Open any productivity blog and you’ll find the same diagram: a neat weekly calendar, every hour assigned a colour-coded block, work flowing without friction from one task to the next. It looks satisfying, and it looks nothing like any real workday I’ve ever had.

Calendar blocking is genuinely useful. But the version most people try (rigid, hour-by-hour, no slack) collapses at the first disruption. One meeting runs twenty minutes over and the whole afternoon is thrown off. The blocks sit there, mocking you, while you scramble to handle whatever just came in.

This post is about a more honest version: a blocking approach that bends when the day goes sideways instead of snapping.

Why the rigid version fails

The classic advice treats your calendar like a flight schedule. Block 9-11 for deep work. Block 11-12 for email. Block 1-3 for the report. It assumes your energy is flat across those hours, that meetings end on time, and that nothing unexpected arrives. None of those assumptions hold reliably.

Research on cognitive performance is pretty clear: our capacity for focused work fluctuates across the day, and it varies between people and between days. Some mornings you’re sharp by eight. Some mornings you’re not sharp until ten, or not at all. A block you put on the calendar last Sunday has no way of knowing that. (There’s a longer look at the science in The science of deep work if you want to go deeper on this.)

Add in the meeting problem. A 2023 Reclaim AI report found the average knowledge worker has eleven meetings per week, and late starts and overruns are common enough to be the norm rather than the exception. If your blocking system assumes meetings end exactly when scheduled, it’s already fragile before you even start.

The two reasons any given plan breaks

Most plan failures fall into one of two categories.

The first is external: a meeting runs long, someone pings you with an urgent ask, a fire needs putting out. These aren’t failures of willpower. They’re the normal texture of collaborative work. Planning as if they won’t happen is the mistake.

The second is internal: your energy doesn’t match what the block expects of you. You scheduled deep analysis work at 2pm, but you’re in a post-lunch dip and can only manage shallow tasks. Pushing through often produces worse work than waiting an hour. But a rigid block gives you no room to adapt.

A good blocking system has to account for both.

The anchor-block method

The core idea is to protect one or two non-negotiable deep-work blocks per day and keep everything else flexible.

An anchor block is a slot you treat as genuinely protected. You don’t schedule meetings over it, you don’t let it slide to accommodate requests, and you defend it the same way you’d defend a standing client call. One anchor block is usually enough for most people. Two is realistic if your work heavily demands focused thinking.

What makes anchor blocks different from regular time blocks is the selection criteria. You pick the anchor based on two things: when your cognitive energy tends to peak, and when external interruptions are least likely. For many people that’s the first ninety minutes of the workday, before Slack fills up. For others it’s after lunch, or in the late afternoon. The point is you know your own pattern better than any template does.

Everything else in the day (the emails, the admin, the lighter tasks, even the secondary deep-work sessions) goes into flexible slots. You plan them loosely. If something runs over or falls away, the flexible slots absorb the impact. The anchor doesn’t move.

This isn’t a new idea. It’s closer to how good project managers think about scheduling: protect the critical path, give everything else some float.

Buffers and fallback rules

Buffers are fifteen-to-thirty-minute gaps you leave between blocks on purpose. They exist to absorb overruns, to let you decompress between context switches, and to give you a moment to re-orient before the next task.

Most people resist buffers because they feel like wasted time. They’re not. A thirty-minute buffer that catches a meeting overrun saves you from an hour of fragmented afternoon work. The math works out.

Alongside buffers, it helps to have simple fallback rules for common disruptions. Not elaborate contingency plans, just quick if-this-then-that decisions you make in advance so you don’t have to make them under pressure.

A few that work well:

• If the morning anchor block gets cancelled or cut short, take the first available flexible slot of ninety minutes or more and use it as a replacement anchor. Don’t try to squeeze deep work into twenty-minute gaps.
• If an urgent ask arrives during a flexible slot, handle it. If it arrives during the anchor block, log it and return to it after.
• If your energy is clearly low at the start of a scheduled deep-work block, spend fifteen minutes on a lighter task first rather than forcing it. A brief warm-up often brings the focus along.

These rules work because they’re decided ahead of time. In the moment, a disrupted day produces decision fatigue. Pre-made rules short-circuit that.

Re-planning mid-day without guilt

When the morning doesn’t go according to plan (and some mornings it won’t) the useful move is a short re-plan around midday rather than abandoning the day entirely or rigidly sticking to a broken schedule.

This doesn’t have to take long. Five minutes, sometimes less. Look at what actually happened in the morning, what’s left, and what realistically fits in the afternoon given your current energy and remaining obligations. Then re-block.

The trap to avoid is treating the deviation as a moral failure. It’s not. It’s information. A meeting ran over: that’s a data point about how meetings behave in your organisation. You were too tired to do focused work at 9am: that’s a data point about your energy pattern that day.

Over time, mid-day re-plans become faster because you start to notice patterns. Which meetings reliably run over. Which mornings you’re sharper. Where the flexible slots actually end up going most often. That accumulated knowledge is more useful than any productivity template.

The time blocking vs. to-do lists post goes into more detail on how to decide what actually belongs on the calendar versus staying on a task list. Worth reading alongside this one.

When the system adapts itself

The manual version of all this (protecting anchor blocks, maintaining buffers, re-planning mid-day) works. It’s also effort. You’re doing the re-scheduling in your head, making judgment calls about where to move things, deciding on the fly what still fits.

Flow does this automatically. When a meeting runs over, it reflows your remaining blocks to fit the actual shape of your day. When your energy signal suggests you’re in a dip, it surfaces lighter tasks rather than pushing you into deep work you’re not ready for. The logic is the same as the anchor-block method: protect the high-value focused time, absorb disruptions in the flexible slots. You just don’t have to manage it manually.

There’s more on how the calendar integration handles meeting overruns specifically in Why your calendar should adapt when meetings overrun.

If you want to try Flow in early access, join the list. It’s pre-launch and I’m building it in the open. Feedback from early signups directly shapes what gets built next.