I Spent a Month Treating Shallow Work as Real Work, and My Deep Work Improved

The vocabulary around concentration divides the day into deep work, which is where value comes from, and shallow work, which is overhead to be minimised, batched, and ideally delegated. I absorbed this framing and then noticed that in my job the shallow category includes the roster, the incident forms, the ordering, and the handover — which is to say, most of the things that actually keep a unit running and that people notice immediately when they are done badly.

So for a month I ran an experiment where I refused to call any of it shallow. Every one of those tasks got a named slot, a stated quality standard, and the same treatment as anything else. What I expected was that this would eat my concentrated time. What happened was the opposite, and I have spent a while trying to work out why.

The first mechanism is that admin done at a bad time takes longer. I had been doing forms at the end of a shift, on the theory that low-value work belongs in low-value time. At the end of a shift I am slow and I make errors, and errors in an incident form generate a follow-up conversation two days later. Moving the same forms to a mid-morning slot cut the time roughly in half and the follow-ups to near zero. The work was not shallow; my attention was. The related planning problem is making room for interruptions instead of treating every scheduled hour as available.

The second is that unfinished admin is not free while it waits. A roster I have not done occupies a small amount of attention continuously — not a lot, but it is present, and it is present during the exact stretches I was trying to protect. When I started finishing these things at a scheduled time, the protected stretches got quieter in a way I noticed within about a week. I do not have a number for this and I am wary of overclaiming, but the subjective difference was large enough that I have not gone back. A Friday review of commitments offers a separate checkpoint for work still left open.

The third mechanism is about resentment, and it is the least measurable and possibly the most important. When a category of your work is officially the lesser kind, doing it feels like a failure to be doing the real thing. That feeling is corrosive and it made me rush, and rushing produced the errors in the first mechanism. Once the roster was simply a task I do on Tuesday mornings, it stopped carrying that charge, and I stopped arriving at my afternoon already slightly annoyed.

I want to be careful not to overstate this into a claim that all work is equally valuable. It is not, and the underlying advice — that most people spend too much of their best hours on things that do not require them — is correct and I follow it. My deep and shallow slots are not interchangeable, and the roster does not get my best two hours.

The correction is narrower: the category “shallow” contains two very different things, and lumping them together caused the problem. There is work that is genuinely low-consequence and should be batched, minimised, and done at whatever time is left over. Email is mostly this. And there is work that is procedurally simple but high-consequence, which looks identical on a to-do list and behaves completely differently — it punishes tired attention, and it does not stop costing you until it is done.

Sorting my supposedly-shallow tasks into those two piles took twenty minutes. Nine were the first kind and seven were the second, and the seven had all been sitting in end-of-shift time for years because the vocabulary I had been using did not have a place for them.