Almost every annual marketing plan contains the same eleven sections. Almost none of them get built in the right order — and order, not content, is what decides whether the plan survives past March.
You already know what goes in a marketing plan. Executive summary, situation analysis, personas, SMART goals, competitor review, channel mix, budget, timeline, KPIs. That list is not a secret and it has not changed in twenty years. If the sections were the hard part, everyone's plan would work.
What actually breaks plans is doing the pieces in the wrong sequence, on the wrong calendar, using a baseline nobody stress-tested. Targets get set before last year's numbers are properly closed. Budgets get requested before anyone decides what to stop doing. The plan lands in January as a 40-page document that gets read once and never reopened.
So this guide is not another list of sections. It's a calendar: what to decide in October, what to decide in November, what to lock in December, and what each phase depends on. If you're planning for a financial year that starts in January, you're at the front of that window right now.
Why the sequence matters more than the sections
Each phase of planning produces an input the next phase needs. Skip or reorder one and the downstream work is built on a guess.
You cannot set a credible target until you've closed and interrogated the baseline. You cannot size a budget until you know what you're stopping — otherwise you're pricing everything you did last year plus everything new, which is how budget requests get halved by finance without discussion. And you cannot design a review rhythm after the year has started, because by then the trigger conditions are being argued about while someone's channel is on the line.
| Phase | When | The one decision it must produce |
|---|---|---|
| 1. Close the baseline | Early–mid October | What actually drove revenue this year, stated with known uncertainty |
| 2. Decide the cuts | Late October | The list of things you are stopping, in writing |
| 3. Tier the budget | November | Committed, flexible and reserve spend, with a defensible case for each |
| 4. Set the mechanics | Early December | The reforecast trigger and who is allowed to pull it |
Everything else — personas, positioning, channel tactics, content themes — is real work, but it's work that slots into these phases rather than competing with them. If your positioning or your customer journey map genuinely changed this year, that belongs in Phase 1 as an input to the baseline, not as a separate December exercise.
Phase 1: Close the baseline before you set a single target
This is the phase most teams rush, and it's the one everything else rests on. Three weeks is not excessive.
The task is not "pull last year's numbers." It's to arrive at a defensible statement of what produced revenue, what merely correlated with it, and how confident you are in each. That distinction is the whole game. A channel that looks efficient because it takes last-click credit for demand created elsewhere will get more budget next year and quietly starve the thing that was actually working.
The 2026 baseline has a specific problem
There is a reason to be more careful this year than usual. For a lot of sites, the relationship between traffic and revenue moved during 2026 rather than holding steady — and a baseline built on a moving relationship extrapolates badly.
AI-generated answers absorbed a meaningful share of informational queries. That means a site can show flat or falling sessions alongside steady conversions, because what disappeared was the top-of-funnel research visit that was never going to convert this quarter anyway. It can also mean the opposite: fewer assisted touches, so the visits that remain convert at a rate that flatters the channel. We covered the mechanics in what AI Mode means for organic traffic and in why zero-click content is forcing a rethink of success metrics.
Either way, "organic sessions grew 18% this year, so let's target 20% next year" is now a sentence with no meaning attached to it. Segment before you multiply. Split organic by query intent, and split the whole baseline by surface rather than lumping everything into one channel line.
The same caution applies to attribution generally. Signal loss has been degrading channel-level attribution for a while, which is why marketing mix modelling came back into fashion and why attribution keeps getting harder to pin down. You do not need a perfect model to plan. You need to know which of your numbers you'd bet the budget on and which you wouldn't.
If you have never run an end-to-end review of where prospects actually drop out, October is the time. Our funnel audit framework is built for exactly this window, and it will surface problems that no amount of channel budgeting fixes.
ClickFunnels 2.0
Build high-converting sales funnels and landing pages without any coding skills
Best for: Funnel Building
Phase 2: Decide what you're stopping — and write it down
Late October. This is the shortest phase and the one that changes the plan most.
Most plans are additive. Last year's activity carries forward by default, and the planning conversation is about what to add on top. Nothing is ever explicitly killed, so the budget request grows every year and the team's attention thins out across a widening set of commitments.
The test of a real plan A plan that only says what you will do is a wish list. A plan becomes a decision the moment it also says, on the record, what you have chosen to stop.
The mechanic is simple and uncomfortable. List every recurring marketing activity — every channel, every content format, every tool, every retainer. Against each, answer one question: if this were not already running, would you start it in 2027 at the cost it carries? Anything that gets a no or a maybe goes on the stop list. Anything you can't cut for contractual reasons goes on a separate list with its exit date, because that's a Phase 3 problem.
Two warnings from experience. First, expect to defend this list more than any other part of the plan, because every line has someone attached to it. Second, don't confuse "underperforming" with "under-resourced" — a channel starved of budget and attention for two years will look identical to a channel that doesn't work, and cutting the wrong one is expensive. The distinction usually comes down to whether it ever got a fair test at a sensible budget, which is a question your competitor analysis can help answer: if competitors are getting results from a channel you abandoned, the channel is probably not the problem.
Phase 3: Build the budget in three tiers, not one number
November. The budget is where plans get their credibility, and the single most useful change is to stop presenting it as one allocation.
Split it three ways:
- Committed. Contracts, headcount, retainers, tooling — spend you cannot exit inside the year without a penalty. This tier is not a strategy decision; it's a fact you're reporting. Know the number precisely and know each exit date.
- Flexible. Channel budgets you can genuinely move between quarters. This is where the strategy lives, and it should be presented as a starting allocation rather than a fixed one.
- Reserve. Money deliberately left unallocated at planning time, for opportunities and problems you can't yet name.
The reserve tier is the one most teams skip, and skipping it has a predictable consequence: every mid-year opportunity has to be funded by cannibalising something that was working. A reserve turns "we can't do that until next year" into a decision rather than a constraint. Even a modest one changes the conversation.
On sizing the flexible tier, resist the urge to distribute evenly across channels for political peace. Concentration usually beats spread, and the case for concentration is easier to make in a planning meeting than mid-year. If you're deciding how aggressively to scale paid, our guide to setting and scaling ad budgets without losing efficiency covers where the efficiency curve typically breaks, and where marketing budgets shifted during 2026 is a useful sanity check against how the rest of the market moved.
One practical note on organic: it's the tier most often under-committed because its payback sits outside the planning year, which makes it easy to defer and expensive to restart. If organic search is meant to carry pipeline in 2028, the investment decision belongs in this plan, not the next one — and if you don't have the internal capacity to sustain it, that's a resourcing question worth settling now rather than in June. A retained SEO partner is one way to make that commitment survive a busy quarter.
Writing the version leadership will actually read
Your plan needs two documents, and most teams only write one.
The long version holds the analysis, the assumptions, the channel detail. Nobody outside marketing will read it, and that's fine — its job is to be the thing you check when someone asks "why did we assume that?"
The short version is one page and it contains only decisions: the targets, the stop list, the three budget tiers, the reforecast trigger. No situation analysis, no persona summaries, no channel tactics. If a finance director can read it in four minutes and repeat your three main bets back to you, it works.
The questions that page has to survive are narrower than marketers expect. What are you asking for, what does it buy, what happens if you get 80% of it, and how will we know by June whether it's working. That last one is really a question about your reporting, which is why the dashboard you build for leadership should be designed alongside the plan rather than after it. Targets you cannot show progress against monthly will lose their budget to targets you can.
Phase 4: Set the reforecast trigger before January
Early December, and it takes an afternoon.
Every plan says it will be "reviewed quarterly." Almost none define what a review is allowed to change or what conditions force one. The result is that reallocation happens when someone senior gets nervous — usually later than the data warranted, and aimed at whichever channel is least well defended in the room rather than whichever is actually failing.
Fix it by writing the trigger conditions in advance, while nobody's performance is on the line:
- A downside trigger. A specific, measurable condition that automatically opens a budget review — a channel missing its cost-per-acquisition target for two consecutive months, say, or pipeline coverage falling below a stated multiple of the quarterly goal.
- An upside trigger. The one everyone forgets. If something beats its target by a wide margin, that should also force a reallocation conversation, and the reserve tier is what lets you act on it.
- A decision owner. Who can actually move money when a trigger fires, and how much they can move without escalating. A trigger with no owner is a note, not a mechanism.
Set your targets so this machinery has something to bite on. Vague goals can't trigger anything, which is the practical argument for the discipline in setting marketing goals and KPIs that drive growth — a target you can't miss precisely is a target you can't act on.
What to leave out
Four things that make plans longer without making them better.
Anything you'd have done anyway. If a channel is running and nobody is questioning it, it belongs in the committed tier as a number, not as three paragraphs of justification.
Tactics at the campaign level. Q3's email subject-line approach is not an annual planning decision. Plans that specify tactics twelve months out either get ignored or, worse, get followed after the market has moved.
Trend commentary. A section on what's happening in the industry is genuinely useful — as a briefing document. Inside a plan it reads as padding, because it doesn't attach to a decision. If a trend matters, it should show up as a changed assumption or a budget line, not as prose.
New launches with no owner and no date. Every plan has one: a promising initiative nobody has actually been assigned. It either gets a name and a start date in Phase 3, or it goes in a parking-lot appendix where it can wait for the reserve tier.
If you're reading this in December
The full four-phase sequence assumes you start in early October. If you're later than that, don't compress every phase evenly — protect Phase 1 and cut elsewhere.
A workable two-week version: spend the first week on baseline and the stop list together, working only with the numbers you already trust rather than trying to rebuild attribution. Spend the second week on budget tiers and the reforecast trigger. Skip the long document entirely and write only the one-pager. You'll have a plan that's thinner on analysis but sound on decisions, which is the better trade if you're short on time.
What doesn't work is the opposite compression — a thorough situation analysis, a beautiful deck, and a budget assembled in the last two days. That's the version that gets approved and then quietly abandoned by February, and it's the more common failure of the two.
The short version
Close the baseline before setting targets, and treat 2026's traffic numbers as needing segmentation rather than extrapolation. Write down what you're stopping before you write down what you're spending. Split the budget into committed, flexible and reserve, and actually fund the reserve. Define what triggers a reforecast, and who can pull it, before the year starts. Write one page of decisions and keep the analysis somewhere else.
The sections were never the hard part. The order is.
Budget built, but not sure the channel split holds up?
We pressure-test the allocation against what your numbers actually support before you commit the year.
Review my channel split →Frequently asked questions
When should you start building next year's marketing plan?
Early October, if your financial year runs to December. The sequence matters more than the start date: you need roughly three weeks to close and interrogate your baseline, two weeks to decide what you are stopping, three weeks to build and defend the budget, and a final fortnight to set your review mechanics. Starting in November compresses the baseline work, which is the part that everything downstream depends on. Starting in August means planning against a Q3 that has not finished happening yet.
Why can't you just extrapolate this year's numbers into next year's targets?
Because for many sites the relationship between traffic and revenue changed during 2026 rather than staying stable. AI-generated answers absorbed a share of informational queries, so a flat or falling session count can sit alongside steady or improving conversions. Straight-line extrapolation from a contaminated baseline produces a target that is either unreachable or embarrassingly easy, and you will not know which until Q2. Segment the baseline by intent and by surface before you multiply anything.
How should you split a marketing budget between committed and flexible spend?
Split it into three tiers rather than one allocation. Committed spend covers contracts, headcount and retainers you cannot exit inside the year. Flexible spend covers channel budgets you can move quarterly. Reserve is money deliberately left unallocated at planning time. The exact percentages depend on how much of your spend is contractual, but a reserve of nothing is the common mistake: it leaves you funding every mid-year opportunity by cannibalising something that was working.
What is a reforecast trigger and why set one in advance?
A reforecast trigger is a threshold you define before the year starts that automatically opens a budget review — for example, a channel missing its cost-per-acquisition target for two consecutive months, or a target being beaten by a wide margin. Setting it in advance removes the politics from the decision. Without one, reallocation happens when someone senior gets nervous, which is usually later than the data warranted and aimed at whichever channel is least well defended.
How long should an annual marketing plan be?
Short enough that people reread it in March. In practice that means one page of decisions — targets, what you are stopping, budget tiers, review triggers — with the supporting analysis kept in a separate appendix nobody is required to read. Long plans are not more rigorous; they are usually a sign that the writer could not decide what mattered, so everything was included and the reader was left to work it out.
