SEO

SEO divides into technical health, content, and references from elsewhere. Almost every account is limited by one of the three, and the useful first move is to work out which, because effort spent on the other two produces nothing while the constraint holds.

A site with crawling or indexing problems cannot be helped by more content, since the content will not be seen. A site that is crawled perfectly and has nothing worth ranking cannot be helped by technical work. A site with both, in a competitive field, is usually limited by whether anybody else refers to it.

This is also the honest answer to the audit that lists ninety issues by severity. Most of them are real and almost none of them are the constraint, and an agency that delivers the list without saying which two things matter has transferred the diagnosis back to the client.

Changes take effect on a schedule nobody controls. Pages have to be recrawled, reassessed and then re-ranked against everything else competing for the query, and the gap between work and result is commonly measured in months rather than weeks.

That gap is where SEO retainers die. The client signed in January expecting something by March, and in March the traffic chart looks identical to the one they were shown at the pitch. Nothing has gone wrong, but nothing visible has gone right either, and a quarterly review with one flat line in it is difficult to survive.

The way through is to report the chain rather than only its last link. Pages published, pages indexed, crawl errors cleared, impressions, average position, click-through rate, then sessions and finally conversions. Each of those moves earlier than the one after it, and showing the early ones moving is evidence in a month where sessions have not.

It works the other way too, which is worth saying at the pitch. A client who pauses the work for a quarter usually sees no immediate drop, concludes it was never doing anything, and cancels. The decline arrives later, after the decision has been justified.

For most of the history of SEO, position and traffic moved together closely enough that reporting one stood in for the other. That has stopped being reliable. AI Overviews and similar generated summaries now sit above the results for many informational queries and frequently answer them outright.

The practical effect is a shape agencies keep encountering: rankings held or improved, impressions held, clicks down. Nothing regressed. The page simply stopped being the thing the searcher had to click to get the answer.

Rich results do a milder version of the same thing. A featured snippet, a local pack or a shopping unit each satisfy some share of searchers on the results page itself. The effect is strongest on queries with a short factual answer and weakest on queries where somebody needs to compare, buy or contact somebody.

The reporting consequence is that visibility and traffic now need to be shown as two things. A client looking at one chart will read a flat click line as a failure, and a client looking at both can see that the work held position in a results page that changed underneath it.

Google has moved its public guidance steadily away from tactics and toward whether the content demonstrably helps the person who searched. The helpful content guidance is the clearest statement of that, and it is written as a list of questions about the content rather than a list of things to do to it.

The Search Quality Rater Guidelines are the other public document worth reading, and they are more useful than their status suggests. Raters do not change rankings directly, but the document describes what Google is instructing humans to call a good result, which is the closest thing there is to a written specification.

E-E-A-T, meaning experience, expertise, authoritativeness and trust, comes from that document. It is not a score and there is no way to check it, which is why agency claims to have improved a client E-E-A-T should be read with suspicion. What it does give is a checklist that is genuinely predictive: is it clear who wrote this, do they have any standing to write it, is the site transparent about who runs it, and does the page show first-hand knowledge rather than a summary of other pages.

For pages that can affect somebody health, finances or safety, that bar is explicitly higher. A financial advice page written anonymously is a structural problem rather than a content quality one, and no amount of editing fixes it while the byline is missing.