In 1991, three friends and I started a software company in Oulu. We had one plan, and we all knew it by heart. We were going to build software products. Write once, sell many times. We built them. A CAD program with a graphical user interface for Windows 2, the first CAD program for Windows on the world market. AutoCAD moved to Windows years later. A system for analyzing and documenting groundwater research. A converter that took 2D geometry from DXF files and turned it into CL data, the cutter location instructions a cutting robot actually follows.
They were good products, written by people who loved writing them. There was no money behind them. Finnish high tech had no investors to speak of in those years, so nobody would fund the marketing a product needs to scale. I paid my own salary by fixing online lottery machines at weekends.
Then the floor gave way. Between 1990 and 1993, Finland lost more than a tenth of its GDP. Unemployment went from about three percent to almost seventeen. Banks failed, and customers stopped buying anything they did not strictly need.
Then Rautaruukki New Technology called. They needed a large software suite for laser measurement systems in steel plants, built by contractors. They paid by the hour, and on time. We said yes. Of course we said yes. That project never ended. We wrote the suite, and then we maintained it for ten years. By the end, it was 800,000 lines of C++, every line written by hand, decades before any AI agent could help.
Ten years. 800,000 lines for one customer. We had engineered a whole product, and it belonged to someone else. Rautaruukki was only the first. Orders soon came in from many other large companies, and each brought a software problem too hard to solve in-house. Hard problems became our competitive edge.
The edge was also the trap. We drifted because we were good at what we did. Every success pulled us further from the products. It took me years to see what came next. We never held a meeting to decide to stop being a product company. Nobody wrote it down, and nobody voted on it. If you had asked any of the four of us at almost any point in those years, we would have told you, sincerely, that we were a product company doing projects for a while. Our code said otherwise.
In 2020, the US Army University Press published a short book called On Strategy: A Primer. In one chapter, J. P. Clark and Francis Park discuss the people who turn strategy into action within large organizations. They borrow a term from the German psychologist Dietrich Dörner: repair service behavior. Every fix seems sensible and urgent, so you fix whatever is in front of you while the real problem goes untouched.
Their warning is a single sentence: “the false comfort of activity seemingly sanctified as purposeful by process will yield a product, but it might not be what is required.”
I read that sentence and recognized my first company. I never described it that well then.
Every project was a good decision. It paid salaries. It kept the team together through Finland’s worst postwar recession. I would make the same call again. What I would not repeat is leaving it unnamed. Now take the idea out of the army headquarters and put it inside a company of four friends. A strategy lives in what people actually spend their week on. When the plan and the week drift apart, people stop reading it. The team stops reading it first, and the founders stop last.
Clark and Park say a good planner must know when and how to step outside the process. I would add one thing. When you leave your strategy, leave it on purpose. Say it out loud. “For the next eighteen months we are a service company, because we have to survive. Here is the signal that tells us to go back.”
A pivot with a reason and an exit condition is a strategy. Drift is just weather. Left unnamed, a temporary decision becomes permanent by default. Good people build careers around it, customers build expectations around it, and the cost base quietly reshapes itself to fit. Every year of maintenance makes the next year more natural. The thing you meant to come back to is still there. You just never find a quarter in which coming back makes sense.
Thirty-five years later, I am leading the same journey in the opposite direction, taking Ginolis from project business to product business. This time the decision has a name, a date, and a plan the board can read.
Here is a test you can run this week. Open your calendar for the last four weeks and mark every block in one of two colors. Use one color for work that builds the strategy you wrote down, and the other for work that pays for today. Then ask your three best people to do the same.
If the colors don't match the plan, you are executing a strategy nobody chose.
I still think about that Windows CAD program. It never found its customers because we never had the money to find them. Then the projects came, and nobody stood up to say what we had become.
A strategy nobody cancels can still die. It just dies without a decision.
Kauko Väinämö, CEO, Ginolis.