Product Strategy Is a Decision About What You Won't Do

Somebody asked you what your product strategy was, and you opened the roadmap. You opened it because the roadmap is the only document you own that answers a question shaped like that one. It has dates on it. It looks like a plan. And for about four seconds, while it loaded, you knew it wasn't the answer.

That roadmap isn't a bad document. Somebody put real work into it, and the work shows: second quarter is the reporting rebuild, third quarter is the mobile app plus single sign-on for the two customers who keep asking, fourth quarter is integrations. Every one of those is a real thing your team is going to build, in an order somebody thought about. The document is competent. It answers a different question than the one you were asked.

Everything on it is yours

Look at what those items have in common. You decide to do the reporting rebuild. You decide what order it goes in, who works on it, and roughly how long you'll let it run before you get nervous. Nobody outside your building gets a vote on a single line of it.

That is why the meeting where you present it goes smoothly. Everybody can agree to a list of things you're going to do, because nobody has to give anything up in order to say yes. Your engineering lead is fine with it: it's a lot of work, but it's real work. Sales is fine with it, because their customer's thing is on there in the fourth quarter, which is late enough that it can still be moved. The agreement feels like alignment. It's arithmetic: if everyone in the room can agree to it, it hasn't turned anything down.

So the roadmap couldn't answer the question. A list of what you're going to build carries nothing about what you'll refuse. The roadmap is doing its own job perfectly well. It tells you when things are going to happen, and you were being asked whether they should.

Forty years of work on the difference

The field has a document that does answer that, and it's been working on the distinction carefully for about forty years. The most-watched explanation of it anywhere is nine and a half minutes of Roger Martin, who ran the Rotman School of Management and wrote Playing to Win with A.G. Lafley, who ran Procter and Gamble. Close to seven million people have watched it.

Planning, in his account, is a list of activities, which is precisely what your roadmap is. Strategy is a theory about how you're going to win. What separates them is who gets to decide whether you were right.

The activities on a plan are all inputs. You buy them, you hire them, you schedule them. Every line on that roadmap is one of them, and you can tick off every single one with complete control the whole way through. A strategy names an outcome, and an outcome depends on customers, who don't work for you and never agreed to any of this. You can ship that entire roadmap, on time, exactly as written, and still lose, and that second result is the one nobody in your building can promise you.

Two definitions, fourteen weeks apart

Before you go and write a word of your own, look at where that definition came from. Harvard Business Review published that Martin video in the summer of 2022. Fourteen weeks earlier, the same channel published a different explanation of strategy, by Felix Oberholzer-Gee, who teaches at Harvard Business School.

His definition isn't Martin's. Martin says a strategy is a theory of how you win, and the tell of a good one is that you can't prove it in advance. Oberholzer-Gee says it's a plan to create value, and value is something you can put a number on. One of those turns on a choice you can't test. The other turns on a number you can estimate. Under the second video there are comments working out which of the two men is right.

I'll leave the two of them standing where they are. Blending them would produce a definition neither man wrote, and the blended version is always the one that sounds most reasonable and commits to least. What's useful for the page you're about to write is the part where they agree, and every version of this I've read agrees on three things: a strategy is a choice, a choice means giving something up, and it should fit on one page. Everything past those three belongs to somebody's particular framework.

The tell at six minutes forty-eight

Six minutes and forty-eight seconds into that nine-and-a-half-minute talk, Martin says a real strategy makes you anxious, and that the anxiety is the tell. You can't prove in advance that it's going to work, and you have to commit to it anyway.

Once you've heard that, the roadmap stops looking like a failure of rigor and starts looking like the sensible thing to reach for. A plan doesn't make you anxious, and that's the appeal of it. Everything on it sits inside your control, so you can check it before you commit. So when somebody asks you for a strategy and you're under pressure, you produce the document that doesn't frighten you. I would have done the same.

Four years of yes

That document costs something, and the cost is easiest to see in a company that never wrote one. Meridian is the composite company I've been carrying through this series. Forty people, six engineers, one designer, about sixty customers, and a founder who's also the chief executive and has held product since the beginning, because there was never anyone else to hold it. Meridian's strategy isn't written down anywhere. Here is what four years of that cost.

In year one they built for the first customer who paid them, a mid-size logistics operator. That was the right call. In year two a bigger company turned up wanting a custom integration into their warehouse system, and the number on it was ninety thousand dollars. They said yes. Of course they said yes. In year three a different large customer wanted a different integration, and they said yes to that one too, partly because they'd built one already and it felt like something they were good at now.

By year four, two of their six engineers spend most of their week maintaining integrations for three customers. That is a third of the engineering capacity of the company, permanently committed, and no one ever sat down and committed it.

Every one of those decisions was correct on the day it was made, which is what makes the pattern so hard to see from inside the company. There is no bad call anywhere in the list.

Eighteen directions

Somebody has described that exact accumulation better than I can, and the company he was describing had fifteen thousand people in it. There's a recording from 1997 of Steve Jobs taking questions at Apple's developer conference. He'd been back a few months, Apple was roughly a year from running out of money, and he'd just canceled a long list of projects. Somebody in the audience stands up and asks him, with real heat in it, why a project called OpenDoc had been killed. It's a hostile question, and some of the people who worked on OpenDoc are sitting right there.

The famous six words come out of that answer. Focusing is about saying no. You've seen it on a poster in somebody's office, usually in a font that has had all the difficulty sanded off it.

The useful part comes about forty seconds earlier. He says Apple had good engineers going off in eighteen different directions, all doing arguably interesting things, and the total came out less than the sum of the parts. Each of those projects made sense on its own. Together they made no sense at all. That is a very different claim from we had bad projects, because it means you can't find the mistake by looking at the decisions one at a time. Jobs doesn't pretend the fix is free, either. He says when you say no, people get angry, they go to the press, and the press writes something unkind about you. He opens that entire answer by apologizing to the people sitting in the room whose work had just been canceled.

The strategy nobody wrote

What's happening at Meridian is close to the opposite of what its founder would tell you. He'd say they haven't got around to writing a strategy yet, meaning the document doesn't exist. But the document isn't the thing.

Meridian has a strategy. You can read it straight off four years of decisions, and it goes like this: we serve mid-size logistics operators, plus whichever large enterprise asked us most recently, and we're running a custom integrations business on the side. That's a real strategy. It's coherent, you can write it down, and right now it's allocating two of their six engineers. Nobody in that company ever chose it. Their inbox wrote it.

If that sounds like your company, the document was never the place to look. A strategy you never wrote down leaves a pattern instead of a document, and the pattern is sitting in the last thirty things you said yes to.

The question that comes before the strategy work

Meridian's founder could start writing that page tomorrow, and I'd stop him. One question comes first, and it's what I ask when somebody shows me a set of products that don't hang together: how did this happen. That's a question about history, and it still has to come first.

The answer changes the job completely. If those products grew out of one original idea and drifted apart, then the coherence is real and it's been lost, and your job is to go and find it. If they arrived separately, there may be no shared spine to find at all, and writing one anyway gets you a document nobody can execute.

The diagnosis is free. It's an afternoon. Take the last thirty decisions of any size: what you built, what you didn't build, who you said yes to. Write the pattern down as though it were a strategy somebody else had handed you. Every one of those yeses was reasonable on the day, and together they add up to something you can read in a sentence. When Meridian's founder did this, the sentence he ended up with was we build for whoever is in front of us, and we're best at whichever thing we shipped last. He read it off his own decisions.

The page that replaces the mission statement

Once you can read your pattern in a sentence, you can write a better one deliberately, and the place most companies put that sentence is a mission statement. A mission statement says what you'd like to be true about your company. Nobody can act on it and nobody can violate it, which is how it survives unchanged for a decade on a wall. Under the Roger Martin video there's a comment from somebody who has clearly spent years in these sessions, calling the mission statement a futile ceremony. That's harsh, and I understand where it comes from. What we're writing is the opposite kind of document, one you could break by Thursday.

It has three things on it.

The first is the focus: who you're for. Meridian's is mid-size logistics operators running twenty to two hundred vehicles. That took the founder about ten minutes, because he already knew it. The focus is rarely the hard part.

The second is the insight, and it's the one that gets left out. A focus with no reason underneath it is just a preference, and preferences lose arguments. The first time somebody more senior wants something else, or a customer with real money wants something else, there's nothing there to hold on to. The insight is whatever you understand about those customers that a bigger competitor doesn't. For Meridian it's this: a twenty-vehicle operator has no IT department, so what wins there is a product that works on the first day, with no project attached to it. The enterprise vendors can't do that, because their whole business model assumes a six-week implementation and somebody on the customer's side to run it. That's a real insight. It might also be wrong, and that's what makes it a strategy. It's the thing you'd have to change your mind about if the market told you otherwise.

The third is the one I couldn't find in a single source I read. You have to name what you're not going to do, and not in general terms either. Specifically, by name, in a sentence somebody could hold you to. Meridian's version reads: we don't build custom integrations for enterprise warehouse systems. That sentence has a name. It's a decline.

It also has a price. That sentence costs the ninety thousand dollars from year two, and the deal from year three, and whatever the next one would have been worth. It's revenue Meridian could book this quarter, and the page says they're not taking it. That's why it's the hardest line in the document, and why it's missing from every strategy document you've ever been shown.

What I can't tell you

I can't tell you that writing that sentence down makes a company more successful. I went looking. There's no study. Nobody has measured companies with written declines against companies without them. And the practitioners this series leans on sell transformation work for a living: Cagan's group does, Castenfors does, and so does the consultancy behind the JP Morgan numbers I used last time.

What I can tell you is mechanical. You can settle an argument with a document that forbids something. That is a fact about the document.

My own third line

I have one of these pages myself. I'm a co-founder of a waste-to-energy company, and ours has the same three lines on it. The third one is the one that costs.

The focus is small refineries. Artisanal, I call them, which is a joke, but the small part isn't. They take waste in and make one of two things: drop-in diesel that meets the ASTM standard, which means it goes straight into an existing engine with nothing modified, or synthetic aviation fuel. We sell it to offtake suppliers, a deliberately small number of customers who take the fuel from us in volume. That's a choice as much as the refineries are, because the alternative is a distribution business, and working out how to get fuel to pumps would eat the company alive long before the refineries ever got good.

The insight underneath all of it is about feedstock, which is what goes into the refinery. Ours comes mostly from cattle farms, and agricultural waste is predictable. It's the same material week after week, which means you can build a refinery tuned to exactly that, instead of one that has to cope with anything.

So the third line says we don't use municipal mixed waste as our feedstock. That means household trash, everything that goes in a trash can. It can be done. Other companies do it, and it is an enormous operational challenge, because what turns up is different every single day.

That sentence turns down more than it looks like it does. If the largest landfill owners in the country came to us and said dig this trash up and turn it into fuel, the answer is no. If a country came to us, Singapore say, and asked us to take its citizens' waste and make energy out of it, the answer is still no. That's the biggest customer either of us could imagine, and the page says we don't do that. Not because it can't be done, but because our refineries are only good when we know what is going into them, and mixed waste is the one feedstock nobody can know in advance.

Why one page and not the whole method

That page of mine is one page for the same reason Meridian's is, and one page leaves a great deal out. Martin's actual method is bigger than what I've described. It's five choices that all have to reinforce each other, running from what winning looks like down to the management systems that keep the whole thing alive. It's good, and I'd use it.

That method also assumes you've got management systems. At forty people you have a founder, a spreadsheet and a group chat. One page is the part of Martin's method that still works when there's no apparatus underneath it, which is a different thing from a beginner's version of it.

One sentence, this week

Open a document, type we do not, and finish it. If you can't finish it, that's your finding.

Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote Founders Who Finish and publishes The Build. More about Dave →