You Don't Choose a Direction Once

When I was around ten, a Scout leader drove two of us to the edge of a forest, handed us a map, a compass, a canteen, and a granola bar, and pointed at a spot a few miles off. Find your way there. It took six hours.

The lesson wasn't how to read a compass. It was why you have to keep reading it. You'd sight a heading, pick a distant tree that lined up with it, and start walking – and then the trees got in the way. You'd step around a thicket, drop into a gully, climb back out, and the tree you'd been aiming at was gone. Every obstacle nudged you a few degrees off, and a few degrees, compounded over a mile, puts you somewhere you never meant to be. The only way through was to stop, over and over, and check the compass again.

I think about that walk whenever someone asks me what they should build next, as though the answer is a thing you decide once and then execute. It isn't. Direction is the compass, not the tree.

Why "decide once" quietly fails

Most teams do set direction once. There's an offsite, a strategy deck, a roadmap that runs twelve months out, and then everyone puts their heads down to execute it. The problem isn't the plan. It's that the terrain moves, and the plan doesn't.

ProductPlan's 2026 State of Product Management survey found that teams rate the alignment between their roadmap and their strategy at just 3.6 out of 5 on average, and only 61% rate it a 4 or better. Nearly two in five teams know their plan and their direction have drifted apart. And the reasons are almost entirely internal, not market surprises: 49% point to resource and capacity constraints, 47% to shifting short-term priorities, and 44% to leadership changing direction. Those are the thickets and gullies. Each one nudges the team a few degrees off the heading, and because nobody stops to re-check the compass, the roadmap slowly stops being a decision filter and becomes a diplomatic document – a record of who argued loudest, not a bet on where the market is going.

Building faster points you the wrong way faster

The obvious hope is that modern tools fix this by making the team quicker. They don't, because speed was never the thing that was broken. Setu Shah, who runs global product strategy at Oracle, put it well in that same report: "Execution got faster, but thinking became the real differentiator. The teams that win aren't the ones that ship the most – they're the ones that choose deliberately."

That matches what the adoption data has said for years: roughly 80% of software features are rarely or never used. When you can build five times as fast, an unchecked heading doesn't get you to the wrong place more efficiently – it gets you further into the wrong forest before anyone notices. Acceleration raises the cost of a stale direction. It doesn't lower it.

Checking the compass, in practice

Re-checking direction isn't re-running the twelve-week strategy project every quarter, and it isn't a 200-page document that's obsolete the moment the market does something it didn't predict. As one founder-operator put it, "you never stop going to market" – a new feature, a new segment, a competitor's move, each one is its own small heading-check, not a detour from the real plan.

What you want is a direction you can interrogate. Something you built on evidence and can revisit when the evidence changes – answer the hard question once, then keep answering it as the world moves, rather than treating the first answer as permanent because re-opening it feels like admitting the first one was wrong. It wasn't wrong. It was true for the forest you were standing in, and you've since walked into a different one.

The teams that hold their heading aren't the ones with the most detailed map. They're the ones who kept stopping to check where they actually were. Direction isn't a decision you make. It's one you keep making, on purpose, before the drift adds up to somewhere you never meant to go.