A 90-day plan is one page holding three to five priorities for the next quarter, each broken down into actions that go straight into this week's diary, with a checkpoint once a month to catch drift early. It takes about two hours to build and about thirty minutes a month to maintain. That is the whole method, and it works because ninety days is short enough that you still remember what you decided and long enough to finish something that matters.
The reason to build one is simple: the annual plan is not failing because it was badly written. It is failing because nothing in it tells you what to do on a Tuesday morning. A quarter is the shortest planning horizon that still allows for real work, and the longest one an owner can hold in mind without a document. Below is how to build the plan step by step, what to put on the page, and the three ways it usually falls over.
Why does ninety days work when the annual plan does not?
Three things change when you shorten the horizon. Urgency becomes real — a deadline twelve weeks out is close enough to feel, and far enough away that you cannot fix it in the last fortnight. Assumptions get tested sooner, so a plan built on a wrong number gets corrected in the quarter rather than in the following December. And the review actually happens, because reviewing twelve weeks of work is a job you can do in an hour.
Done four times a year, the compounding matters more than the plan. Four honest reviews and four resets produce more movement than one ambitious document, because each cycle starts from what really happened rather than from what you hoped would happen. This is the same rhythm we use in business coaching, and it is the reason progress gets measured rather than felt.
Step one: what actually happened last quarter?
Before you plan forwards, close the loop backwards. Write down what you said you would do last quarter, then what happened, then why. Not a paragraph — a line each. Most owners find one of four causes behind anything that did not move: it never got into the diary, it depended on someone who was never told, it was blocked by something you did not control, or it was never really a priority and you knew it at the time.
This step feels optional. It is the one that makes the next plan better, because your planning improves only when you find out how your own estimates behave. If you consistently plan four times more than you deliver, that is not a character flaw — it is a number, and now you can plan against it.
Step two: which three to five priorities go on the page?
A priority is something that, done well, changes the business. Not a task, not a wish, and not routine work you would be doing anyway. Three is better than five. The plan has to constrain as much as it commits, because the value is in what you have agreed not to do this quarter.
Test each candidate with three questions. Would finishing it move a number you care about — margin, hours, cash, capacity? Can it be finished inside twelve weeks, or at least reach a defined milestone? And does it need you, or is it on the list because nobody else has been asked? Anything that fails the first question is activity. Anything that fails the third belongs to somebody else with a deadline.
Step three: how do you turn a priority into a weekly action?
Every priority needs at least one action that could be done in the next five working days, with a named owner and a date. If you cannot name that action, the priority is not a plan — it is an intention, and intentions do not survive a busy month.
Work backwards. Twelve weeks, then the milestone at week six, then what has to be true by week two, then what happens this week. Put the recurring time in the calendar as a real appointment rather than a hope: not "work on pricing" but Tuesday 9 to 11, repeating. Protected time is the mechanism; everything else is intention.
An invented illustration of the shape — not a client, and not a claimed result. An owner sets a quarterly priority of "improve margin". Useless as written. Worked back, it becomes: reprice the twelve lowest-margin jobs, at an average uplift of £180 a job, worth roughly £2,160 a month if they all hold. Week two: pull the job list and cost the bottom twelve properly. Week four: agree new prices and write the standard quote. Weeks five to ten: two repricing conversations a week. Week eleven: count how many held. Same ambition, but now it is six conversations a fortnight in a diary, and by week five you already know whether it is working.
Step four: what happens at the monthly checkpoint?
Thirty minutes, once a month, on the same date. Four questions: what moved, what did not, what is now in the way, and what changes for the next four weeks. Write the answers down. The checkpoint exists to catch drift while it is still cheap — a priority that has not moved in four weeks is recoverable, and one that has not moved in twelve is a wasted quarter.
Be careful with one thing. Dropping a priority because the facts changed is good management. Dropping it because it turned out to be uncomfortable is how plans quietly die. The test is whether something changed outside your own appetite for it.
What does the finished plan look like?
- One page. If it needs two, you have too many priorities.
- Three to five priorities, each written as an outcome with a number and a date.
- One measure per priority that you can check weekly without building a report.
- The first action for each, with an owner and a date inside the next five days.
- The three checkpoint dates and the end-of-quarter review date, already in the calendar.
- A short list of what you have deliberately decided not to do this quarter.
Keep it visible. A plan in a folder you open once a quarter has no effect on behaviour; the same page on the wall or pinned in your task list gets read weekly by accident, which is the point.
What usually goes wrong?
Three failures account for most of it. Too many priorities, so nothing gets the attention that would have finished it. No weekly action, so the plan never touches the diary. And no review, so the same quarter repeats with new wording. If your plan is going wrong, it is almost certainly one of those three rather than something exotic.
There is a fourth that owners rarely name: the plan assumes a version of the week you do not actually have. If the diary is already full of delivery, a quarter of ambitious priorities is arithmetic that does not work. That is worth facing directly — either something comes out of your week, or the plan is fiction. Our guide on building a business that runs without you deals with the capacity side, and the Planning Hierarchy model shows how the quarterly plan is supposed to hang off the longer-term one. If the problem is that the goal itself never turns into activity, start with goals that create momentum instead.
Get the next one in your inbox
One practical, plain-English guide for business owners each week. No spam, unsubscribe any time.
Frequently asked questions
How long should building a 90-day plan take?
About two hours for the first one, and closer to ninety minutes after that. Half of it is the honest review of the quarter just finished, and half is choosing priorities and working them back into weekly actions. Do it in one sitting, away from the office, before the quarter starts rather than a fortnight into it. If it is taking a full day, you are almost certainly trying to plan the whole year in quarterly clothing, or you are writing a task list instead of choosing priorities. Maintenance afterwards is thirty minutes a month at the checkpoint, plus about ten minutes a week to check the measures.
What if the quarter goes off course by week four?
That is the checkpoint doing its job, and it is far better than finding out in week twelve. Take the four questions in order: what moved, what did not, what is in the way now, and what changes for the next four weeks. Then decide honestly whether the facts changed or your appetite did. If a supplier fell over or a big client left, rebuild the plan on the new facts. If nothing changed except that the work was harder than expected, keep the priority and fix the execution, usually by protecting time you have been giving away. A quarter half delivered and reviewed properly still beats a quarter abandoned in silence.
Should the team have their own 90-day plans?
Yes, once yours exists and they can see how theirs connects to it. Each person's plan should be expressed as things they control rather than results they only influence: on-time completion, quotes turned round inside two days, a documented process finished. Keep them to two or three priorities, on the same quarterly dates as yours, reviewed in the same monthly rhythm. The point is not paperwork; it is that everyone can answer what they are trying to finish by the end of the quarter and what good looks like. If two people give different answers about the same priority, you have found the gap the plan was meant to close.
Do I still need an annual plan as well?
Yes, but keep it short and treat it as direction rather than instruction. One page describing where the business is heading over the next one to three years, the numbers that matter and the two or three big things that have to change. The 90-day plan is where that direction becomes action, and each quarter should visibly serve the longer-term picture. The failure is when the annual plan is thirty pages of detail nobody reads, and the quarter is invented from scratch every twelve weeks with no connection to it. Short at the top, specific at the quarter, weekly where the work happens.
The same priority has rolled over three quarters running. What now?
Stop rolling it. A priority that has survived three quarters without moving is not a priority, and carrying it forward a fourth time only teaches everyone that the plan is decorative. Decide which of three things it actually is. If it genuinely matters, it goes to the top of this quarter with protected diary time and the first action booked inside five days, and something else comes off the page to make room. If it matters but not to you, hand it over with a named owner, a budget and a date. If neither is true, delete it and say so out loud at the review. The relief of formally dropping something is usually the sign it should have gone two quarters ago.