Getting Things Done

The problem it solves

Most people run their commitments from memory for longer than memory can cope with, and the arrangement tends to hold until the number of them passes a threshold that passes unnoticed. After that things get dropped, and the ones that get dropped are rarely the ones anybody would have chosen. They're whichever happened to be out of mind at the moment they needed doing.

The less obvious cost is paid even when nothing is dropped. A commitment that exists purely in your head tends to present itself at the wrong moments, in the middle of other work or at three in the morning, and rarely at the moment you're in a position to act on it. A good share of the unease people describe as being busy is this, a background process of remembering that runs whether or not anything gets done.

Lists are the obvious answer and most people already keep some. The trouble is usually that the lists don't hold everything, so the head keeps a copy as a backup, and a list you don't fully trust saves you none of the effort it was supposed to.

The idea

Getting Things Done is David Allen's method for moving commitments out of your head and into a system you trust enough to stop remembering on its behalf. Its starting point is the open loop, meaning anything you've agreed to do or be responsible for (including agreements made only with yourself) that isn't finished yet. An open loop keeps drawing on attention until it's either done or recorded somewhere you'll reliably look. That second half is better supported than it sounds, since in experiments on unfinished goals, making a specific plan for one removed the intrusive thoughts and the drag on other work that it otherwise caused.

Allen organises the method in five stages. The first edition called them collect, process, organise, review and do, and the names below are the ones he uses now.

Enlarge: Clarifying each item is the step that turns a pile into a system.

Five stages in a row, with a weekly review arc from reflect back to organise. Below clarify, a spine of questions. Not actionable: bin it, or park it as reference or someday. More than one step: a project. Two minutes or less: do it now. Someone else's: delegate, and track it as waiting for. The rest is deferred, to the calendar or the action list.

Clarifying each item is the step that turns a pile into a system.Close

Clarifying each item is the step that turns a pile into a system.

Capture means collecting everything that has your attention into a small number of inboxes, paper or digital, rather than leaving it in your head or scattered across sticky notes, chat threads and half-remembered conversations. It's deliberately indiscriminate, because stopping to decide what something means at the moment it arrives is what makes people skip capturing it.

Clarify is where the value is, and it's the stage most people skip. For each item, decide whether it's actionable at all. If it isn't, it gets binned, filed as reference or parked on a someday list. If it is, decide what the very next physical action would be, at the level of asking finance for last quarter's figures rather than sorting out the report. Anything that needs more than one action to finish is a project, and it's recorded as one alongside the next action that moves it forward. Anything that takes less than two minutes gets done there and then, because tracking it would cost more than doing it. What's left is either handed to somebody else and tracked as waiting for, or deferred, onto the calendar if it has to happen on a particular day and onto an action list if it doesn't.

Organise puts each clarified item where it belongs. Next actions go on lists, often grouped by where or with what they can be done. Projects go on a project list, things you're waiting on from somebody else go on a waiting-for list, and ideas you might pursue go on a someday/maybe list. The calendar is kept for things that have to happen on a particular day, and reference material is filed somewhere separate from all of these.

Reflect is mainly the weekly review, and it's what makes the rest trustworthy. Once a week you empty the inboxes, go through each list, check that each project has a next action, chase what you're waiting for and look at what's coming. Without it the lists drift out of date within a few weeks, the head quietly goes back to keeping its backup copy, and the system becomes one more place where things are written down and not looked at.

Engage is doing the work, choosing from the next-action lists according to where you are, how much time and energy you have and what matters most. It's the stage the method says least about, which matters, and it comes up again below.

The distinction between a project and a next action carries more weight than it looks. Most stalled to-do items are projects written down as though they were actions, such as the offsite, the promotion case or a parent's birthday, and they stall because nobody can do a project. You can only do the next action on one, and until that action has a name the item sits on the list producing guilt rather than progress.

How to use it

Start with a single sweep. Go through your head, your inboxes, your notes and your desk and write each commitment down in one place, large and small, work and otherwise. Most people are surprised by how long the list is, and the relief tends to arrive before anything on it has been done, which is the effect the method is built around.

Then clarify the whole pile in one sitting, asking of each item whether it's actionable and, if so, what the next physical action is. Be strict about the wording, since an action you can't picture yourself doing (look into the budget, think about the offsite) will be skipped each time the list is read. Do the two-minute ones as you go.

Keep the organising simple. A handful of lists in whatever tool you already use is enough to start with, and the calendar stays reserved for things that really have to happen on a given day, because a calendar padded with intentions stops being believed as quickly as a stale list does.

Book the weekly review as a recurring appointment with yourself, an hour at the end of the week, and protect it more than anything else in the system. It's the step people drop first and the one whose absence brings the rest down.

Where it breaks down

Every concept here has one. It is the section most summaries leave out.

It handles inputs and says little about what matters. The method is thorough at making sure nothing is lost and much weaker at telling you which of a long list of next actions deserves today. Allen does write about longer horizons of goals and purpose, though the day-to-day system assumes priorities arrive from somewhere else, and a well-run system can process the wrong work very efficiently. It pairs naturally with something that ranks, such as the Eisenhower matrix or a daily set of most important tasks.

The upkeep is real, and the weekly review is where most people give up. The setup takes a day or two, and after that the system runs on a weekly hour that's easiest to skip when the week has been bad, which is when it's needed most. Skip a few and the lists fall out of date, and a list you don't trust is worse than no list, since it adds a place to check without removing the need to remember.

It was designed around one kind of job. The method fits a knowledge worker handling a steady stream of discrete requests, messages, meetings and small commitments. It fits less well where the work is a few deep projects that each need long stretches of thought, since breaking a hard design problem into next actions tends to produce a list of trivial steps around a large blank that no list can hold.

It treats an organisational problem as a personal one. A good system helps one person cope with a flood of messages and requests, and it leaves the flood where it was. Where the volume comes from the way a team or company allocates work, a better inbox for each person is a workaround, and the people best at the method can end up absorbing more of the disorder than anybody else.

People over-invest in the tooling. Choosing apps, designing tag schemes and migrating between systems feels like progress and produces none, and a whole software market has good reason to encourage it. The habits carry the method, and a notebook is enough to run it.

The evidence is mostly for the mechanism rather than the method. The claim that unrecorded commitments cost attention, and that a specific plan relieves that cost, has reasonable experimental support. Whether people who adopt the whole system become more effective has barely been tested (its most thorough academic defenders said as much in 2008), and the wider research on time-management practice finds clearer effects on wellbeing and feeling in control than on how well people perform.

In one line

Write down each commitment, decide its very next physical action and review the whole lot weekly, so your head can stop being the place you keep the list.