Google Ranking Factors
Ranking factors are not a checklist of two hundred switches. Google publishes no weighted list of factors; it publishes a list of systems, each solving its own problem. This lesson covers what is actually published, how to turn it into a working model, and why a strong domain does not carry a weak page.
We work from the guide to Google Search ranking systems and the explanation of how Search works. Anything absent from those is marked in the lesson as somebody's guess.
Ranking factors are not a checklist of 200 switches. Google uses many systems and signals to choose the most relevant, useful, and reliable result for a specific query.
How to Think About Ranking
Beginners often look for one secret: add the keyword, buy links, speed up the site, or improve E-E-A-T. In reality, a page ranks when it passes several layers at once: it can be found, understood, matched to intent, considered high quality, and shown comfortably to the user.
What Google actually publishes
"200 factors" lists exist because the real list looks different and does not make a headline. Google publishes a guide to its ranking systems — not factors, not weights, but mechanisms, each with its own job. Some of the ones named there:
| System | What it does |
|---|---|
| Link analysis and PageRank | Judges the significance of pages from the links between them — one of the oldest systems |
| BERT and neural matching | Work out how the words of a query relate to each other and to the page |
| Passage ranking | Lets a single section of a long page answer a narrow query |
| Freshness systems | Lift recent material where the query calls for it |
| Deduplication and site diversity | Remove repetition and stop one domain taking the whole first page |
| Original content systems | Try to show the source above the retellings of it |
| Spam detection | Finds policy violations, from link schemes to scaled content |
The "retired systems" section is worth a look of its own. Panda, Penguin, Hummingbird and the helpful content system sit there: not cancelled, but folded into the core ranking systems. So advice like "get ready for the next HCU" describes a world that no longer exists — it is not a separate update but part of the general machinery.
What the guide does not contain: weights, an order of importance, and the number 200. Anyone calling a factor "the third most important" is adding that themselves.
Seven layers: a model for work, not a description of Google
The list of systems answers "what is inside", but not "where do I start". For work, a different breakdown is easier — by the stage at which a page can fall out. This is a model built for convenience, not a description of the algorithm; Google names no such layers.
| Layer | What Google checks | How to improve it |
|---|---|---|
| Accessibility | Can the page be crawled and indexed? | robots.txt, noindex, canonical, sitemap, 200 status, internal links. |
| Relevance | Does the page match the words and meaning of the query? | Title, h1, structure, topical coverage, clear answers, user language. |
| Intent | Does the page solve the user's task? | Compare the SERP: guide, product, category, local result, video, list, or quick answer. |
| Quality | Is the content helpful, original, accurate, and people-first? | First-hand experience, facts, examples, complete answers, no fluff or copied content. |
| Authority | Does the site or page show recognition in the topic? | Natural links, mentions, brand reputation, expert content, topical authority. |
| Usability | Can users comfortably use the page? | Mobile UX, speed, Core Web Vitals, readability, no intrusive elements. |
| Context | Does the best answer change by country, language, device, freshness, or settings? | Localization, freshness, correct language, regional data, updates. |
Ranking starts before the content
If the page is blocked from crawling, the canonical points elsewhere, or the server returns an error, quality never gets assessed. The technical base is not "extra magic", it is the entry ticket.
| Problem | What you see | Where to check |
|---|---|---|
| The page is not indexed | No impressions, the URL is absent from search | URL Inspection and the indexing report |
| Google chose a different canonical | Impressions go to the wrong URL | The same URL Inspection; the rules are in the documentation on canonical URLs |
| The page is blocked in robots.txt | The crawler does not enter and the content is never read | robots.txt documentation |
| Weak internal linking | Important pages are buried deep | internal link context analysis |
| A slow or awkward page | People do not wait for it to load | the page experience guidance |
Walking that list for one URL is what the URL auditor does, and for the whole site the site audit. The glossary covers the terms: canonical and Core Web Vitals.
Relevance Is Not Keyword Density
Google needs to understand that the page answers the query. That does not mean repeating the keyword ten times. Relevance comes from page language, headings, structure, entities, examples, and how completely the question is answered.
- Bad: "buy laptop buy laptop laptop price cheap laptop".
- Better: a category page with filters, models, prices, reviews, specs, delivery, and a clear title.
- For an article: the topic is explained step by step with definitions, examples, mistakes, limitations, and a short answer near the top.
Intent can beat "perfect optimization"
When someone searches "best CRM for small business", the results are usually comparisons, lists and selection criteria. A landing page for one CRM can be fast, handsome and expert — and still miss what the query is for. You test that by opening the results, not by reasoning about them.
| Query | Likely intent | What the page needs |
|---|---|---|
what is robots.txt | Explain | A definition, the syntax, examples, the usual mistakes. |
buy running shoes | Buy | A category or product page, price, sizes, delivery, filters. |
cloud pos or online pos | Compare | A table, selection scenarios, the limits of each option. |
dentist near me | Choose nearby | A map, address, reviews, opening hours, booking. |
The intent analysis sorts your queries by type, and the SERP analyzer shows which format already won in your niche. The glossary covers the term: search intent.
Quality and E-E-A-T
Google states the goal as helpful, reliable content made primarily for people. What that means in practice is set out in the helpful content guidance, along with the list of questions it suggests asking of your own page.
E-E-A-T — experience, expertise, authoritativeness, trust — is not a separate factor with a weight. It is a set of traits raters assess material by, and a target the systems are tuned towards. It matters most where a mistake costs the reader dearly: health, finance, law, safety.
It works for an ordinary site too: a named author, real examples, current data, visible contacts and a clear editorial logic. The next lesson covers E-E-A-T in detail; the glossary covers the term: E-E-A-T.
Links and authority: important, but not a cure
Link analysis is one of the systems named in the guide, and nobody argues about whether links work. The argument is usually about something else: whether a strong domain carries a weak page. The guide answers that directly, and the answer is inconvenient for both sides.
Two things follow. First: a weak page on a strong domain can lose — "everyone knows us" is not an argument. Second: a strong page on a young site can win, and that is not an exception to the rules but how the system is built.
What not to do is buy links: that is named as a violation in the spam policies. The glossary covers the term: PageRank.
Bad Strategies
- Hunting for "200 factors". Most lists mix confirmed ideas, guesses, and outdated advice.
- Optimizing only the title. A title matters, but it cannot compensate for weak content or wrong intent.
- Treating speed as the main factor in every case. Speed matters, but a fast empty page is not the best answer.
- Buying links instead of earning quality. It is risky, especially when the content does not deserve citations.
- Copying the top result blindly. Understand why it ranks: intent, format, brand, links, freshness, depth, or locality.
Diagnosing Ranking Drops
| Symptom | Likely layer | First step |
|---|---|---|
| Impressions drop close to zero | Indexing or technical access | Check GSC, robots.txt, noindex, canonical, server errors. |
| Impressions exist, CTR drops | Snippet or SERP features | Compare title/description, ads, AI blocks, featured snippets. |
| Page is stuck in positions 6-10 | Relevance, depth, authority | Compare with top 3: format, completeness, links, experience, freshness. |
| Mobile traffic underperforms desktop | Usability and speed | Check mobile UX, CWV, content visibility, interactivity. |
| A new competitor pushes you down | Intent, quality, freshness, authority | Review their format, proof, structure, date, links, and brand. |
Practice: find your layer
Take a page that is not growing and walk the layers top down. Stop at the first one it fails — going further is pointless.
- Accessibility. Run the URL through URL Inspection: is the page indexed? Is the right canonical chosen?
- Relevance. In the performance report, look at which queries the page already shows for. If they are the wrong ones, the problem is here.
- Intent. Open the main query and see which format holds the top. If yours is a different format, the text is not the issue.
- Quality. Compare your page with the top three: what do they have that you do not — data, examples, an author, currency.
- Everything else. Only now argue about links, speed and details.
The wins and losses report shows what rose and fell over a period, and the page analyzer breaks down a specific page.
How to tell you have understood ranking
The sign is not knowing the systems by heart but that the question changed. It used to be "which factor will lift this page"; now it is "which layer does this page fall out at".
Three questions to check yourself. Can you name the layer your problem page gets stuck at? Can you tell what is published from what somebody assumed? Do you catch yourself thinking "we have a strong domain, it should carry this"?
What not to expect: a list of actions that works on any page. The layer you are stuck at differs per page, and the work starts by finding it, not by picking a tactic.
What comes next
The next lesson is E-E-A-T in detail: what trust in a piece of material is made of and how much of it Search can see. The previous lesson, on the anatomy of a SERP, is here.