SEO is a season, not a single match. Even in the age of AI

Football helped me explain the time involved

The original title was “What does SEO mean? Let me explain it through… football”. I was looking for a familiar image to make sense of work that, from the outside, can look like a series of unrelated tasks: improving a page, writing content, checking a report or earning a link.

My table placed club ownership alongside the client, the coach alongside the professional, the website alongside the team, and match statistics alongside analytics. Some comparisons were straightforward. Others stretched the metaphor: off-site SEO became the transfer market, while responsive design became participation in more competitions.

It was an introductory way to describe the subject, not a technical model. Reading it today, I can see where the matching game oversimplified things. I have no interest in defending every row simply because I wrote it.

One passage still sounds familiar. I described the work as a “strategic, effective, meticulous and lengthy process”. I would lose a few adjectives, but not the idea of a process. That was the important part, and it still is.

One match matters. A season asks for more

A team can play a brilliant match without having the consistency needed for a whole season. It can also lose a game without making all its preparation worthless. This is where the metaphor remains useful to me.

On the web, an individual change can make a real difference. Removing an indexing block can finally make a page eligible for indexing. Rewriting content can make an offer clearer. A relevant link can help someone discover a resource. I do not dismiss these changes; I question expecting any one of them to do the work of an entire project.

Football has a set fixture list. Search has no season with a fixed date for results. Using a season to explain continuity therefore does not mean promising a reward after a predetermined number of months.

I would also change the old article’s objective. “Being found first” was too vague. I do not need a website to appear for every possible search. I want it to offer something relevant to people looking for that particular service, product or information.

Continuity does not mean waiting for the website to do everything on its own. It means continuing to observe, choose and check.

I keep the coordination, not the table

Today I would not assign a football role to every part of a website. I would rather explain what the project needs to bring together.

Technical foundations help make content reachable and understandable: structure, server responses, URLs, navigation and page behaviour. The content then needs to say something useful and distinctive. Intent gives both a direction: what is this person looking for, and why should this page help?

I do not reduce authority to the number of ranking keywords. I look at the expertise I can document, the identity of the person speaking and the quality of the information. Internal and external links need a reason and a context; I do not treat them as a collection of points to accumulate.

Time allows me to observe how these elements are discovered and used, compare meaningful periods and make changes justified by the evidence. It does not automatically make a weak page useful. Publishing more material without direction can increase a website’s volume without making its value any clearer.

That is the part of the metaphor I am happy to recover: different components working towards a shared purpose. The rest can stay in the archive.

Being indexed is not being chosen

In the 2020 article, “being found” bundled together stages I now prefer to separate when speaking with someone commissioning a website. The distinction helps prevent an intermediate signal being presented as a commercial result.

A page can be indexed without receiving useful visits. It can receive visits without generating enquiries. And an enquiry can come from a recommendation or social media without involving organic search.

I explain the earlier stages in the guide Why your website does not appear on Google. I am not repeating its diagnosis here: my concern is the risk of mixing up stages when assessing the work.

Statistics should support decisions

The old article also paired analytics with football statistics. The comparison holds as long as we look beyond the scoreboard. A number becomes useful when I know which question I am asking it.

I use Search Console ↗ to read impressions, clicks, queries and pages in Google Search. Indexing checks answer a different question: whether and how the search engine has included a page in its index.

With Plausible ↗, I look at what is measured on the website: traffic sources, entry pages and actions configured as events or goals. It is not a second Google ranking table. It observes a different stage.

The two tools should not be expected to produce identical figures: they measure different things and use their own criteria. I read them together to understand the journey without assigning information to a report that it does not contain. Clicking a contact button, for example, does not establish that a new client has been acquired.

When comparing periods, I consider the changes made, traffic sources and the amount of available data. A short-term movement may suggest something to check; on its own, it does not justify a definitive explanation. For me, measurement mainly helps choose the next intervention with fewer assumptions.

AI does not erase the previous field

In 2020 that table could not describe searching through the generative assistants I use today. In 2026 someone may discover a business within an AI answer before looking for its website or getting in touch.

This broadens what I need to observe. It does not make SEO obsolete. Google’s documentation on AI search features ↗ confirms that SEO foundations remain relevant to AI Overviews and AI Mode, and that meeting the requirements does not guarantee inclusion.

I do not automatically apply Google’s rules to every assistant. Systems do not all consult the same sources, in the same way or at the same time. That is why I distinguish a citation I have checked from an assumption about how it was produced.

When I also work on making information understandable to these systems, often described with terms such as GEO and AEO, I start with clear information, verifiable identities, accessible content and coherent links. A new label is not proof that all previous work has become irrelevant.

I can control what I publish and check the answers I observe. I cannot instruct a system to cite me, nor turn a citation into a guarantee of visits or clients. AI adds another field of play; it does not change that boundary.

Sound foundations do not automatically create demand

In 2026 I rebuilt my working foundations and the way I design websites. I tested them on my own projects and also use them for client websites. I have already written about that rebuild; here I am interested in what happens once a sound foundation is in place.

Building technical SEO, performance, accessibility and measurement into a website creates conditions I can control. It does not automatically create interest in an offer, change when someone decides to buy, or replace the trust needed to commission a project.

On my website I bring together service pages, Guides and Notes, relevant links and content distribution. I check that they describe the same profession and read the stages separately, without treating an impression as an enquiry. I document the work through checks and comparisons over time, keeping evidence separate from interpretation.

For someone commissioning a new website from me, that distinction matters. The quality of the build and the tools used to measure it have a defined scope. Commercial results also depend on the offer, audience, competition and decisions that good technical structure cannot make on their behalf.