Strategy is a diagnosis

Confession: for years, the word "strategy" made my eyes glaze over.

Every strategy document I'd seen looked about the same. A vision. Some values. Four to seven goals with arrows. A slide with a mountain on it.

Then I read Richard Rumelt's Good Strategy / Bad Strategy (2011), and it gave me words for what had been bugging me. This year I got to put it to work as process lead for strategy work at Nav. Here's what stuck.

The kernel

Rumelt argues that good strategy has a simple core. He calls it the kernel, and it has three parts:

1. A diagnosis. What's going on here? A good diagnosis names the critical challenge and simplifies an overwhelming situation into something you can actually act on.
2. A guiding policy. The overall approach to dealing with that challenge. It doesn't spell out every step. It points in a direction and rules things out.
3. Coherent actions. Concrete steps that carry out the guiding policy and support each other. Coordinated, resourced, real.

Notice the order. Diagnosis comes first. Not goals. Not vision. Diagnosis.

You can't prescribe anything until you know what's wrong. Doctors figured this out a while ago.

Four signs of bad strategy

Rumelt also describes four hallmarks of bad strategy. Once you've seen them, you'll spot them everywhere:

• Fluff. Big words dressed up as insight. "We will drive customer-centric, value-based transformation." Great. What does that mean on a Tuesday?
• Failure to face the challenge. If you don't name the problem, nobody can judge whether the strategy solves it.
• Mistaking goals for strategy. "Grow 20%" is a wish. A goal says what you want. A strategy says how you'll get past what's stopping you.
• Bad strategic objectives. Either a long, unprioritized list of things to do (Rumelt calls it a "dog's dinner") or "blue-sky" objectives that just restate the dream without saying how to get there.

One of Rumelt's points that I find oddly comforting: bad strategy isn't just the absence of good strategy. It usually comes from avoiding hard choices. Choosing means saying no to someone, and that's uncomfortable. Fluff is what uncomfortable looks like on a slide.

What I learned applying it at Nav

Since spring 2026 I have been process lead for strategy work in the Section for employer services at Nav. The job is to help the leadership team clarify their strategic frame, priorities and choices. We work through a series of workshops, I have interviewed product and domain leads, and in May we brought in the wider section at a gathering of around 70 people.

What they landed on is theirs to tell. But the process taught me a lot.

Start with the problem, even when everyone wants to start with goals. Goals feel productive. They're also the easiest place to hide. We spent the early workshops building a shared understanding of the situation before talking about where to go. It feels slow. It's also what makes everything after it faster.

Interviews before workshops. One-on-one conversations surface things that never come out in a big room. People are more candid when there's no audience. It also meant I walked into workshops already knowing where the disagreements were.

Disagreement is data. When two leaders describe the challenge differently, that's not the process failing. That is the diagnosis work.

Write the diagnosis in plain language. If it doesn't fit on a page and survive being read out loud, it's not done.

A guiding policy has to rule something out. If nothing gets harder to justify after you've agreed on it, it's not a policy. It's a mood.

Why designers should facilitate strategy

Problem framing is design work.

It's what we do in every discovery phase. We resist jumping to solutions. We dig into what's really going on. We make the mess visible. We check our understanding against reality. Rumelt's diagnosis is basically a problem statement with higher stakes and fancier attendees.

Designers bring a few useful habits to the table:

• Comfort with ambiguity. The fuzzy front end is our home turf.
• Making things visible. Putting a situation on the wall so a group can reason about it together.
• Facilitation. Years of workshops, retros and co-creation sessions add up.
• Healthy suspicion of assumptions. "How do we know that?" is a strategic question.

My psychology background helps here too. A leadership team is a group, and groups fall into the same traps as everyone else: anchoring on the first idea, deferring to the most senior voice, agreeing too fast to avoid conflict. Knowing the traps doesn't make you immune, but it helps you design around them.

Practical tips

If you're about to facilitate a strategy process, here's my cheat sheet:

• Introduce the kernel early. Shared vocabulary saves hours of talking past each other.
• Ask "what's stopping us?" more often than "where do we want to be?"
• Use silent writing before discussion, so the loudest voice doesn't set the frame.
• Sort draft objectives into "goal" or "strategy". Most turn out to be goals. That's a useful moment, not an embarrassing one.
• Test the guiding policy: what does it make easier to say no to?
• Check for coherence: do the actions support each other, or are they just a list?
• Keep a parking lot for good ideas that aren't strategy. They deserve a home, just not this one.
• Plan how it reaches everyone else. A strategy that only the leadership team understands isn't finished.

The short version

Strategy isn't a document. It's a diagnosis you're willing to act on.

Start there, and the goals tend to sort themselves out.