Good strategy starts with a clear diagnosis and an action agenda

Executive overview

Most "strategies" are lists of goals or ambitions — not strategies. A strategy is a design for overcoming a specific, high-stakes challenge.

The kernel of good strategy has three elements: a diagnosis of the situation, a guiding policy, and coherent actions. Remove any one and it stops being strategy.

Start with the challenge, not the plan. Ask: what makes this hard? Then build an action agenda — not a strategy document — around what you can actually address.

The most common strategic failure is mistaking a list of goals for a strategy.

The three elements of the kernel

• Diagnosis: an understanding of the challenge, not just a description of what you want. Ask "what makes this hard?" not "what do we want?"
• Guiding policy: how you're dealing with the situation — the core directional choice, not a wish list
• Guiding policy isn't goals; it's a decision about what you are and are not going to do
• Coherent actions: the things you will actually do — must implement the policy and must not contradict each other
• "We will grow and increase return on equity" is incoherent; growth and ROE often trade off
• All three must be present; missing one means it isn't a strategy

What bad strategy looks like

• A list of performance goals ("we will grow revenue, cut costs, and improve NPS") — these are ambitions, not strategy
• Fluff: abstract, high-sounding language that obscures the absence of real choices
• Diagnosis-free action: jumping from a problem statement to a solution without understanding why the problem exists
• Obama's education example: "the US is falling behind in education" is correct, but "more people should go to college" is not a diagnosis of why 15-year-olds can't do elementary math
• Mistaking coordination aspiration for strategy ("the 17 intelligence agencies will work together more effectively")

Focus and power

• Focus is the fundamental source of power in strategy; diffusing effort across too many priorities destroys it
• Power comes from asymmetry: something you have, know, or can do that competitors cannot match equally
• Sources of power include: reputation, relationships, proprietary knowledge, being first, network effects, a specific customer base
• Power is not permanent — IBM's deep corporate relationships became a liability when computing moved to cloud
• The magnifying glass analogy: source of power (sun) + focus (lens) + right target (black thread) = result
• When evaluating any strategic bet, ask: where is the asymmetry that makes the odds not 50-50?

Network effects as a modern source of power

• Network effects: the more users, the more valuable the product — first articulated around the telephone
• Applies to software platforms, social media, marketplaces, and increasingly AI (more training data → better model)
• Amazon's size creates a switching cost; Google's search data makes Bing structurally hard to catch
• Two-sided markets (credit cards, marketplaces) layer buyer and seller network effects
• Generative AI may produce the strongest network effects yet seen — scale of training data likely to compound incumbents' advantages
• Twitter demonstrates network effects can survive almost everything: leadership change, name change, 80% staff reduction

Diagnosing well

• Diagnosis is not describing what you want; it is understanding what is actually happening and why
• "We're not growing fast enough" is a value statement; "we're not growing because X is holding us back" is a diagnosis
• Complex situations have multiple competing interpretations; you must resolve the argument before you can act
• Practice diagnosis by studying historical situations — the Great Depression, the fall of Nokia, Afghanistan — and forming your own interpretation before reading the consensus view
• The most important intellectual tool: think again. Your first frame is a life raft, not necessarily the truth

The crux: finding the hardest problem

• The crux is the hardest part of the challenge — in climbing, if you can't do the crux, don't do the climb
• In strategy: find the challenge that is both most important (close to your ambition) and actually addressable
• Design insight comes from immersing yourself in what makes the problem hard, then seeing a path through
• Examples: I.M. Pei's glass pyramid — the constraint (don't obscure the Louvre) generated the solution (transparent building)
• SpaceX: focused on why reusability was hard (re-entry burns the vehicle), which surfaced the propulsive landing idea
• Insight isn't magical. It comes from sustained attention to the difficulty. It may arrive in the shower, not the boardroom

How to create an action agenda (not a strategy)

1. Surface ambitions — what do you want? Accept all of them
2. Filter by what is achievable now — which ambitions can you make progress on today?
3. Ask what makes that hard — find the real challenge
4. Choose the crux — the most important and most addressable challenge
5. Define action steps — things you will do, not goals you will achieve
6. Check coherence — do the actions reinforce each other, or fight each other?

• Don't call it a strategy; call it an action agenda
• Avoid mission, vision, and values documents as preconditions — they're a different literary form
• Timelines: large company action steps may span 2-3 years; startup action steps may span 6-12 months

Organizational dynamics as the enemy of strategy

• Strategy inside an org requires resolving disagreements about diagnosis and making uncomfortable choices
• People in organizations have private interests that don't surface clearly: contracts, headcount, conceptual agendas
• Nokia lost its smartphone lead partly because a matrix org so diffused power that no one could authorize building a touchscreen despite the CEO demanding it
• The growing emphasis on "leadership" and personal development has displaced the harder skill of actually deciding and directing
• Someone has to say: "we're doing this, Bob is in charge of A, John is in charge of B" — and fewer CEOs are willing to do it
• When everyone's priority gets included, the word "priority" loses meaning (a tower controller giving "priority" to three planes on runway five)

Strategy for startups

• Pre-PMF founders are making a bet under uncertainty — be clear about the nature of the bet
• Research on Silicon Valley startups: initial product-market hypotheses usually don't work; the winners pivot
• The skill is being simultaneously convinced you will win and willing to shift when signals change
• Treat early strategy as a search: truffle hound, not march on a fixed objective
• Value denies exercise: what should people be able to buy but can't? That gap is an opportunity
• Perfect-object exercise: how should this thing actually work? Enumerate requirements, find contradictions, find solutions
• Important + achievable = the founder's job: find a problem that matters and that you can actually solve

On history as strategic training

• There is no science of strategy — no equations of stress on a beam
• Strategic thinking is largely analogical reasoning from prior human experience
• Read biographies and business histories, not strategy theory
• Go back to primary sources: read the Wall Street Journal front page from September 1928 forward, day by day, and form your own view before reading anyone's theory
• Recommended: Only the Paranoid Survive (Andy Grove), Rockefeller biographies, Steve Jobs biography, Playing to Win (Roger Martin)
• The further back in history, the leaner the documentation; recent history is richer