How does Google rank websites? The honest answer starts with what Google itself has written, because most of what circulates is a list of “ranking factors” with no source behind it. Google’s own explanation is short. It groups its signals into five families and says, in broad terms, what each one does. It does not say how much any of them counts.
That gap is where the industry lives. Leaked documents, court testimony and correlation studies all try to put numbers on the five groups. Some add real detail. None of them turns the groups into a formula, and you should be wary of anyone who says otherwise.
How does Google rank websites, in its own words
Google’s page on how Search ranks results names five groups of signals. The table sets out each one with Google’s own wording, so you can see how much is stated and how much is left open.
| Signal group | What Google says | What it leaves open |
|---|---|---|
| Meaning of your query | “we first need to establish what you’re looking for – the intent behind your query” | How intent is classified |
| Relevance of content | “The most basic signal that information is relevant is when content contains the same keywords as your search query” | What counts beyond keywords |
| Quality of content | Signals “that can help determine which content demonstrates expertise, authoritativeness, and trustworthiness” | Which signals, and their weight |
| Usability | “When all other signals are relatively equal, content that people will find more accessible may perform better” | How close “relatively equal” must be |
| Context and settings | “your location, past Search history, and Search settings” | How much results vary per person |
Two things stand out. First, the order is not arbitrary: meaning comes before relevance, because Google cannot judge whether a page answers a query until it knows what the query is asking. Second, usability is explicitly a tiebreaker. Google says accessible content “may perform better” only when everything else is close.
Meaning and relevance: the first filter
The first two groups decide which pages are in the running at all. Google says it must establish “the intent behind your query” before it returns results. A search for a product name might want a shop, a review or a manual, and the pages that compete are the ones that fit the intent Google settles on. Our guide to search intent covers how to read that from the results page itself.
Relevance comes next, and Google’s description is plainer than many SEOs expect. The “most basic signal” is that the content “contains the same keywords as your search query”. That is not an invitation to repeat a phrase. It is a reminder that a page which never uses the words people search for is asking Google to infer too much. Google calls it the most basic signal, not the only one, so treat the keyword as the entry ticket, not the whole test.
Quality: where most of the argument happens
Once Google has a set of relevant pages, it has to order them. This is where the quality group does its work, and it is the group Google describes least precisely. Google says it looks for signals “that can help determine which content demonstrates expertise, authoritativeness, and trustworthiness”. It does not list them.
That vagueness is why so much ranking advice is really quality advice in disguise: author bios, citations, links, reviews, brand mentions. Some of those are plausible proxies; none is confirmed as a direct input by this page. Our piece on E-E-A-T examples shows what demonstrated expertise looks like on a page, which is a more useful target than chasing any single proxy.
It is also worth stating what quality is not. Google says: “We never provide special treatment to advertisers in how our search algorithms rank their websites, and nobody can pay us to do so.” Ad spend does not buy organic position, whatever a sales call suggests.
Usability and context: tiebreakers and personalisation
Usability gets more attention than its wording justifies. Google’s sentence is conditional: “When all other signals are relatively equal, content that people will find more accessible may perform better.” A fast, mobile-friendly page beats a slow one when the two are otherwise matched. It does not lift a thin page above a strong one.
That does not make speed irrelevant. In competitive results many pages are close on relevance and quality, and a tiebreaker decides real positions there. Our pieces on page speed and SEO and mobile-first indexing cover the practical side. Just do not expect a performance project to fix a content problem.
Context is the group people forget when they check rankings. Google uses “your location, past Search history, and Search settings”. Two people typing the same query can see different results, so a single manual check is a sample of one.
What the leak and the trial added, and what they did not
Google’s page describes categories, not weights. Two later sources added texture. The 2024 documentation leak exposed attribute names from Google’s internal systems; our analysis of the Google algorithm leak sets out what it does and does not show. Trial testimony described a system that uses click data, covered in our piece on NavBoost.
Both are useful. Neither is a formula. A list of attribute names tells you what Google stores, not how much each one moves a ranking, and testimony about one system tells you little about how it is balanced against the others. If you see a confident list of exactly how many ranking factors Google uses, it did not come from Google’s ranking page, which gives no such count. Our round-up of SEO myths covers that one and its relatives.
The limitation is real and you should plan around it. Nobody outside Google can see the weights, they change, and they vary by query type. What you can do is work through the five groups in order. Does your page match the intent of the results that rank? Does it use the words searchers use? Does it show real expertise and earn trust? Is it accessible on the devices people use? Then compare your page against the pages that already rank for that query, because they are the only evidence you have of how Google weighs the groups for that result. That comparison is slower than reading a factor list, and it is the only method grounded in what Google actually ranks.
Frequently asked questions
What does Google use to rank websites?
Google names five signal groups: the meaning of the query, the relevance of the content, the quality of the content, usability, and the searcher’s context and settings.
What are the most important Google ranking factors?
Google does not publish weights. Meaning and relevance decide which pages compete; quality orders them; usability acts as a tiebreaker when other signals are relatively equal.
Can you pay Google to rank higher?
No. Google says it never gives advertisers special treatment in how its algorithms rank their websites, and that nobody can pay it to do so.
Why do my rankings look different from a colleague’s?
Context is one of the five groups. Google uses location, past Search history and Search settings, so two people can see different results for the same query.
The takeaway Google ranks on meaning, relevance, quality, usability and context, in roughly that order of gatekeeping. It publishes the categories, not the weights, so the pages already ranking are your best evidence of how they balance for any query.
