SEO
SEO: what you can influence and what you cannot
SEO is the practice of making a site easier for a search engine to crawl, index and judge relevant. This guide separates what you can directly control, what you can only influence, and what is outside your hands entirely — and explains how to test a claim rather than believe it.
Search engine optimisation is best understood as three concentric circles: what you control outright, what you influence, and what happens to you.
What you control outright
Everything on your own server. Whether pages return the right status codes. What
your robots.txt allows. Whether important content is in the initial HTML.
Which pages exist and how they link to each other. What each page actually says
and how completely it answers the question it targets. How fast the server
responds. These are engineering and editorial decisions, and they either happen
or they do not.
What you influence
How a search engine judges the result. You can make a page more relevant to a
query, but you cannot decide that it is the most relevant. You can earn links,
but you cannot control who links to you. You can write a <title> and a meta
description, and the engine may still rewrite them for a given query. You can
set rel=canonical, and the engine may still pick another URL.
Treating influence as control is the source of most disappointment in SEO.
What is outside your hands
Competitors improving. Query intent shifting. Algorithm updates. How much of the results page is given to organic results at all for your queries. Whether a query starts triggering an AI Overview that answers it without a click.
How to evaluate any SEO claim
The niche is full of confident assertions with no evidence behind them. Three questions handle most of them:
- What is the primary source? Is this in a search engine’s own documentation, a published patent or paper, or is it one person’s inference repeated until it sounded official?
- Is it correlation or causation? Ranking correlations are easy to produce and almost never establish cause. A study showing that ranking pages have more of something does not show that adding more of it causes ranking.
- Can it be tested on a site you control? If a claim cannot be tested, it cannot be relied on. If it can, test it before you act on it at scale.
The order of operations
For a site that is not performing, work the pipeline in order rather than starting with content. Confirm the pages are crawled. Confirm the crawled pages are indexed. Then, and only then, work on relevance and competitiveness. Fixing content on pages a crawler cannot reach changes nothing at all.
Articles in this guide
The articles for this guide are being written. The guide above covers the whole topic in outline, and the article index lists everything published so far.