Build a 90-Day Marketing Plan Your Team Can Actually Run

You have the findings. Maybe they came from an audit, maybe from your own review of the accounts. There are more problems on the list than your team can fix at once, and some of them depend on each other. The decision in front of you is what to do first, what can run in parallel, who owns each piece, and how you will know in 90 days whether to keep going, change course or stop.

Most 90-day plans fail for one of two reasons. They are a calendar of activities with no link to a business outcome, or they schedule work before the things it depends on are ready, such as launching new campaigns before tracking can measure them. This guide builds the plan around dependencies and decision points instead of dates.

This is a planning guide for paid acquisition work after a diagnosis. It is not a new-job 30-60-90 plan, and it is not a generic startup template.

Write the business constraint in one sentence

Before listing tasks, write down the one thing the next 90 days must change. Examples: “We cannot tell which campaigns produce booked consults.” “Cost per qualified lead has doubled and we do not know why.” “Paid traffic converts on the main site but not on the new landing pages.” If you have three sentences, pick the one that blocks the other two.

Choose an outcome and define qualification

Pick one number the plan will report against, and define it precisely. “Leads” is not a definition. A qualified lead, a booked consult, an enrolled family or a first order is. Write down:

• What qualifies (for example, a new caller or form submitter who books an appointment).
• What does not (existing customers, wrong numbers, spam, out-of-area requests).
• Which system counts it: the CRM, booking system or point-of-sale system, not the ad platform alone.

Record the baseline and what remains unknown

Write down the current value of that number and the period it covers. Then write down what you cannot see. If calls are not tracked, the baseline undercounts. If the CRM does not record source, you cannot yet tie outcomes to campaigns. Unknowns are part of the baseline. They tell you which work has to happen before you can judge anything else, and they warn you that a jump in the number after tracking is fixed may be new measurement rather than new demand.

Turn findings into dependencies

Take every finding and ask: what has to be true before this fix can be made or measured? The answers sort your list into an order.

Measurement and access before confident budget decisions

You cannot judge a budget change if the conversion data is wrong. That makes measurement repairs (conversion tracking, call tracking, deduplication, CRM source fields) the first dependency for most plans. Access is the dependency before that: the team needs working access to ad accounts, analytics, tag manager, the CRM or booking system, and the pages that receive traffic. Late access is the most common reason a plan slips in its first two weeks.

Creative, page and approval dependencies

New creative needs a brief, a concept approval, production time and sometimes a shoot with talent, locations and usage rights. New pages need copy, design, development, tracking and approval. In regulated categories such as healthcare, legal and clinical review sits in the critical path for both. Write down how long each of these takes in your organization, not in theory. If approvals take two weeks, the plan needs two weeks.

Dependency map (hypothetical example)

Read it as a chain: access first, then tracking and briefs together, then production and account cleanup, then tests, then a decision. Anything that can run in parallel should; anything that cannot should not be scheduled as if it could.

Build the first work cycle

The first cycle, often the first three to five weeks, has four jobs:

1. Repair priority defects. Fix tracking, broken forms, unanswered call routing, obvious wasted spend (for example, search terms with no commercial intent). These are the changes most likely to matter and least likely to need a test.
2. Gather evidence. Once tracking is reliable, let it collect data. Note the date measurement changed so later comparisons account for it.
3. Approve concepts. Get creative concepts and page outlines through approval while tracking is being fixed, so production can start as soon as it is ready.
4. Establish reporting. Agree the report format, cadence and owner now. A useful monthly report shows spend, cost per result by campaign and creative, calls and forms separately, the agreed number, and a short note on what changed and why.

Avoid launching major new campaigns during this cycle unless the tracking they depend on is already working.

Run the next tests, not every possible test

The findings will suggest more tests than your budget and production capacity can support. Pick the two or three with the largest expected effect on the agreed number, and the clearest way to read the result.

Assign budget and production capacity

For each test, write down the budget it needs, the creative or pages it needs, and who produces them. If production can deliver a fixed number of new assets per month, the test plan has to fit inside that number. Spreading budget across too many tests makes each one too small to read.

Budget figures in the plan are assumptions, not forecasts. Write them as ranges with the reasoning behind them (for example, “assumes current cost per click holds within a range and the new page converts no worse than the old one”).

Define an inconclusive result and a stopping rule

Before launching each test, write down three things:

• What counts as a win: the change in the agreed number that would justify rolling it out.
• What counts as inconclusive: too few outcomes to tell, or a difference too small to act on. Decide in advance what you will do (extend, redesign, or drop it).
• When to stop early: a spend ceiling or a clear failure signal, such as a page that breaks tracking or a creative that drives volume with no qualified outcomes.

Low-volume accounts need this most. If a campaign produces a handful of qualified outcomes a month, a 90-day window may not produce enough to separate a real difference from noise. In that case, test bigger changes, judge on leading indicators you have defined in advance, and say plainly that the result is directional.

Plan the review and next allocation decision

The 90-day review is a decision meeting, not a status update. It should answer:

• Is the measurement now reliable enough to trust? If not, that is still the priority.
• What did each test show, and which results are conclusive?
• Where should the next 90 days of budget go?
• Which assumptions in the plan were wrong?

Sales-cycle lag matters here. If a lead takes 30 to 60 days to become a customer in your business, outcomes from campaigns launched in month two may not arrive until after the review. Report by lead date and by outcome date, and do not treat a missing outcome as a failure until enough time has passed. Ninety days is a planning cycle, not a guarantee that the evidence will be complete.

The one-page plan

Every item fits on one line. If it does not, it is not specific enough.

One-page planning worksheet (hypothetical example: not an actual client plan, no results implied)

The business in this example is fictional: a multi-location service business whose main constraint is “we cannot tell which campaigns produce booked appointments.”

Next to the plan, keep a short decision log:

Decision log (hypothetical example)

The log matters more than it looks. It shows why the plan changed, which is the best defense against a review that judges week-one assumptions with week-twelve hindsight.

Decide who executes it

A good plan names an owner for every line. Then decide how each owner gets the work done:

• Internal team. Best when the skills exist in-house and the team has capacity. Usually the cheapest option and the one that keeps knowledge inside the company.
• Freelancers. Good for bounded tasks such as a page build, an edit batch or a tracking fix, if someone internal can brief and check the work.
• A specialist agency. Fits when one discipline (search, social, tracking or creative) is the real gap and the rest is handled.
• An integrated agency. Fits when page, creative and media problems are tangled together and coordinating several vendors has become the bottleneck.
• Mixed delivery. Often the realistic answer: internal leadership and approvals, a specialist for media, freelancers for production.

Whoever executes, the plan should be yours. Elite Créatif’s own audit, for example, describes its 90-day plan as “yours to run with any team” (Marketing Audit). That principle applies to any plan: if it only works with the people who wrote it, it is a proposal, not a plan.

If you already have a sound plan with owners, dependencies and decision gates, you probably do not need another audit. Start executing and use the review to decide what comes next. If you need the diagnosis before the plan, review what the Marketing Audit includes.