A 90-day marketing plan that fits a small team

Marketing plans usually fail for one of two reasons. They assume resources that do not exist — a plan requiring twenty hours a week from a person with four. Or they start with production before anything is decided, so three months of content lands on a site that cannot convert it and against queries nobody chose. This plan is built to avoid both.

It assumes one or two people, four to six hours a week, and no budget beyond tools you already pay for.

Before day one: two decisions

Neither takes long, and skipping them makes everything after arbitrary.

1. The outcome number. One: orders, qualified leads, trials, bookings. Everything in the plan either moves it or gets cut.
2. The two channels. Chosen by where your customers already spend attention, not by where you are most comfortable. Two — a third is added in month four at the earliest.

Month 1 — fix and measure

No new content this month, which is the part everyone skips and the reason months two and three normally underperform. You are removing the leaks and putting measurement in place so the rest of the plan can be judged.

Week 1: measurement

• Analytics and Search Console installed, verified, and actually reporting.
• Conversion tracked as an event — not "visits", the thing you count as success.
• A "how did you hear about us" field on your form or checkout. Self-reported attribution catches what analytics cannot.
• Write down today's numbers. Without a baseline nothing that follows is measurable.

Week 2: the technical leaks

Run the indexability layer of a technical audit — noindex tags, canonicals, redirects, orphan pages, whether your content exists in the raw HTML. Fix what it finds. This is usually a day of work and it is the highest-return day in the quarter, because everything else is downstream of pages being indexable.

Week 3: the pages that already exist

• The page your traffic lands on most: does it state what you do, who for, and what to do next, above the fold?
• Your top ten commercial pages: unique titles, real copy, and answers to the questions people ask before buying.
• Category or service pages: most are an empty grid, and they match how people search.
• Anything that ranks 8–20 already: those are edits with the fastest possible payoff.

Week 4: decide the targets

Build the target list — twenty to thirty queries you can plausibly win, sourced from Search Console, your support inbox, sales calls and the method for small sites. Group them into pages. Order them by winnability, not by volume. This list is the input to month two, and month two is wasted without it.

Month 2 — produce, in batches

Now you create, against a decided list, at a cadence you can hold on a bad week.

• One substantial page a week against the target list. Four in the month, each answering one query completely.
• Three social posts a week on your two channels, from recurring formats, batched fortnightly.
• One lead magnet, if email is a channel for you — one page, narrow enough to qualify.
• Internal links added as you go: every new page links to two existing ones, and one existing page links to it.

Two rules make this survivable. Batch by task — write four captions in one sitting, not one on four days. And keep a week of finished work in the bank, so the week everything goes wrong does not become the week the plan stopped.

Month 3 — judge, then adjust

Keep producing at the same cadence, and spend one session on an honest review.

1. Query coverage: how many queries do you now appear for at all, versus month one? This moves first and predicts everything else.
2. Average position on the target set. Movement from 40 to 18 is real progress even though it produced no clicks yet.
3. The month-one fixes: did the pages you edited move? These should have, and if nothing did, re-check indexability before doing anything else.
4. Channel by channel: which of the two produced anything traceable? Reallocate, do not add.
5. The outcome number, against what you wrote down in week one.

Then decide exactly one change for the next quarter. One. More than that and the following review cannot attribute anything — which is the difference between measuring and explaining.

What this looks like weekly

• Two hours: the week's substantial page, or the fortnightly content batch.
• One hour: social scheduling and visuals, batched.
• Thirty minutes: replies, comments, messages. Daily in small pieces, not weekly in one.
• Thirty minutes: one internal-link pass and one small fix from the audit list.
• Monthly, thirty minutes: the review, against a written expectation.

Four to five hours. That is the honest cost of a marketing operation that compounds, and the reason plans requiring twenty hours produce nothing: they are not executed at all.

What to expect, honestly

• Month 1: no traffic change. You fixed leaks and built a baseline. This is the month people quit.
• Month 2: impressions start rising on the pages you fixed. Almost no clicks yet.
• Month 3: the month-one edits move positions; the month-two pages are still settling. First traceable outcomes, usually small.
• Months 4–6: the compounding starts. Pages published in month two reach stable positions and internal links start distributing authority.
• On a new domain, add two to three months to all of the above. It is slower and it still works.

The plan's only real requirement is that it survives contact with a bad month. That is why the cadence is set at what you could hold during your busiest week, and why month one exists at all.

Frequently asked questions