I picked this up expecting a business book.
I got a mirror.
Everyone has one.
Every company deck has a slide called Our Strategy. Every roadmap gets signed off with “this aligns with our strategic goals.” Every OKR doc starts with a vision statement.
And almost none of it is strategy.
That’s the central argument of Richard Rumelt’s book. And once you see it, you can’t unsee it.
Rumelt is specific about this. Bad strategy isn’t just the absence of good strategy. It has its own shape.
It looks like:
Kodak is the classic example of the last one. It’s the 1970s. Digital photography is taking over. What’s Kodak’s strategy? Double down on film. Fight digital at every turn. Protect the old business. The bitter irony: Kodak invented digital photography. They just refused to diagnose what was actually happening.
Reading this as a designer, I recognised every hallmark.
I’ve sat in rooms where the “strategy” was:
“We want to be the best product in the space.”
That’s not a strategy. That’s a desire.
This is where the book earns its keep.
Rumelt calls it the kernel.
Three parts:
1. Diagnosis: what’s the actual challenge? 2. Guiding policy: how will you deal with it? 3. Coherent actions: what will you actually do?
That’s it.
The diagnosis names the problem clearly. Not “we need to grow” but why you’re not growing, and what specifically is the obstacle.
The guiding policy doesn’t solve the problem. It sets the approach. It rules things out. It creates constraints that focus effort.
The coherent actions are the decisions that flow from it. They reinforce each other. They’re specific. They’re coordinated.
The clearest example in the book is Apple in the late 90s.
Steve Jobs returns and the company is weeks from bankruptcy. Product line is a mess: fifteen different desktop models, no consumer could tell them apart.
Diagnosis: months from bankruptcy, chaotic product lineup bleeding cash and focus.
Guiding policy: radical simplification. Cut everything non-essential. Then wait for the next big thing.
Coherent actions: fifteen desktop models become one. Manufacturing moves to Taiwan. All attention turns toward mobile.
The rest is history.
That’s a kernel. Notice how each action follows from the policy, and the policy follows from the diagnosis. Nothing is decorative.
Product design is full of strategy debt.
You see it in roadmaps with fifty items and no hierarchy. You see it in discovery that finds the problem but never shapes a response. You see it in design systems built without a guiding principle.
The kernel is actually just good product thinking with a clearer name.
Diagnosis is research. It’s the phase where you figure out what’s actually wrong, not what you wish was wrong.
Guiding policy is the decision about what kind of product you’re building and what you’re willing to sacrifice to do it well. This is the thing most teams skip.
Coherent actions are the design decisions. Not a list of features. Decisions that work together, that point in the same direction, that follow from the policy.
When all three are present, a team knows why they’re doing what they’re doing. When they’re absent, every decision is a debate.
There’s also a risk that’s specific to designers and researchers: being so close to the work that you can’t see whether the work itself is pointed in the right direction.
You can be excellent at the tree level (sharp insights, tight interaction design, thorough testing) while the forest is going in the wrong direction entirely. The strategy above you determines how much your work actually matters. If you’re not helping shape it, you’re stuck executing someone else’s, and living with the consequences.
One thing the book covers that I hadn’t really named before: the idea of a proximate objective.
A good strategic objective is one that’s actually achievable. Close enough that you can orient real actions behind it.
The moon landing is a good example. The goal wasn’t “beat the Soviets in space.” It was: land a person on the moon by the end of the decade. Specific. Binary. Achievable. And it focused an entire national effort.
That’s proximate because it gave everyone something concrete to point at. It ruled things out. NASA wasn’t trying to win at everything. Just one thing.
Rumelt talks about leverage.
Good strategy doesn’t try to do everything. It finds the place where focused effort produces disproportionate results.
Find the pivotal point. Apply concentrated force there.
He uses David and Goliath to make this concrete. Goliath has size and brute force. Nobody can match him in a straight fight. But he has weaknesses: exposed, unarmoured. David has speed, agility, and a long-range weapon.
David doesn’t try to win on Goliath’s terms. He finds the asymmetry and exploits it.
That’s leverage.
For a product team, that’s the difference between:
Most teams do the former. Because it feels more productive. Because it keeps everyone happy. Because saying no is hard.
Strategy, in Rumelt’s framing, is mostly about saying no.
I’ve read a lot of product and business books. Most of them describe things I already half-knew.
This one changed how I think about a word I’d been using wrong for years.
Strategy isn’t a vision. It’s not a goal. It’s not a list of things you want to do.
It’s a clear-eyed answer to a hard problem, with a coherent set of choices that follow from it.
Most of what gets called strategy is noise.
The real thing is rarer than it looks. And worth a lot more than people think.