Lesson 7 of 8

The strategy kernel

9 min read Strategy 5-question check Flashcards included

Lesson 8 is about writing an engineering strategy. This lesson is the test that tells you whether what you wrote is one.

What a strategy is

Rumelt’s definition: a cohesive response to an important challenge. A coherent set of analyses, concepts, policies, arguments and actions responding to a high-stakes challenge.

Note what is bundled in. Many people treat strategy as a big-picture direction, separate from any specific action. Rumelt says that definition creates the chasm between strategy and implementation, and that most strategy work becomes wheel-spinning once you accept it.

Which explains a complaint you have heard. Executives who say they have a sophisticated strategy process but a huge execution problem have usually confused strategy with goal setting. When the process is a game of setting performance targets, so much market share, so much profit, there remains a yawning gap between ambition and action. Strategy is about how an organization will move forward.

A leader can set goals and delegate figuring out what to do. Rumelt’s response is that this is fine, and it is not strategy, so skip the spin and call it goal setting.

The kernel

The kernel
The logical structure shared by every good strategy: a diagnosis, a guiding policy, and a set of coherent actions.
Diagnosis
An account of what is going on. It names or classifies the situation, links facts into patterns, and directs attention toward some issues and away from others.
Guiding policy
The approach for dealing with the obstacles named in the diagnosis. A signpost marking the direction forward, without defining the details of the trip.
Coherent actions
Feasible, coordinated policies, resource commitments and actions designed to carry out the guiding policy.
names the difficulty to address carried out by and rules these out the part that expires so the policy changes too The situation Diagnosis: what is going on here? Guiding policy: the approach Coherent actions Ruled out Revisit when circumstances change
The situation
Diagnosis:what is goingon here?
Guidingpolicy: theapproach
Coherentactions
Ruled out
Revisit whencircumstanceschange
Three parts, one direction. The dashed edges are the two things a written diagnosis buys you.

Diagnosis is a decision, not a description

Rumelt’s colleague sat in on ten of his classes and observed that there is really only one question being asked in every case: what is going on here?

A great deal of strategy work is not deciding what to do. It is the more fundamental problem of comprehending the situation.

At a minimum, a diagnosis names or classifies the situation, links facts into patterns, and suggests that more attention be paid to some issues and less to others. A good one can transform how you see the situation. Classifying it as a certain type also opens access to how analogous situations were handled before.

Take his Starbucks example. In 2008 the company had flat or declining same-store traffic and return on assets down from 14 percent to about 5.5 percent. The candidate diagnoses:

  • a problem in managing expectations
  • a search for new growth platforms
  • an eroding competitive advantage

None of these is an action. Each suggests a range of things that might be done and sets other classes of action aside. Picking between them is the strategic act, and it happens before anyone has proposed doing anything.

Two more reasons to write it down. An explicit diagnosis lets you evaluate the rest of the strategy, because a guiding policy is only good or bad relative to the difficulty it addresses. And it lets the strategy be revisited when circumstances change, because you can see which premise expired. An unwritten diagnosis leaves you with a plan you can neither justify nor safely retire.

Guiding policy and coherent actions

The guiding policy specifies the approach for dealing with the obstacles the diagnosis called out. Rumelt’s image is a signpost: it marks the direction forward without defining the details of the trip.

Coherent actions are feasible, coordinated policies, resource commitments and actions designed to carry out that policy. His insistence: they are not implementation details, they are the punch in the strategy. A strategy that fails to define plausible, feasible immediate actions is missing a critical component.

Coordinated is the load-bearing word. In the scenario, two teams cite one pillar and one team cites another to reach opposite conclusions about the same migration. The actions are not coherent with each other, and nothing in the document could have caught that, because nothing in it was derived from a shared diagnosis.

Four hallmarks of bad strategy

Bad strategy is not the absence of good strategy. It has a life and logic of its own, a false edifice built on mistaken foundations. It comes from avoiding analysis of obstacles because negative thoughts are believed to get in the way, from treating strategy work as goal setting rather than problem solving, and from avoiding hard choices so as not to offend anyone, which produces a document that covers all the bases instead of focusing resources.

HallmarkWhat it looks like
FluffBuzzwords and needless complexity over restated obvious ideas
Failure to face the problemNo defined challenge, so quality cannot be assessed at all
Mistaking goals for strategyStatements of desire rather than plans for overcoming obstacles
Bad strategic objectivesObjectives that miss the critical issues, or are impracticable

Fluff. Rumelt’s example is an Enron presentation on bandwidth markets: a diagram with the surface appearance of analysis that turned out to be a stew of half-truths, complex drawings and buzzwords. The “strategies” listed were an electronic trading platform, an over-the-counter broker, an information provider. These were not strategies, they were names, like butcher, baker and candlestick maker. His test: a hallmark of true expertise is making a complex subject understandable, and a hallmark of bad strategy is unnecessary complexity masking an absence of substance.

Failure to face the problem. A strategy is a way through a difficulty. If the challenge is not defined, you cannot assess the quality of the strategy, which means you cannot reject a bad one or improve a good one. This is the fatal one, because it disables the machinery that would catch the other three.

Mistaking goals for strategy. Statements of desire rather than plans for overcoming obstacles.

Bad strategic objectives. An objective is set as a means to an end. It is bad when it fails to address critical issues, or when it is impracticable.

The test

You now have one, and it takes a minute.

Find the diagnosis. If there is no sentence saying what is going on, you are holding goals. Then ask what the guiding policy rules out. If nothing, everyone can cite it for everything, which is the scenario. Then check that the actions are coordinated with each other rather than merely listed under the same heading.

Nobody objecting to your strategy is not a sign that it is good. It is usually a sign that it costs nobody anything, which means it will not settle the argument you wrote it to settle.

The next lesson takes this structure into an engineering organization, where Larson names Rumelt’s three parts directly, and where the awkward common case is that the company strategy you are meant to derive from does not exist.

Source: Richard Rumelt, Good Strategy / Bad Strategy, Ch. 3 'Bad Strategy' and Ch. 5 'The Kernel of Good Strategy'

Check your understanding

Answer to reveal the explanation. Nothing is scored.

1An engineering strategy reads: 'Reach 99.99% availability, ship the platform migration, and grow the team to 90.' What is missing?

Rumelt's term is mistaking goals for strategy: statements of desire rather than plans for overcoming obstacles. Specificity does not rescue it. You cannot tell from this document what the company's actual difficulty is, which means you cannot tell whether these three things address it or whether they conflict.

2A CEO says the strategy is sound but the company has an execution problem. What does Rumelt say is usually happening?

His argument is that coherent actions are not implementation details, they are the punch in the strategy. Defining strategy as broad direction divorced from action manufactures the chasm that the word 'execution' then names. If the leader sets targets and delegates figuring out what to do, that is goal setting, and it should be called that.

3Why does Rumelt insist the diagnosis be an explicit, written element rather than shared understanding?

Two functions, both lost if it stays implicit. Nothing can be judged without it, since a guiding policy is only good or bad relative to the difficulty it addresses. And when the situation changes you have no way to know which parts of the strategy should change with it, so the whole thing either survives unearned or gets thrown out.

4Starbucks in 2008 had flat traffic and return on assets down from 14% to 5.5%. One executive calls it 'an eroding competitive advantage', another 'a search for new growth platforms'. What does the choice between them do?

Rumelt's point exactly: none of the candidate diagnoses is by itself an action, but each suggests a range of things that might be done and sets other classes of action aside. Naming the situation is where the choosing happens. Everything downstream is constrained by a call most organizations never realise they made.

5Which statement matches Rumelt's view of bad strategy?

He treats it as an active construction with identifiable causes: avoiding analysis of obstacles because negative thinking is thought to get in the way, treating strategy as goal setting rather than problem solving, and dodging hard choices to avoid offending anyone. That last one produces a document covering all the bases, which is the opposite of focusing resources.

What are the three elements of Rumelt's kernel?
A diagnosis of the situation, a guiding policy for dealing with the critical difficulties, and a set of coherent actions to carry out that policy.
What is a diagnosis, and what question does it answer?
It answers 'what is going on here?' At minimum it names or classifies the situation, links facts into patterns, and says which issues deserve more attention and which less.
Why must the diagnosis be written down explicitly?
An explicit diagnosis lets you evaluate the rest of the strategy, and lets the strategy be revisited and changed when circumstances change. Without it there is nothing to judge the strategy against.
What is a guiding policy?
The approach to dealing with the obstacles named in the diagnosis. Rumelt's image is a signpost: it marks the direction forward without defining the details of the trip.
What are coherent actions, and why are they not implementation detail?
Feasible, coordinated policies, resource commitments and actions that carry out the guiding policy. Rumelt calls them the punch in the strategy; a strategy without them is missing a critical component.
What are the four hallmarks of bad strategy?
Fluff; failure to face the problem; mistaking goals for strategy; and bad strategic objectives.
What is fluff, in Rumelt's sense?
Restated obvious ideas dressed in buzzwords and unnecessary complexity, giving the surface appearance of analysis. His test: expertise makes a complex subject understandable, mediocrity makes a simple one complex.
Why is failure to face the problem fatal rather than merely incomplete?
If the challenge is not defined you cannot assess the quality of the strategy, and therefore cannot reject a bad one or improve a good one.
Rumelt's explanation for complaints about 'execution problems'
The executives complaining have usually confused strategy with goal setting. Setting performance targets leaves a gap between ambition and action, and strategy is about how the organization will move forward.
Is bad strategy just the absence of good strategy?
No. It has a life and logic of its own, a false edifice on mistaken foundations. It comes from avoiding analysis of obstacles, treating strategy as goal setting, or dodging hard choices so as not to offend anyone.