Good Strategy/Bad Strategy: Rumelt for Business Execution
Richard Rumelt's Good Strategy/Bad Strategy is the book most directors wish they had read before their last planning offsite. It is blunt about what fails: goals without diagnosis, vision statements without choices, and "strategies" that are really lists of things leadership would like to be true.
Good strategy, Rumelt argues, has a kernel: a diagnosis of the real challenge, a guiding policy for dealing with it, and a set of coherent actions that align the organisation. That structure maps almost exactly to how growing businesses should run strategy execution: direction, objectives, KPIs, and linked work reviewed on a rhythm.
This article translates Rumelt's kernel for UK SME leaders and shows how teams using Elevale can turn good strategy from a document into an operating system.
Bad strategy: what Rumelt warns against
Bad strategy is recognisable once you know the pattern:
- Fluff: buzzwords that sound strategic but imply no action
- Failure to face the challenge: goals that ignore the hardest constraint
- Mistaking goals for strategy: "grow 30%" is not a strategy
- Bad strategic objectives: a long wish list with no coherence
Many SMEs produce bad strategy in kinder packaging: a polished deck, enthusiastic workshop, and twelve priorities that nobody can rank. By week six, daily work has absorbed the quarter and directors cannot say which choice mattered.
Read the three traps that keep good businesses stuck for how activity, strategy resets, and accountability gaps reinforce bad strategy in practice.
The kernel of good strategy
Rumelt's kernel has three parts. Each maps to a layer of execution.
1. Diagnosis
What is the real challenge? Not the presenting symptom, but the underlying difficulty leadership must address. Diagnosis requires honest assessment of capabilities, market position, constraints, and what is not working in the current model.
Elevale layer: strategic direction and live KPIs make diagnosis visible between offsites. Directors see whether the business is healthy while priorities stall, or priorities move while health deteriorates.
2. Guiding policy
The overall approach for dealing with the challenge. Not every tactic, but the choice that rules out other paths. Guiding policy creates focus: we will win here, this way, and not scatter effort across every opportunity.
Elevale layer: company OKRs translate guiding policy into quarterly outcomes leadership can inspect. Three to five objectives with named sponsors, not fifteen ambitions on a slide.
3. Coherent action
Policies and actions that reinforce each other. Coherence means initiatives do not fight each other, resources align to the diagnosis, and functional teams can explain how their work serves the approach.
Elevale layer: tasks and projects linked to key results. Orphan work is incoherent action: busy, defensible locally, disconnected from the kernel.
Good strategy vs bad strategy in SMEs
| Bad strategy signal | Good strategy alternative |
|---|---|
| "Become the leading provider of..." | Diagnosis: we lose deals on speed, not price |
| Twelve company priorities | Three objectives with trade-offs named |
| OKRs copied from a template | Key results tied to diagnosis and KPIs |
| Projects green, outcomes flat | Coherent actions linked to key results |
| Annual deck nobody opens | Weekly reviews on exceptions and decisions |
Sources of power Rumelt describes (and how to use them)
Rumelt also describes sources of strategic advantage: leverage, focus, growth through design, and competitive insight. For SMEs, the practical sources are usually simpler:
- Focus: concentrate resources on the constraint that limits growth or margin
- Chain link: improve the weakest link in delivery or acquisition, not every link equally
- Design: align incentives, measures, and ownership so behaviour matches the kernel
Software does not create focus. It makes focus visible and harder to accidentally undo mid-quarter.
Applying Rumelt in a leadership quarter
Quarter start (diagnosis + policy):
- State the challenge in one paragraph leadership agrees is true
- Write guiding policy: our approach this quarter is...
- Set three to five company OKRs that express the policy in outcomes
- Confirm KPIs that show health while those outcomes progress
Weekly (coherent action):
- Review objective exceptions, not every task
- Cut or re-scope work that cannot link to a key result
- Log decisions that change resource allocation
Quarter end (honest diagnosis again):
- Did the kernel hold, or did bad strategy creep back in as fluff and extra priorities?
- What evidence from KPIs and OKRs informs next quarter's diagnosis?
See quarterly business review agenda for a practical review structure.
Proximate objectives: strategy in steps, not leaps
Rumelt also emphasises proximate objectives: goals close enough to be feasible, far enough to matter. SMEs often fail by setting objectives that sound inspiring but cannot be reached with current capabilities. Leadership burns a quarter proving the objective was fantasy, then resets with less trust.
Proximate objectives connect diagnosis to coherent action in achievable chunks. This quarter we fix onboarding conversion, not "become market leader." Next quarter we scale what worked. Elevale quarterly OKRs should read as proximate objectives tied to evidence, not distant wishes.
Pair proximate objectives with strategy vs OKRs vs KPIs so directors know which layer answers which question.
Example: Rumelt kernel for a 40-person UK agency
Diagnosis: revenue is flat because utilisation is high but average project margin fell. New business wins deals leadership cannot deliver profitably. The constraint is senior delivery capacity and scoping discipline, not lead volume.
Guiding policy: win fewer, better-fit clients and protect senior time on highest-margin work. Stop pursuing low-fit RFPs even when pipeline looks thin.
Coherent action: company OKR on margin recovery; sales key result on qualified pipeline quality; delivery key result on senior utilisation on top accounts; pause two marketing initiatives that drove low-fit leads.
Without the kernel written plainly, the agency would have launched another lead-gen campaign (incoherent action) while the real challenge worsened.
How Rumelt complements other execution books
Rumelt answers what good strategy is. Other books answer how to run it under pressure:
- The Goal finds the system constraint Rumelt's diagnosis should name
- 4 Disciplines of Execution protects the kernel from the whirlwind
- Clausewitz explains why coherent action degrades without visibility and rhythm
Return to the strategy books hub for the full reading map.
Common mistakes when using Rumelt
- Renaming the old plan "diagnosis" without new insight
- Guiding policy that says "focus on customers" without ruling anything out
- Coherent action on paper while teams run incoherent tasks in the project tool
- Treating Rumelt as criticism of the team instead of criticism of the strategy document
Kernel checklist for your next planning session
Before you approve company OKRs, leadership should answer yes to each:
- Diagnosis names the hardest constraint or challenge, not a symptom
- Guiding policy rules out at least one attractive path
- Each OKR traces to the policy in one sentence
- Active projects without key-result links are flagged for pause
- KPIs show whether the business can afford the bet this quarter
If any answer is no, you have a workshop output, not a kernel. Fix it before teams cascade work. Rumelt's standard is strict because bad strategy wastes the quarter before anyone admits the plan was hollow.
Next steps
- Rewrite this quarter's plan as diagnosis, guiding policy, and three coherent OKRs
- Audit active projects: which support the kernel, which are orphan work?
- Read what is strategy execution software for the system layer
- Explore more books in our strategy books for business execution hub
Start your 14-day free trial and connect Rumelt's kernel to OKRs, KPIs, and reviews in one platform.