011 min
What breaks without Next Action
A task sits on a list for weeks and nobody can say what would happen next. The list says "organize company offsite" or "fix onboarding flow." Both describe a finished state, not a move a person can make today. Faced with a sentence like that, the brain has to do two jobs before any work starts: decide what the very next physical step is, and then do it. Most people stall at the first job and call the stall procrastination. An open-ended item like this also feeds the planning fallacy, because there is no concrete step to estimate against, only a vague outcome.
David Allen's Getting Things Done names this failure precisely: the task was captured, but never turned into an action. An open project sits as a vague outcome, generating a background sense that something is unresolved and no obvious place to pick it up. In triage, the same gap shows up in review meetings. A backlog item reads as a topic, not a step, so every conversation about it turns into a debate about direction rather than a five-minute task somebody could take on. The cost is not the missing hour of work. It is the repeated decision cost paid every time the item is looked at again, still unclear, still undone.
022 min
How Next Action works
Allen's own definition sets the test:
The next action is the next physical, visible activity that needs to be engaged in, in order to move the current reality of this thing toward completion.
Two words carry the whole mechanic: physical and visible.
Physical means a next action is something a body can do: type, call, walk, write, click, ask. "Think about," "plan," and "figure out" fail this test. They describe mental work with no observable end point, so a person can sit with them for an hour and never know whether the step is finished.
Visible means someone watching would be able to tell the step happened. "Decide on a vendor" is invisible. A decision can be made silently, inside a person's head. Nobody else can confirm it occurred, not even the same person a week later. "Email the three vendors with the shortlist criteria and a reply deadline" is visible: the email exists, timestamped, or it does not.
| Fails the test | Passes the test | |
|---|---|---|
| Physical | "Think about the offsite" | "List three venues in the shared doc" |
| Visible | "Decide on a vendor" | "Send the shortlist email with a reply deadline" |
| Neither | "Fix onboarding" | "Open a ticket describing the drop-off step in the signup funnel" |
The mechanic runs on a single recurring question: what is the very next physical, visible thing that has to happen for this to move? Applied once, it turns "prepare for the review" into "gather last quarter's commits." Applied again to that answer, it can turn into something even more concrete, such as running a specific command or opening a specific file. The questioning stops the moment the answer is something a person could do in the next few minutes without deciding anything else first.
The test also draws a line the word "task" does not draw on its own. A task list can mix outcomes ("launch the feature"), decisions ("pick a name"), and genuine next actions ("file the DNS change ticket") in the same list. Nothing about the wording alone says which is which. Sorting a list against the physical-and-visible test is what turns a list of intentions into a list of things a person can actually start.
031 min
Where Next Action came from
David Allen spent the 1980s consulting on productivity with businesses, including a stint consulting at Lockheed's human resources department, before he wrote the practice down as a system. He published Getting Things Done: The Art of Stress-Free Productivity in 2001, and released a revised edition in 2015 to reflect changes in information technology during the preceding decade.
The next action sits inside the workflow's "clarify" step, one of five steps Allen names in the book, where a captured item gets turned into either "not actionable" or a decision about what its very next physical step is. Allen's own claim for the practice is direct: he calls the physical-and-visible test "perhaps the most fundamental practice of this methodology." That claim is specific to the clarify step inside a five-step workflow, not a general productivity slogan detached from the rest of the system.
042 min
Real-world examples
Getting Things Done is a checkable case on its own terms. The David Allen Company runs a certification academy and licenses Global Partners worldwide to teach Getting Things Done courses and coaching. The book itself has been published in more than 28 languages since 2001. None of that proves the physical-and-visible test works. It does show the test is not a private habit — it is documented practice that a licensed industry has built itself around.
Independent of GTD, psychologist Peter Gollwitzer's research on implementation intentions backs the same mechanism from a different direction. Gollwitzer's 1999 paper in American Psychologist studied plans of the form "whenever situation x arises, I will do y," and found that people who formed one were more likely to actually act on the linked goal than people who only intended to. In Gollwitzer's own words, "implementation intentions delegate the control of goal-directed responses to anticipated situational cues, which (when actually encountered) elicit these responses automatically." A next action applies the same mechanism at the level of one task: naming the physical, visible step in advance removes the moment where a person would otherwise have to decide what to do, which is the moment procrastination usually happens.
An illustrative case shows the shift in practice. A PM opens a backlog item that reads "improve checkout." It is not physical or visible — nobody could watch someone "improve" something. Turning it into a next action means asking what the very first move is: read the last ten support tickets that mention checkout, and list the three most common complaints in the ticket tracker. That sentence can be watched happening. The PM, or anyone else on the team, can tell whether it got done.
052 min
How to run Next Action this week
Pick one list already in use: a backlog, a personal to-do list, or a project's task board. Run this pass over it.
- Read each item and ask: is this already a next action, or does it describe an outcome? Redesign the settings page is an outcome. Sketch three settings-page layouts in Figma is a next action.
- For every outcome, ask what the very next physical, visible step toward it is. Write that step with a verb a body can perform: open, write, call, send, file, ask.
- Check the step against both halves of the test. If it fails "physical" (it is really a mental state like "think about" or "figure out"), break it down one level further. If it fails "visible" (it is really a private decision like "decide"), rewrite it as the visible act that follows the decision: the email that announces it, or the ticket that records it.
- If the step takes less than two minutes, do it immediately instead of writing it down. This is Allen's own two-minute rule, meant to keep the list from filling with trivia.
- Sort what is left by where the action happens: a calls list, an at-my-desk list, and an in-this-doc list. This is what GTD calls organizing by context, so a person looking at the list while on a call only sees calls.
- Park anything that turns out not to be a real next step for now (GTD's someday-maybe list), so the working list stays short enough to actually scan.
The output is a list where every remaining line is something a person could do in the next few minutes without a further decision. That is the artifact: not a plan, and not a prioritized ranking, just a list of concrete starts. Triage runs a version of the same pass under time pressure, using a different question: not physical and visible, but what needs doing before anything else can. The underlying move is the same: turn a topic into a step.
062 min
Next Action done badly
The most common failure is writing a smaller outcome instead of a real action. Set up the meeting looks concrete next to improve the process. It still hides a decision: who to invite, what time, and what the meeting is even for. The tell is that two people reading the same line would take different first steps. A real next action leaves only one thing to do.
A second failure runs the other way: breaking work down so far that the list turns into hundreds of tiny steps, and maintaining the list becomes its own project. GTD's two-minute rule exists to catch this — a step that small should be done, not filed.
A third failure is writing the action but leaving the context out. The step is physical and visible in theory but unusable in practice: call the vendor with no phone number, no vendor name recorded elsewhere, and no indication of what to say. The step passes the test on paper and still cannot be started.
A fourth is letting a vague item sit unclarified because clarifying it feels like the actual work. A backlog full of items like "look into performance" behaves the way Parkinson's Law predicts: with no concrete first step and no real constraint on scope, each item can absorb as much time as is left before the deadline, because nothing about the wording limits what "look into" would take.
071 min
Next Action vs. nearby concepts
A to-do item is any entry on any list, written at any level of clarity. "Call mom," "redesign the app," and "buy milk" are all to-do items. A next action is a to-do item that has passed the physical-and-visible test: it names one observable move, not an outcome or a category of work. Every next action is a to-do item. Most to-do items are not yet next actions.
Most Important Task answers a different question. It asks which one thing, among several valid next actions, matters most to do today. Next Action asks whether a given item is actionable at all. One is a priority ranking. The other is a clarity check. The two work in sequence, not in competition: a list has to be made of real next actions before ranking which one is the most important task is even a meaningful exercise. Ranking three outcomes by importance just produces a ranked list of things nobody can start.
A backlog item that has been through forcing function treatment is related but distinct again: a forcing function makes an action unavoidable by removing the alternative, where a next action just makes the very next step legible. Writing a next action down and putting it somewhere visible can work as a mild forcing function, but the concept itself is about clarity, not compulsion.
?5 questions
