An SEO plan is a dated, sequenced list of decisions: what you are going to rank for, which pages will do it, what has to be fixed before those pages can rank, and what you will measure to know whether it worked. Most documents people call an SEO plan template are just a list of tactics with no order and no owner, which is why they get abandoned in month two. The skeleton below is a real 90-day SEO planning template — five phases, each with an input, an output, and a completion test.
Copy the skeleton, replace the bracketed parts with your own site, and delete anything you genuinely cannot staff. A plan you finish beats a plan that covers everything.
The 90-day SEO plan skeleton
This is the copy-able part. Everything after it explains how to fill each phase in.
- Phase 1 — Audit (days 1 to 14). Input: your live site plus Search Console. Output: one ranked list of problems, each tagged blocking, damaging, or cosmetic. Done when every URL template on the site has been checked, not just the homepage.
- Phase 2 — Keyword map (days 8 to 21). Input: your product vocabulary, customer language, and Search Console query report. Output: a table of primary keyword, intent, target URL, and secondary phrases. Done when every keyword has exactly one owning URL and no URL owns two intents.
- Phase 3 — Technical fixes (days 15 to 45). Input: the blocking and damaging items from Phase 1. Output: crawlable, indexable, fast pages. Done when a fresh crawl returns zero blocking items.
- Phase 4 — Content calendar (days 21 to 90, continuing). Input: the keyword map. Output: dated briefs and published pages, pillar first, supporting posts after. Done when the calendar is populated at least six weeks ahead at all times.
- Phase 5 — Measurement loop (weekly from day 30). Input: Search Console, rank tracking, and analytics. Output: one weekly note recording what moved, and one monthly decision to continue, expand, or kill each cluster.
The order matters more than the contents. Publishing into a site that blocks crawlers, or writing before the keyword map exists, wastes the most expensive resource in the plan, which is writing time.
Phase 1: audit before anything else (days 1 to 14)
You cannot plan around problems you have not found. Two weeks is enough to check everything that matters on a site under a few thousand URLs.
Work through this list and write each finding into one document:
- Indexation. How many pages does Search Console report as indexed, and how does that compare to the number you intended to publish? A large gap in either direction is the single most valuable finding in an audit.
- Crawl blocks. robots.txt rules, noindex tags, canonical tags pointing somewhere unexpected, and pages reachable only through JavaScript or a search box.
- Duplicate and thin pages. Tag archives, paginated listings, near-identical location or product variants, and posts under a few hundred words that answer nothing.
- Titles and meta descriptions. Missing, duplicated, truncated, or brand-first when they should be keyword-first.
- Heading structure. Multiple H1s, skipped levels, or headings used purely for visual styling.
- Internal links. Which pages have almost no inbound internal links? Orphaned pages rarely rank regardless of quality.
- Speed and Core Web Vitals. Field data in Search Console beats lab scores; treat anything failing on mobile as a real problem.
- Existing performance. Which queries already produce impressions but sit in positions eight to twenty? Those are the fastest wins in the entire plan.
Tag every finding as blocking (the page cannot rank at all), damaging (the page ranks worse than it should), or cosmetic (nice to have). Cosmetic items do not enter the 90-day plan. A crawler-driven check such as our SEO score tool will surface most on-page and indexability items quickly, but the judgement calls about thin content and duplicated intent are yours.
The audit output that actually gets used
One document. Each row is: finding, URL or URL pattern affected, severity tag, estimated effort in hours, and owner. If a finding has no owner it is not in the plan, it is a wish.
Phase 2: build the keyword map (days 8 to 21)
The keyword map is the spine of the entire SEO plan. Everything downstream — briefs, calendar, internal links, reporting — reads from it.
Build it in this order:
- List your money terms. The three to ten phrases that describe what you actually sell. These usually belong on service, product, or pricing pages, not on the blog.
- List the problems you solve, in customer language. Not your marketing vocabulary. The words people use before they know your category exists.
- Multiply by modifiers. best, alternatives, vs, cost, pricing, for beginners, near me, template, checklist, examples, how to. This mechanically generates most of a mid-size keyword set.
- Validate demand and phrasing. Autocomplete, People Also Ask, and your own Search Console query report confirm that people phrase it the way you wrote it.
- Judge winnability. Read the top ten results for each shortlisted phrase. If every result is a major publisher with deep coverage, park it. The method is worked through in how to find low competition keywords.
- Assign one URL per keyword. This is the step everyone skips, and skipping it is what causes cannibalisation six months later.
Worked examples across every intent type are in SEO keyword examples, and the competitive side — finding what already ranks for people you compete with — is in competitor keyword analysis.
Keyword map columns
| Column | What goes in it |
|---|
| Primary keyword | One phrase, the exact wording searchers use |
| Intent | Informational, commercial, transactional, navigational |
| Target URL | The one page that owns this keyword |
| Page type | Pillar, supporting post, comparison, product, service |
| Secondary phrases | Two to five variants sharing the same intent |
| Status | Not started, briefed, drafted, published, updated |
Phase 3: technical fixes (days 15 to 45)
Fix in severity order, not in difficulty order. The temptation is always to clear the easy cosmetic items first because it feels productive.
Blocking items come first: anything that stops a page being crawled, indexed, or rendered. Accidental noindex tags, robots.txt disallow rules covering live sections, canonicals pointing at the wrong URL, and pages returning soft 404s all belong here. A single wrong line in robots.txt can neutralise an entire quarter of content work.
Damaging items come second: duplicate titles, missing descriptions, broken internal links, redirect chains, orphaned pages, and mobile performance failures. These reduce performance rather than eliminate it, so they can be batched.
Structural work comes third: the site architecture itself. Category pages that exist, internal links that follow the keyword map, breadcrumbs, and a sitemap that lists exactly the URLs you want indexed and nothing else.
Schedule technical work before the bulk of content publishing, because every fix applied to the template benefits every page published afterwards. Applying the same fix to two hundred existing posts costs far more than applying it once to the template.
Phase 4: the content calendar (days 21 to 90)
With a keyword map and a crawlable site, content becomes scheduling rather than guesswork.
Sequence within each topic cluster:
- Pillar page first. The broad page that defines the topic and links out to everything else.
- Three to six supporting posts. Each answers one specific question from the map and links back to the pillar.
- Commercial page last. The comparison, alternatives, or pricing-adjacent page, published once supporting content exists to link into it.
A realistic first-quarter cadence for a small team is one pillar and two supporting posts per month, plus one refresh of an existing page. That is nine to twelve pages in 90 days — enough to establish one or two clusters properly, and far more useful than thirty scattered posts that each stand alone.
Every calendar row should carry a brief before it carries a writer. Briefs are what stop a post drifting off the keyword it was commissioned for; the structure to use is in content brief template. If you want the finished shapes to copy, the post-type patterns are worked through in blog post examples.
What belongs in each calendar row
- Publish date and owner
- Primary keyword and target URL from the map
- Page type and word target
- Internal links in and out
- Brief status and draft status
Phase 5: measurement (weekly from day 30)
SEO measurement fails in two directions: checking rankings daily and panicking, or checking nothing for six months and having no idea what caused what.
Weekly, review: impressions and clicks by query group, new queries appearing, pages moving into or out of the top twenty, and indexation of anything published that week. Fifteen minutes.
Monthly, review: cluster-level performance, which briefs produced pages that ranked and which did not, and whether the technical backlog is shrinking or growing. Then make one decision per cluster — continue, expand, or stop.
Quarterly, review: whether the keyword map still reflects the business, which pages need refreshing, and what the next 90-day plan should contain.
Do not expect meaningful ranking movement inside the first 90 days for competitive terms. What you should see by day 90 is a clean technical baseline, a populated calendar, indexed pages, and early impressions on long-tail phrasing. That is the correct milestone. Traffic follows in the quarters after.
Adapting the plan to your situation
New site with no history. Weight the plan towards long-tail informational content and get the technical foundation right first time. Skip competitive money terms entirely in quarter one.
Established site with traffic decline. Weight towards audit and refresh. The fastest wins are almost always pages that already rank in positions eight to twenty and need better intent match, not new pages.
Ecommerce. Category pages carry most of the commercial weight, so faceted navigation, duplicate variant URLs, and thin category copy dominate the technical phase.
Local service business. Add business profile completeness, review acquisition, and location page quality to the plan. National keyword competition matters far less than local relevance signals.
Where the plan usually breaks
The three failure modes are consistent. First, no owner per line, so nothing gets done and no one notices. Second, content production outruns the keyword map, so pages start competing with each other. Third, the plan is measured on rankings alone, so a quarter of genuine foundation work looks like failure and gets cancelled.
The fourth, quieter failure is capacity. Most plans assume writing throughput nobody actually has. If the calendar says twelve pages a quarter and the team produces four, the plan is fiction. Either cut the calendar to the real number or change how production happens. AutoSEO exists for that constraint — it generates SEO-structured articles from a keyword map, audits and fixes on-page issues, publishes to thirteen CMS platforms, and tracks rankings and AI-assistant visibility afterwards. It does not remove the need for the map or the briefs.
The strategy document that sits above this plan — goals, ICP, pillars, KPIs — is in SEO strategy template, and the quarter-by-quarter view for years two and beyond is in SEO roadmap template.
Frequently Asked Questions
What should an SEO plan include?
Five things, in order: an audit with severity-tagged findings, a keyword map assigning one URL per keyword, a prioritised technical fix list, a dated content calendar built from the map, and a measurement loop with weekly and monthly reviews. Anything else — outreach, digital PR, CRO — is an addition to that spine, not a replacement for it.
How long should an SEO plan cover?
Plan in 90-day blocks with a rolling twelve-month view. Ninety days is long enough to complete an audit, fix the technical baseline, and publish a first cluster, and short enough that the assumptions behind it are still true at the end. Anything planned in detail beyond a quarter tends to be rewritten before it is executed.
What is the difference between an SEO plan and an SEO strategy?
The strategy states what you are trying to win and why — the audience, the topics you intend to own, the position you are claiming. The plan states what happens on which dates and who does it. Strategy changes rarely; the plan is rewritten every quarter. A plan without a strategy becomes a task list with no direction, and a strategy without a plan never ships.
How do I write an SEO plan for a brand new website?
Spend the first fortnight on foundations rather than content: crawlable architecture, clean URL structure, titles and descriptions, sitemap, and analytics. Then build a keyword map weighted almost entirely towards long-tail informational phrases you can realistically win, and publish one tight cluster rather than scattered posts. Competitive commercial terms belong in quarter three or later.
How often should an SEO plan be reviewed?
Weekly for a fifteen-minute check on indexation and query movement, monthly for cluster-level decisions, and quarterly for a full rewrite of the next 90 days. The weekly review is for catching breakage — a page dropping out of the index, a fix that regressed — not for reacting to ranking noise.
