team
The offsite produced four priorities for the quarter. Somebody found an action plan template with twelve columns, filled in the first three rows properly, and the rest of the grid holds fragments. By week three the document has not been opened. The work has not stopped, but what people are doing now bears only a loose relationship to what was agreed, and nobody can point to the moment where the two separated.
Downloading a different template will not change this outcome. Most published action plan templates are competent. What they cannot supply is the part that makes a plan work, which is a named moment when somebody reads it back and asks what happened. A template is worth using because it forces a small number of decisions to be made explicitly. Choosing one is mostly about which decisions those are.
An action plan sits between a goal and a schedule, and it gets confused with both.
A goal document says what should be true at the end. Increase repeat purchase rate. Pass the audit. Open the second location. It contains no owners and no dates, and it is not a plan.
A project plan describes work that has been scoped, with dependencies, durations and usually a timeline view. It assumes the approach is settled.
An action plan is the step in between. It names the specific actions that will move the goal, who owns each one, when each is due, and how anyone will know it worked. It is short by design, usually one page, and it exists for work where the approach is not yet a project. That is why the same structure appears in community organising guides, in corrective action plans after an incident, and in thirty, sixty and ninety day plans for someone new in a role. The context differs and the skeleton does not.
The distinction matters because it sets the level of detail. An action plan with forty rows has become a bad project plan. An action plan with three rows and no dates is a goal document with extra formatting.
Most templates offer between eight and fourteen columns. Seven fields do nearly all of the work, and two of the seven are routinely left out.
| Field | What it forces | Common failure |
|---|---|---|
| Goal statement | One sentence everyone agrees on | Two goals joined by an and |
| Action | A specific thing someone does | Written as a theme, not an action |
| Owner | One person who moves it | A department, or two names |
| Due date | The action competes with dated work | Every row due at quarter end |
| What it needs | Money, access, a decision, someone's time | Left blank until the week it blocks |
| Measure of done | Anyone can judge it finished | Confused with the goal's metric |
| Status | Stale rows become visible | Ten statuses, so none are updated |
The two that get skipped are what it needs and measure of done, and they are the two that decide whether the plan survives. An action with no stated requirement is an action that will stall silently, because the requirement, usually a decision from somebody outside the room or a budget nobody has approved, only becomes visible when the deadline arrives.
Measure of done is worth separating from the goal's own metric. The goal might be measured in repeat purchase rate. The action, rewrite the post purchase email sequence, is done when the new sequence is live for all customers. Mixing the two produces actions that can never be closed, because the metric has not moved yet.
Three statuses are enough: open, done, dropped. Dropped is the one most templates omit, and its absence is why old plans fill with rows nobody intends to do. A plan that is mostly noise stops being read, which returns the whole exercise to where it started.
Rather than collecting templates, it helps to recognise which of three situations is at hand, because each needs a different shape.
| Shape | Situation | What makes it different |
|---|---|---|
| Single page plan for one goal | A quarter priority, a launch, a campaign | Grouped under the goal, with a measure for each action |
| Time phased plan | Someone new in a role, or a staged rollout | Actions grouped by period rather than by theme, first period much more detailed |
| Corrective plan | After an incident, a complaint, or a failed audit | Root cause recorded next to each action, and a verification step after the fix |
The time phased shape, commonly written as thirty, sixty and ninety days, has one rule that decides whether it is useful. The first period is specific and the later ones are deliberately coarse, because detail written for day seventy will be wrong. Plans that give all three periods equal detail are a sign that nobody expects to follow them.
The corrective shape carries an extra column that the others do not need. After each fix there has to be a check that the fix held, with its own date and owner. Without it, the plan records intentions and cannot demonstrate anything to an auditor or a customer.
Four rules cover almost all of the difference between a plan that moves and one that does not.
Start with a verb and name the object. Improve onboarding is a theme. Rewrite the three welcome emails and send them for approval is an action. The first cannot be finished, and therefore will not be.
Give every row a different date. When every action is due at the end of the quarter, everything arrives in the last week and half of it slips. Spreading the dates is also the only honest way to find out whether the plan fits the available time.
Check the plan against one person's capacity, not the team's. Four actions owned by the same person in the same fortnight is the most common reason a plan fails, and it is visible in the document before anyone starts. Reading the owner column and counting is a thirty second check that prevents a month of drift.
Name the first action for each goal explicitly. Not the most important one, the first one. Plans stall at the start more often than in the middle, and the first action is where the sequencing errors show up, usually as a dependency on a decision somebody has not made.
Plans get into trouble when the goal's number and the actions' progress are treated as the same measurement. They move on different clocks.
The goal has one number, checked monthly at most. Repeat purchase rate, days to close the books, number of complaints. It responds slowly, it is affected by things outside the plan, and reading it weekly produces noise that leads to changing the plan for no reason.
The actions have a count, checked weekly. How many are open, how many are done, how many have been dropped. This says whether the plan is being executed, which is a different question from whether it is working.
Keeping them separate makes one situation legible that otherwise causes arguments. When every action is done and the goal's number has not moved, the plan was wrong about cause rather than badly executed. That is genuinely useful information, and it is only available if the two measures were never mixed. The opposite case, a number that improved while half the actions were never started, is worth just as much attention, because something other than the plan is doing the work.
One habit to avoid: assigning each action a percentage of the goal. The weights are guesses, they make a spreadsheet look precise, and they encourage finishing the actions with the largest made up weight rather than the ones that matter.
This is the failure that matters, and it is structural rather than a matter of discipline.
The plan lives in a document. The work lives on a board, in a tracker, or in an inbox. During the hours when people decide what to do next, they are looking at the second thing. An action that exists only in the first is invisible at the moment it needs to be seen, and invisibility, rather than reluctance, is what kills most of them.
There are two workable responses, and they can be combined.
The first is to keep the plan as the source document and copy each action into the place where its owner already works, with a link back to the plan for context. The plan then holds the reasoning and the measures, and the board holds the actions. This works provided the plan is not edited afterwards, because two editable copies will disagree within a fortnight.
The second is to stop keeping the plan in a document at all. The goal becomes a board or a label, each action becomes a card with an owner and a date, the measure of done goes in the card, and the plan is read by opening the board. Whether this is comfortable depends on whether the tool can hold the fields a plan needs, including things like a measure and a requirement, and whether the same cards can be seen against dates rather than only as a list. Checking that a tool supports custom fields and more than one view of the same cards before moving a plan into it avoids discovering the gap halfway through the quarter.
Either way, one rule holds. The plan needs a named date and a named person for the review, written into the plan itself. A plan with no review date is a record of what a group of people believed on one afternoon.
Fifteen minutes a week is enough, and the agenda does not vary.
Read every open action aloud with its owner and date. Anything overdue gets one of three outcomes, decided in the meeting: a new date with a stated reason, a different owner, or dropped. The fourth option, leaving it overdue and moving on, is what turns a plan into a document nobody trusts.
Two patterns are worth watching. An action whose date has moved three times is not late, it is blocked on something the owner does not control or it is not going to happen, and saying so is more useful than a fourth date. And an owner with many open actions has a capacity problem the plan has made visible, which is one of the better reasons to keep owners as individuals rather than departments.
Once a month the plan itself is worth questioning rather than the rows in it. If the goal has changed, the actions underneath it are not worth completing, and finishing work that no longer serves the goal is a more expensive failure than dropping it.
Take the current plan, delete every column nobody has filled in twice, and add the two that are probably missing: what each action needs, and how anyone will know it is done. Then put a weekly fifteen minute review in the calendar with a name against it, and move each action to where its owner already works so it is visible on a normal day. If the actions, their dates and the conversation about them should sit in one place rather than three, Pinateca is free for up to five people and ten boards.
Seven fields cover nearly everything: the goal in one sentence, the action written as a verb and an object, one owner, a due date, what the action needs from outside, how anyone will know it is done, and a status of open, done or dropped. The last two are the ones most templates leave out, and they are the reason plans stall silently. Extra columns tend to be half filled, and a half filled plan stops being trusted.
An action plan names the actions that will move a goal, with owners and dates, and is usually one page. A project plan describes work that has already been scoped, with dependencies and durations, and assumes the approach is settled. An action plan that has grown to forty rows has effectively become a project plan and should be treated as one.
Because the plan lives in a document and the work lives somewhere else. During the hours when people choose what to do next they are looking at a board or an inbox, so an action that exists only in the document is invisible at the moment it matters. Fixing this means either copying each action to where its owner already works or keeping the plan in that place from the start.
Weekly for the rows, monthly for the plan itself. The weekly review reads every open action aloud with its owner and date, and each overdue item gets a new date with a reason, a new owner, or is dropped. The monthly check asks whether the goal still holds, since completing actions that no longer serve the goal is a more expensive mistake than dropping them.
It can be written that way, and it will behave as though nobody owns it. One name means one person responsible for moving it, which does not stop several people doing the work. If the right owner is genuinely unclear, that is the thing to resolve before the plan is finished rather than leaving the column blank.