, capabilities, market positioning, or business model. These might include: programmatic M&A (not just one or two deals but a sustained acquisition strategy), significant resource reallocation (not marginal budget shifts), major capital expenditure (not incremental capex), productivity improvements (meaningful headcount or cost reduction), or differentiation improvements (substantial changes to products, services, or customer experience). The book identifies 10 performance levers across these categories, each with evidence of what kinds of shifts can meaningfully change trajectory. Small, incremental adjustments—even well-executed ones—rarely move the needle on overall company performance. The power law of company performance means that a few big moves create most of the value.

The book also emphasizes the social side of strategy—why organizations fail to make the big moves they need to make. Psychological biases (loss aversion, status quo bias), political dynamics (various parts of the organization protecting their turf), and inertia (the difficulty of changing deeply embedded processes and culture) all work against big moves. The book includes sections on building coalitions, overcoming resistance, and managing the transition to new ways of working. This acknowledgment that strategy is as much about organizational dynamics as analytical insight is particularly valuable and practical.

Key Concepts

The Hockey Stick Trap

The tendency to project optimistic growth trajectories that assume future success without specifying what must change to achieve it. Most hockey stick plans fail because leaders anchor to historical performance and underestimate the change required. Realistic strategic planning requires honest assessment of what must change and how much effort that change will require.

The Social Side of Strategy

Strategy fails not because it's analytically flawed but because of organizational dynamics: resistance to change, political conflicts, inertia, and cognitive biases. Effective strategy implementation requires building coalitions, creating urgency, securing leadership alignment, and managing the human and organizational dimensions of change.

The 10 Performance Levers

Specific areas where companies can drive meaningful performance change: (1) Explore and shape markets, (2) Acquire and merge, (3) Divest, (4) Make strategic capital investments, (5) Improve products and services, (6) Reduce costs systematically, (7) Deploy digital and technology, (8) Reshape the portfolio, (9) Improve the talent base, (10) Improve the operating model. Effective strategies typically involve simultaneous moves across multiple levers.

Big Moves

Significant, difficult decisions that require meaningful resource commitment and organizational change. Examples include entering a new market via acquisition, exiting a business line, implementing a new technology platform company-wide, or fundamentally restructuring the organization. Small, incremental moves rarely move the needle on overall company performance.

The Power Law of Company Performance

A small number of big moves create most of the value. Rather than optimizing dozens of small initiatives, effective strategy identifies a few critical moves that will fundamentally change the company's trajectory and concentrates effort there.

Application to AI Transformation

Strategy Beyond the Hockey Stick provides a crucial reality check for AI strategy: many organizations make hockey stick plans for AI ('By year 3, AI will reduce costs by 20%, increase revenue by 15%, and enable entry into three new markets') without specifying the big moves required to achieve these ambitious targets. The book forces organizations to be honest: if you want AI to fundamentally change your company's trajectory, what big moves will you need to make? This might involve significant investments in data infrastructure, restructuring of business processes, acquisition of AI talent or companies, reorganization of decision-making authority, or even business model transformation. These are not small, incremental changes—they are big moves that require sustained commitment, organizational disruption, and resource reallocation.

The concept of the 10 performance levers is particularly relevant to AI strategy because AI is not a single lever—it intersects with multiple levers. Deploying digital and technology (lever 7) is the obvious AI application, but AI also enables: market exploration (lever 1, through better customer insights), acquisitions (lever 2, identifying targets or integrating acquired AI capability), divestiture (lever 3, understanding which businesses to exit), capital investment decisions (lever 4, more data-driven), product improvement (lever 5, through AI-assisted design and personalization), cost reduction (lever 6, through automation), reshaping the portfolio (lever 8, understanding which business mix creates value), talent improvement (lever 9, through better hiring and development), and operating model improvement (lever 10, through process redesign). Organizations should not treat AI as just a technology lever but should ask how AI enables movement across multiple levers. For example, AI-enabled market intelligence (lever 1) should inform portfolio decisions (lever 8), which should inform talent strategy (lever 9). This integrated approach is more likely to create the 'big moves' that meaningfully shift company trajectory.

The book's emphasis on the social side of strategy is especially important for AI because many AI strategies fail due to organizational resistance, not technical problems. Employees resist AI-enabled process changes because they fear job loss or disruption. Business leaders resist shifting resources from existing programs to AI because existing programs have stakeholders and political support. The book's framework for building coalitions, managing resistance, and securing leadership alignment is therefore essential. An AI transformation is not just a technical and strategic change—it's an organizational and cultural change. The organizations that succeed at AI transformation are those that explicitly manage these social and organizational dimensions, not just the technical ones. This is captured in the book's emphasis on chapters on organization and culture change alongside the technical framework.

Key Quotes & Ideas to Reference

Most strategic plans follow a hockey stick pattern that never materializes. Realistic plans acknowledge what must actually change and the difficulty of changing it.

Small, incremental improvements rarely move the needle on overall company performance. Strategy requires identifying the few big moves that will fundamentally shift trajectory.

The social side of strategy—building coalitions, managing resistance, securing alignment—is as important as the analytical side. Many strategies fail due to organizational dynamics, not analytical flaws.

Performance change requires coordinated movement across multiple levers, not optimization of a single dimension. The companies that move the needle are those that identify how different strategic moves reinforce each other.

Play Bigger

Al Ramadan, Dave Peterson, Christopher Lochhead, Kevin Maney (2016)

Core Argument

Play Bigger presents a provocative thesis: the most successful companies do not just build better products or compete more effectively within existing categories—they create and dominate entirely new categories. Category design is a deliberate discipline, not a happy accident. Companies like Netflix (streaming entertainment), Salesforce (cloud CRM), and Tesla (electric vehicles, but more broadly, the 'car as a computer' category) didn't just win in existing markets—they created new ones. The book argues that category creation and domination is more valuable and durable than within-category competition because the category king captures 76% of a category's value, regardless of how many competitors emerge later.

The authors describe category design as a set of deliberate practices: identifying a 'big contrarian idea' that challenges conventional thinking in your industry, naming the new category clearly and compellingly, educating the market about why the problem matters and why your solution is correct, and positioning your company as the inevitable leader of the category. This is not marketing in the traditional sense (promoting your product) but rather market creation—teaching the market that a new category exists and that your company is its king. Category design also involves what the authors call 'lightning strikes'—major events, partnerships, or competitive moves that crystallize market recognition of the category.

The book emphasizes that category design is inherently strategic and inherently risky. Creating a new category means that your initial market is small—you're asking customers to adopt something new and unfamiliar. Many category design attempts fail. But the upside, if successful, is extraordinary: not just a larger market but market leadership and pricing power. The book also notes that category design is increasingly important in technology-intensive industries because technology creates the possibility of new categories faster than in traditional industries. An organization with the capability to identify and define new categories has a sustainable advantage.

Key Concepts

Category Design

The deliberate discipline of creating a new market category through identifying a contrarian insight, naming the category, educating the market, and establishing your company as the category king. This is different from product design (designing a better product) or marketing (promoting an existing product). Category design is about market creation.

Category King

The leader of a new category who captures the majority of value (research shows 76% of category value) and maintains pricing power. The category king is not the company with the best product but the company that most successfully defines and shapes customer perceptions of the category. Being first to market is helpful but not sufficient—establishing your vision as the definition of the category is what matters.

The Magic Triangle

The intersection of Company (your capabilities, vision, resources), Product (what you're creating), and Category (the new market you're defining). Successful category design requires alignment of all three: you need a company with conviction and capability to create a product that serves a new category that customers didn't know they needed.

Lightning Strikes

Major events, partnerships, announcements, or competitive moves that crystallize market recognition that a new category exists. Lightning strikes are often not the product launch itself but rather subsequent events that confirm the category's importance (e.g., a major enterprise customer adoption, a regulatory shift, a partnership that validates the category).

The Flywheel Effect

As the category king establishes market position, it becomes easier to continue dominating: customers associate the category with the king, new entrants struggle to redefine the category, partners preferentially align with the king, capital flows more readily, and talent is attracted. This creates a self-reinforcing cycle that makes category leadership increasingly durable.

Application to AI Transformation

Play Bigger challenges organizations to ask a provocative question about AI: Are we just using AI to improve an existing business, or can AI enable us to create a new category? Many organizations default to the former—'How can AI make our manufacturing more efficient? How can AI improve our customer service?' These are valuable questions, but they don't address the category design question. The contrarian question is: 'What new category does AI make possible that didn't exist before? What problem could we solve, or what value could we create, with AI that is fundamentally different from what the industry has traditionally offered?' For example, companies like Stripe reimagined payment processing (not just by making it more efficient but by creating an API-first, developer-focused category), and OpenAI created the generative AI assistant category. These are category design plays enabled by AI, not just AI-driven efficiency.

For organizations in traditional industries (manufacturing, finance, healthcare, etc.), the category design question might look like: Could AI enable us to compete in a fundamentally different way? For example, a manufacturing company might ask: 'Could AI shift us from selling products to selling outcomes? Could we create a new category of 'AI-assisted manufacturing advisory' where we sell not just equipment but AI-driven optimization services?' A financial services company might ask: 'Could AI enable us to create a new category of 'autonomous portfolio management' where we shift from selling investments to selling outcomes?' These category design questions are more ambitious than incremental AI use cases, but they have higher potential return.

Chapters 4-5 of 'AI Transformation Made Simple' on 'Sources of Power' and 'The Advantage Test' implicitly raise the category design question. The Advantage Test asks whether your AI initiative creates a defensible, sustained competitive advantage or just matches what competitors will inevitably do. Category design is the ultimate source of sustained advantage because it places you as the definer of a new market. The book should explicitly encourage leaders to ask: Beyond using AI to improve the existing business (which is important), does our AI capability position us to create a new category? This requires different resource allocation, different risk tolerance, and different strategic thinking than optimization-focused AI. But for organizations with the capability and vision to execute it, category design is the highest-return AI strategy.

Key Ideas to Reference

The category king captures 76% of a category's value. Being first in the category is less important than being perceived as the definer and leader of the category.

Category design is about changing what customers think is possible, not just offering them a better product. It requires identifying a contrarian insight and the courage to pursue it.

Most companies compete within existing categories. The most valuable companies create new ones. Category design is increasingly possible in technology-driven industries.

Lightning strikes crystallize market recognition of a new category. These are not always the product launch but rather subsequent events that confirm the category's importance and your company's leadership.

High Output Management

Andy Grove (1983/1995)

Core Argument

Andy Grove's foundational management text begins with a deceptively simple insight: a manager's output is the output of their organization. This means that management effectiveness is not about the manager's personal productivity but about multiplying the productivity of others. The book applies production principles (used in manufacturing) to the 'production' of management: how do you create a system where organizations predictably produce output? Grove uses metaphors of assembly lines, quality control, and process optimization to explain management practices. This framework has proven remarkably durable—the book remains in print more than 40 years after its initial publication because its core insights about leverage and organizational output remain valid.

Grove identifies several key management leverage points: indicators and measurements that tell you about organizational health (these are your 'black box' of management—if you can't measure it, you can't manage it), meetings that serve specific purposes (one-on-ones for coaching and feedback, staff meetings for information sharing and group decision-making), and task-relevant maturity (the principle that the same management approach doesn't work for all people—more experienced people need less direction, while less experienced people need more). The book emphasizes that management is not just about setting direction; it's about establishing processes and systems that ensure direction is executed. A manager who delegates but doesn't monitor is not doing their job—they must establish measurement systems that tell them whether execution is on track.

Grove's framework has been particularly influential in technology companies, especially startups and high-growth organizations where the pressure for output is intense. The book acknowledges that management can be a source of drag on organizational output if not done well, but it can also be a powerful multiplier. The difference between high-output management and poor management is often not a matter of intelligence or effort but of systematization—establishing the practices, indicators, and processes that enable an organization to reliably produce output. This is not about micromanagement (which is a sign of bad task-relevant maturity assessment) but rather about systems that enable autonomy while maintaining accountability.

Key Concepts

Output = Output of Your Organization

A manager's personal productivity is largely irrelevant. What matters is the output of the team and organization they manage. This reframes management from personal contribution to organizational multiplication. A great manager might spend more time on management activities that enable others' productivity (coaching, removing obstacles, establishing systems) than on personal work.

Leverage and Managerial Activities

Some managerial activities have high leverage (affecting the output of many people) while others have low leverage (affecting only one person). Meetings, training, goal-setting, and hiring are high-leverage activities. Operational tasks are typically low-leverage. A manager should allocate time toward high-leverage activities. The leverage multiplier is the number of people affected by your action divided by the effort required.

Production Principles Applied to Management

Management can be thought of as a production process: inputs (resources, goals, capabilities), manufacturing/transformation (the management system that converts inputs to output), quality control (measurement and feedback), and output (organizational results). Improvements to the management process improve organizational output just as improvements to manufacturing processes improve factory output.

Indicators and the 'Black Box' of Management

Indicators are quantitative measures that tell you about organizational health and progress toward goals. Grove emphasizes that management is largely a 'black box'—you can't see directly what's happening inside the organization, so you must establish indicators that reveal it. Without indicators, you're flying blind. Good indicators are leading (predictive of future results) and controllable (the organization can influence them).

Task-Relevant Maturity

The principle that management approach should vary based on the person's experience and maturity in their role. A new employee (low task-relevant maturity) needs specific direction, close monitoring, and frequent feedback. An experienced employee (high task-relevant maturity) can operate more autonomously but still needs alignment and support. Applying the wrong management approach (e.g., micromanaging an experienced person or delegating without support to an inexperienced person) creates problems.

Application to AI Transformation

Grove's concept of leverage is directly relevant to AI strategy because AI is perhaps the ultimate leverage tool in modern organizations. An AI system that augments the work of a human can multiply that person's output—a data analyst with AI tools can accomplish more; a customer service representative with an AI assistant can handle more interactions at higher quality; an executive with AI-powered decision support can make decisions faster and more accurately. Chapter 11 of 'AI Transformation Made Simple' on 'Measuring What Matters' applies Grove's indicator framework to AI measurement: you must establish clear, measurable indicators of whether AI is actually improving organizational output. These indicators should be leading (predictive of future performance) and should measure organizational output, not just AI system metrics. For example, 'model accuracy' is an AI metric, but 'decision quality' or 'time to decision' are organizational output metrics.

Grove's emphasis on task-relevant maturity is particularly valuable for AI adoption strategy. Different parts of the organization have different levels of maturity with AI. Some teams are advanced (early AI adopters, strong data skills) while others are beginners. The same 'one size fits all' AI adoption approach won't work. Advanced teams might need minimal guidance—provide them the AI tools and capability and let them innovate. Less mature teams need more support: training, templates, clear processes, closer monitoring. A manager's job in AI transformation is to diagnose task-relevant maturity and apply the appropriate level of direction and support. Additionally, Grove's point about micromanagement applies: giving experienced AI teams a prescribed set of use cases and requiring approval for everything stifles innovation. But providing novice AI users no guidance or training sets them up for failure. The right approach varies by maturity level.

Grove's production principles framework is useful for understanding AI transformation as a systematic change, not just a technology rollout. You can think of AI transformation as a production process: Inputs (data, AI talent, business problems to solve), Manufacturing (the organizational change process that integrates AI into decision-making and operations), Quality Control (measurement and feedback on whether AI is improving output), and Output (measurable improvements in organizational performance). Many AI transformations fail because they focus on inputs and technology without systematizing the organizational transformation (manufacturing step) or establishing clear measurement (quality control). Grove's framework highlights that all steps matter. Additionally, Grove's emphasis on indicators is central: the organizations that succeed at AI measure whether AI is actually improving organizational output. Those that only measure AI system metrics (model accuracy, inference speed) or only count AI projects launched often find that AI isn't translating to business value. The measurement framework must target organizational output.

Key Ideas to Reference

A manager's output is not their personal productivity—it is the output of their organization. This fundamentally changes how you think about your job as a manager.

Management leverage is the ratio of impact to effort. High-leverage activities affect many people; low-leverage activities affect few. A manager should allocate time toward high-leverage activities.

Indicators are your window into the 'black box' of management. Without measurements, you cannot know whether the organization is on track or whether your management practices are working.

Management approach should match task-relevant maturity. Different people need different approaches. Applying the wrong approach—whether it's too much direction or too little—creates problems.