How I run my week

My week runs on five mechanisms: one single focus chosen in advance, a fixed daily execution block, a friction log I fill as I go, batched sessions for similar work, and a 30-minute weekly review that sets up the next cycle. Nothing in it requires unusual discipline, special software, or getting up at five. It is the operational layer of the personal operating system — the essay explains why the pieces exist; this one shows them running.

I am writing this partly because "systemised week" conjures images of colour-coded calendars and monastic routines, and the reality is much duller. Dull is the point. A week that needs enthusiasm to run is a week that fails several times a quarter.

One focus, chosen before the week starts

Every week has exactly one focus — a single project that counts as the week's work. It is chosen at the previous week's review, not on Monday morning, because Monday morning is the worst possible moment to make that decision: the inbox is loudest and judgement is coldest the further you are from the moment of execution.

Everything else — client delivery, admin, calls — still happens. But the focus is what moves first each day, and on Friday the week is judged by one question only: did the focus move? A week where twenty things progressed slightly and the focus did not is, by definition, a failed week. That scoring rule sounds harsh and is the single most clarifying thing I do.

The daily execution block

The first working hours of my day are a fixed execution block, identical every day, with a set order: most important work first, most avoided work second. The avoided item is placed deliberately — avoidance is the most reliable signal I have of where the real difficulty lives, and difficulty deferred compounds like interest.

The rules of the block are few and absolute: no inbox before it, no meetings inside it, and the focus project always leads. When the block is fixed, then starting costs no decision — I sit down and the calendar has already chosen. This matters more than any productivity technique I have tried, because the expensive part of hard work is not doing it; it is the daily renegotiation about whether to begin.

The friction log

Throughout the day I keep a running file open, and anything that drags, breaks, or gets dodged goes in as one line, at the moment it happens. "Spent 30 minutes finding last quarter's figures." "Avoided the pricing email again." "Call overran because no agenda."

No analysis, no fixing, no tidy formatting — capture only. The log is a sensor, not a journal. Its entire job is to carry accurate evidence to Friday, because by Friday, memory has already rewritten the week into something more flattering. The log is where decision rules come from: a friction that shows up three times is a missing rule announcing itself.

Batching similar work

Similar work is gathered into single sessions rather than scattered across the week. Writing happens in one block, not five fragments. Filming happens in one afternoon — setup and teardown cost the same whether I record one video or four, so recording four is close to free. Calls cluster on set days, leaving other days unbroken.

The logic is switching cost. Every change of mode — writing to calls, calls to editing — burns time and sharpness on re-entry. Commonly the re-entry tax on fragmented work is larger than the work itself. Batching is just refusing to pay the same toll five times.

The 30-minute weekly review

Friday afternoon, thirty minutes, fixed order:

1. Numbers first. Pipeline inflow, proposals out, revenue booked — a handful of figures, looked at every single week. Numbers go first because they are the least negotiable part of the review, and because looking at inflow weekly is how you catch problems months before revenue reports them. This is a feedback loop with the delay shortened, applied to one person.
2. Friction second. Read the log. Pick the most expensive recurring item. Write or amend one rule that would have prevented it. One — not a reform programme.
3. Next week's focus last. With the numbers and the friction fresh, choose the one project that most deserves next week. Write it down. The review ends there.

Thirty minutes is enough because the log and the numbers did the gathering all week. Reviews that take two hours are usually doing archaeology that capture-as-you-go would have made unnecessary.

What deliberately is not scheduled

Roughly a third of my week has nothing in it. This is not slack in the pejorative sense; it is the load-bearing part of the design. Client emergencies, overruns, the task that turns out to be triple its estimate — all of that lands in the unscheduled space instead of landing on the execution block.

A fully scheduled week is a system with no buffer, and a system with no buffer transmits every shock straight to its most important components. When the buffer absorbs the surprise, then the block survives; when there is no buffer, the block is always the first thing sacrificed. Founders who schedule 100% of their hours are not more productive — they are one surprise away from a fortnight of catch-up.

What it adds up to

Nothing above is clever. One focus, one block, one log, batched work, one short review, and empty space on purpose. The system's only real virtue is that it runs on bad weeks — tired weeks, interrupted weeks, weeks where motivation never showed up. Those are the weeks that decide a year, and they are exactly the weeks a willpower-based routine forfeits.

---

Keep the thinking coming: The Format — a fortnightly letter on business and personal systems. Subscribe by email.