You Can't Plan a Quarter When the Tools Change Every Month

I built a sensible quarterly plan in April, and by the end of the month a chunk of it was quietly obsolete, not because we were wrong about the goal but because the tooling had moved underneath it. Two of the hard problems on the plan had become roughly trivial, and one of the easy ones had turned into a question mark.

I have made a lot of plans in my career. I had never watched one go stale that fast, for that reason. I have even watched a model release change what was hard overnight in the middle of a piece of work: we scoped a GPT-4 to GPT-4o migration on a live platform as a careful project, and capability that was expensive when we wrote the plan was cheaper by the time we executed it.

It made me rethink what planning even is when the ground moves under it every few weeks. The plan was not badly made. It was made for a world that stopped existing shortly after I drew it.

Planning has always been a bet against uncertainty, but the uncertainty used to sit in familiar places. Would the estimate hold, would the client change their mind, would something break. You could plan around those because they were the known weather.

What is new is uncertainty about your own capability. You genuinely do not know what your team will be able to do cheaply in eight weeks, because the tools that determine it are changing faster than the plan is meant to last. You are planning a quarter on top of a capability floor that keeps shifting, and no amount of detail in the plan makes the floor hold still.

The old instinct is to plan harder: more detail, more certainty, tighter commitments. That instinct is now actively wrong. A detailed twelve-week plan is a confident statement about a world you cannot see, and its main effect is to make you slow to react when the world turns out different. The more precisely you planned, the more you have invested in a picture a single release can outdate.

So what I have shifted toward is planning the direction firmly and the route loosely. The goal for the quarter, the thing we are actually trying to make true by the end, I still commit to hard, because it does not depend on which tools exist.

But the specific path, which features, in which order, built which way, I now hold deliberately loose, because locking it in early buys precision I cannot honour and cannot use. The plan names the destination and stays honest that the road will be redrawn more than once on the way. That is not a weaker plan. It is a more truthful one.

In practice, holding the route loose means committing to fewer things further out and more things close in. The next few weeks I plan in real detail, because the ground under them is roughly stable. The far end of the quarter I describe as an intention, not a schedule, and I say so plainly. It is not vagueness. It is matching the precision of the plan to the visibility I actually have.

It also means the artefact with the most detail is no longer the one doing the most work. A tidy Gantt chart full of confident boxes looks like planning and is increasingly the opposite: a monument to certainty you do not have. The real planning is in knowing which parts to commit to and which to leave open on purpose.

The harder change is cultural, and it is the part I am still working on. Clients and stakeholders take comfort in detailed plans, and telling them here is the goal, and we will decide the route as we learn what the tools can do can sound like you have not done your homework.

So part of the job now is teaching people to take confidence from the direction rather than the detail, which runs against a decade of everyone being trained to trust the plan with the most boxes on it. I have started being explicit about it: here is what will not change, here is what I am deliberately leaving open, and here is why leaving it open is the responsible choice, not the lazy one.

None of this means planning less. If anything it means thinking harder, because holding a firm direction while keeping the route genuinely open is more demanding than filling in a spreadsheet of certainties. It just means the confidence has to come from somewhere more honest than a picture that will not survive the month.

The reward for getting this right is that you stop being surprised. A plan built on false certainty breaks loudly and late. A plan built on a firm direction and an open route just absorbs the next change and keeps pointing the same way. It bends instead of snapping, and that turns out to be worth far more than the comfort of a detailed chart.

So the question I keep asking myself, and I will put it to you: are your plans still a source of confidence, or have they quietly become a source of drag, a detailed picture of a world that stopped existing a month after you drew it?