How to do SEO: a 12-week plan from zero

Answers to how to do SEO usually arrive as a checklist: fix your titles, publish content, earn links. The list is not wrong, but it leaves out the one thing that matters most — the order. For a team starting from zero, the real difficulty is knowing which job comes before which, because some jobs produce nothing at all when the work they depend on has not been done. This article turns how to do SEO into a 12-week working plan arranged by dependency. Whether or not you bring in an agency, the logic of the sequence stays the same.

Why 12 weeks, and why this order

The sequence is a dependency chain, not an arbitrary calendar. Measurement comes first, because without it you cannot tell afterwards whether any improvement actually worked. Technical accessibility comes second, because the best content on a page that cannot be crawled or indexed produces nothing. Content comes third, because query matching only becomes meaningful on a page that is reachable. Authority work is left until last, because attracting outside attention to a page you have not yet fixed spends the investment in the wrong place.

Twelve weeks is a practical choice. Changes on the search side take time to be recrawled and reassessed, and week-to-week fluctuations are misleading. A three-month window is usually enough to separate the effect of your work from the noise. What follows is a work plan, not a promise of outcomes: rankings vary with competitive intensity, a site's history and the sector, and no engagement can promise a specific position.

Weeks 1-3: measurement and inventory

The first three weeks serve one purpose: building a basis on which, months later, you can answer the question "did it work?" Verify the Search Console and analytics setup, confirm property ownership, then export and store the starting data. That record matters, because Search Console retains data for a limited period, and three months from now the only thing able to answer "where did we start?" is the copy you take today.

In the same period, produce a URL inventory: which pages exist, which of them are meant to appear in search, and which are not. Teams that skip this distinction spend hours in later weeks troubleshooting pages that were never meant to be indexed at all. We covered which Search Console report answers which question in our Search Console guide.

It helps to add a short competitor reading beside the inventory: search your target queries and look at what kind of pages hold the first page. If the results are mostly guides, that query will not be satisfied by a service page; if they are mostly product or service pages, entering with a blog post is hard. This reading needs no tool, and it settles in advance what each page will be over the following nine weeks.

By the end of week three you should hold: a verified measurement setup, an exported baseline, a target page list, and a one-sentence purpose for each of those pages.

Weeks 4-6: clearing technical obstacles

From week four the question becomes: are the pages on the target list actually indexed? The page indexing report gives a status for every URL, and the vocabulary of those statuses is worth learning. "Crawled - currently not indexed" and "Excluded by 'noindex' tag" are entirely different problems and go to different people: one is a content job, the other a development job.

Three groups of problem typically surface here: sections blocked by accident, canonical tags that cancel themselves out, and duplicates created by publishing the same content at several addresses. All three are silent; the page opens in a browser, a user sees it, a search engine does not. We set out the diagnosis and repair order in our indexing problems article, and putting the whole infrastructure under a regular audit is what our technical SEO service covers.

By the end of week six you want a single answer for each page on the target list: is it indexed, if not for what reason, and who owns that reason.

Weeks 7-9: content and query matching

Once the technical obstacles are gone, attention turns to what each page actually answers. The exercise that helps most here is simple: for every target page, write in one sentence the question that page should answer, then read the page and check whether it really answers it. In practice most pages fail not on technique but on the weakness of that match: the page is about the topic, but does not give the answer the query expects.

In the same period the performance report is read as a second source: which queries does a page get impressions for, and do those queries line up with its purpose. Pages that get impressions but no clicks are usually a title and description problem; pages that get no impressions at all are usually a content-to-query mismatch. Making that distinction stops the following weeks from being guesswork. The systematic side of this is the subject of our content SEO work.

A common mistake in these three weeks is setting out to write a new article for every target page. On most sites the faster gain comes from reworking pages that already exist and already get impressions so that they fully answer the question they target. A new page is only needed when nothing in the inventory corresponds to that question.

Weeks 10-12: authority, re-measurement and handover

The last three weeks run two jobs together. First, organise the internal link flow into the pages you have now fixed, and produce real reasons for outside citation: an original data compilation, a tool, a written method. Second and more important, put the baseline taken in week one beside today's data. What changed, what did not, and which change showed up on which page.

At the end of week twelve the work does not stop; it changes shape. The project tempo gives way to a monthly routine: a measurement reading, one technical audit pass, and a limited number of content decisions each month. We covered how that monthly rhythm is built in our monthly work plan article. When you decide which part of the plan to keep in-house and which to source outside, you can look at what our services cover.

What this plan does not give you

The plan makes no promise about rankings, and it cannot. Google states plainly in its own documentation that it does not guarantee that a page will be crawled, indexed or served, even one that follows the rules. What the plan gives you is different: after twelve weeks, you can say what is working on the site and what is not from your records rather than from guesswork. For a team starting from zero, that is the realistic gain of a first quarter.

Sources

• Google Search Central — SEO Starter Guide: basic SEO practices and the topics not worth prioritising.
• Google Search Central — Search Essentials: technical requirements, spam policies and key best practices.
• Search Console Help — Page indexing report: indexing statuses and what they mean.