The Task You've Avoided for Three Weeks Isn't a Priorities Problem

I’ve watched founders run a company competently for a quarter while carrying the same three-line item on their list from week to week. It’s never the biggest thing on the list. It’s usually the one that would take about 40 minutes.

The interesting part is what those tasks have in common, because it isn’t size, difficulty, or how important they are. It’s that each one makes you feel something specific when you look at it, and nobody has ever said what.

The dread arrives before the work

There’s one study worth knowing about here. Lyons and Beilock put people who were anxious about maths into an fMRI scanner and watched what happened when they were warned that a maths problem was coming. Regions of the brain associated with visceral threat and the experience of physical pain lit up during the warning. They did not light up while the maths was actually being done.

The dread is front-loaded. Anticipating the thing hurts, and doing the thing mostly doesn’t.

Two honest caveats before anyone builds a life around that. It was maths anxiety in a scanner, not your board update, and the leap from one to the other is made far more confidently in the popular retellings than in the paper. The commonly quoted figure that discomfort clears after about 20 minutes is a rule of thumb rather than a measured constant, so treat it as an order of magnitude: minutes, not the rest of the afternoon.

What survives both caveats is a prediction you can check against your own week, which is that starting is reliably worse than continuing. You already know this. Every time you’ve forced yourself into something dreaded and found yourself fine ten minutes later, you ran the experiment.

The consequence for how you plan your week is bigger than it looks. Any system that depends on you wanting to start is a system that fails, because the wanting arrives after the starting rather than before it. You’ve been waiting on a signal that only ever shows up once you no longer need it.

Your avoidance list is a risk register

Avoidance doesn’t spread evenly across a founder’s list. It clusters, and it clusters somewhere expensive.

Look at what actually sits there for a month. Telling an investor the number is under plan. The conversation with the hire who isn’t working out. Putting the prices up. Opening the due diligence folder. Reading the churn report properly rather than glancing at the top line. Every one of those carries a feeling, and every one of them is also among the highest-consequence items in the business.

So the list of things you’ve been avoiding is a reasonably accurate register of your real risks, compiled by your nervous system and read by nobody. Whenever I’ve gone through one of these with a founder, the item that’s been sitting there longest is the one with another human being on the other end of it.

That’s the part that makes this worth writing about rather than filing under self-help. The cost of the avoidance isn’t the 40 minutes. It’s that the three or four decisions with the most riding on them are the ones being deferred, quarter after quarter, by exactly the mechanism that’s supposed to protect you from harm.

Every fix you’ve tried aims at the wrong layer

Here’s the problem with almost every tool a founder reaches for: it assumes you don’t know what to do.

Prioritisation frameworks rank the list. Time blocking finds a slot for the top of it. A new app moves the whole thing somewhere better designed. All three are aimed at a decision you made weeks ago. You know what the task is, you know it matters, and you know roughly how long it takes. None of that has ever been in question, which is why none of those fixes hold for longer than about a fortnight.

The thing genuinely in the way has a name, and it’s usually something like “I don’t want to write down a number that will exist in a document forever” or “I don’t want to find out I’ve been wrong about this since March”. Until that gets said out loud, everything built on top of it is aimed one layer too high.

The four moves

Four things, in order. The order matters, because each one changes the answer to the next.

1. Name what you’re avoiding, not what you’re not doing. “I’m avoiding the board update” names the task. “I’m avoiding writing down that we’re 18% under plan in a document that gets read by six people” names the feeling. The second one can be worked with. The first one is just the list again. Don’t accept “no time” as an answer either, because you’ve found time for worse this week.

2. Make the entry point too small to refuse. Not a scaled-down version of the task. The first physical move: open the file, type the heading, nothing else. The useful test is whether your first action still contains a decision. If it does, cut it down again, because you’ll avoid the decision just as happily as you avoided the task.

3. Stop deciding when. Every time the start is a decision, you get a fresh opportunity to not make it, and you’ve been taking that opportunity for three weeks. So hang the work off something that already happens without you arranging it. The end of an existing weekly meeting. The first coffee. The walk back from the school run. A calendar reminder is not one of these, because you’ll learn to dismiss a reminder within about nine days.

4. Write the re-entry plan before you need it. You’ll miss a day. That’s fine and it was always going to happen. What kills the whole thing is that the miss gets read as proof the system doesn’t work for you, and by the following Tuesday there’s no system at all. So decide now what re-entry looks like, make that action smaller than the starting action rather than the same size, and ban catching up. Whatever you missed is gone.

When the answer is to delete the task

The uncomfortable version, and I’d rather say it than sell you a protocol that doesn’t apply.

Sometimes three weeks of avoidance is information rather than a fault. If nothing has broken, nobody has chased, and you genuinely cannot name what the delay is costing, the honest move is to take it off the list and stop paying rent on it. Carrying an item you’re never going to do has a real cost in attention, and it also makes every other thing on the list feel heavier.

The test I’d apply is whether the cost is invisible or absent. Those are different. An unchased invoice, a contract auto-renewing that nobody read, a conversation being delayed with someone who’s already decided to leave: all invisible, none absent. If you can name the cost, it stays and you run the four moves on it. If you can’t, delete it today and get the attention back.

I’ve written the four moves up as a skill file you can download. It runs the stages in order, and there’s a gate at the front for when you arrive carrying a list rather than a task, because a list is how the session turns into another hour of prioritising and no work at all. It’s in the Claude Skills Library as The Unstick Protocol, free, no email needed. Upload it once to your Claude project and it triggers when you say you’re stuck.

It won’t make you want to start. Wanting was never on offer. The first 20 minutes are, and that turns out to be the only part you actually had to solve.