Approach
When making things is cheap, teams finish work they would never have started when it was expensive. The polish is what makes it stick. A quarter of shipped, documented, demoed deliverables is harder to walk away from than a quarter of good intentions, and every sprint spent on the wrong thing becomes one more reason not to admit it was the wrong thing. I work at the point before that, where a team still has to decide what it wants to be true.
How I work
Most of what I do early is find the thing nobody in the room has said out loud. Usually it is an assumption everyone thinks is shared, and nobody has checked. Direction problems look like disagreement about what to build, and turn out to be disagreement about what the thing is for.
What to build is rarely the real question. Where a company is defensible, what it can charge for, which customers are worth the longer sales cycle: these decide what is worth building long before any of it reaches a screen. At Grönloop that meant arguing against building the product interface, because the value was in how data moved between organisations and none of that was proven yet. Advising against work is part of the work.
I teach the same thing in class. Designing a course is designing a system: what has to be understood first, what can be left out, what people should be able to do afterwards that they could not do before. I have built several of the courses I teach, and the hardest decision each time was what to leave out. Students who learn to ask this about a brief keep asking it about everything else.
The measure is what happens afterwards. I want a team that argues better without me in the room, with the reasoning written down somewhere it can be picked up and disagreed with. Anything that only works while I am there has not worked.
What I leave behind
What lasts after an engagement is rarely the thing that was asked for at the start. A deck gets read once and then quoted out of context.
A position, in writing
What the company is, what it is not, and what it is deliberately giving up in order to be that. Short enough to read in one sitting, specific enough that someone can disagree with a particular sentence instead of the general mood.
The criteria, not the conclusion
A decision outlives its usefulness faster than the rule that produced it. At Grönloop the useful part was not the recommendation to hold off on the interface, it was the three conditions under which building it becomes the right call. Those outlast me, and the team can apply them without asking.
Something that runs
Sometimes I build the thing instead of describing it. BTO is a live grading tool I built and shipped for a thesis committee at the Estonian Academy of Arts. A working version settles arguments that a document can only prolong.
Where this is going
The same thing that happened to building is happening to work. What is cheap to produce stops being what anyone gets paid for, and it keeps moving. In fourteen years the tools, the titles and the deliverables have turned over completely more than once, and each time the people who moved with it were the ones whose value was never in the tool.
What held was narrower than I expected. Judgment about what is worth doing. Understanding how a business actually makes money and where it is exposed. Being able to write something down clearly enough that someone can act on it or argue with it. Reading a room well enough to know why the stated objection is not the real one. None of those belong to an industry, which is why none of them expired.
None of this has been tidy, and I have stopped treating that as a problem to fix. The mess is not the stage before the plan arrives, it is the working condition. What I am building toward is the agency to keep deciding inside it rather than being assigned a place in it. Anyone who can tell you exactly what they will be doing in three years is describing a job, not a direction.
There is nothing new under the sun, but there are new suns.