How Google Works: Crawling, Indexing and Ranking

Four stages between publishing a page and someone finding it. Knowing which stage a problem sits in is what turns SEO from guesswork into debugging.

Concept

A page goes through four distinct stages before anyone can find it in search. Most SEO problems are really a failure at one specific stage, and the fix depends entirely on which one.

A four stage pipeline: discover a URL from links and sitemaps, crawl it while obeying robots.txt, render and index it by extracting title, headings, links and structured data, then serve and rank it. Below, three ways a page drops out and a list of the markup that controls each stage.
The pipeline, and the markup that controls each stage.

1. Discovery

Before anything else, the URL has to be found. There are three routes:

  • Links from pages already known, internal or external. This is the main one.
  • An XML sitemap submitted or referenced from robots.txt.
  • A direct request through Search Console.

A page with no links pointing at it and no sitemap entry is an orphan. It exists, and nothing will ever find it. Publishing is not enough.

2. Crawling

Googlebot requests the URL. Several things can stop it:

  • robots.txt disallows the path.
  • The server returns an error, or is too slow.
  • The page requires a login.
  • The site has exhausted its crawl budget on low value URLs.

Crawl budget is roughly how many pages a crawler will fetch from your site in a period. On a small site it is never a constraint. On a large one, thousands of near duplicate filtered URLs can consume it and leave real pages uncrawled.

3. Rendering and indexing

The HTML is parsed, JavaScript is executed, and the page is analysed: the title, headings, body text, links, images, structured data and canonical tag.

Rendering is worth a note. Google does execute JavaScript, but rendering happens in a separate pass and can be delayed. A page whose content only exists after a client side fetch may be crawled, seen as empty, and rendered days later. Content that matters should be in the HTML the server sends.

The page is dropped at this stage if it carries noindex, if a canonical points elsewhere, or if it is judged a duplicate of something already indexed.

4. Serving and ranking

For a given query, candidate pages are selected and ordered. The ordering uses a large number of signals; the ones you influence directly through markup are:

  • Whether the content matches the query.
  • The title, headings and structure.
  • Links pointing at the page and the text of those links.
  • Page experience: Core Web Vitals, mobile usability, HTTPS.
  • Structured data, which affects how the result is displayed rather than where it ranks.

Where problems actually live

SymptomStageCheck
Page not in search at allDiscovery or crawlingSitemap, internal links, robots.txt
Crawled but not indexedIndexingnoindex, canonical, thin content
Indexed under a different URLIndexingCanonical tag
Indexed but ranks poorlyRankingContent, title, links
Ranks well, nobody clicksServingTitle and description

Diagnosing the stage first saves a great deal of wasted effort. Rewriting content will not help a page that is blocked in robots.txt.

Search Console

Google Search Console is free, and it is the only place you can see what Google actually did with your pages rather than what you assume.

The reports worth knowing:

ReportAnswers
URL InspectionIs this specific page indexed, and if not, why?
PagesWhich pages are indexed and which were excluded, with reasons
PerformanceWhich queries show your pages, and the clicks each gets
Core Web VitalsReal visit performance data
SitemapsWhether your sitemap was read and how many URLs it contained
EnhancementsWhether structured data was understood

Verify ownership by adding a meta tag to the head or a file to the site root:

<meta name="google-site-verification" content="...">

Set it up on the day a site launches, not months later. The historical data only starts from verification.

What the markup controls

<head>
  <title>Design Course - Riverside College</title>
  <meta name="description" content="A three year studio based design course in Pune.">
  <link rel="canonical" href="https://example.edu/courses/design">
  <meta name="robots" content="index, follow">
</head>
<body>
  <h1>Design course</h1>
  <main>
    <article>
      <h2>What you will study</h2>
      <p>...</p>
    </article>
  </main>
</body>

Every line here maps to something in the pipeline: the title becomes the blue link, the description becomes the snippet, the canonical decides which URL is credited, the robots tag decides whether the page is listed, and the headings tell the crawler what the page covers.

What SEO is not

  • The keywords meta tag. Ignored by every major search engine for well over a decade.
  • Keyword density. Repeating a phrase a fixed number of times is not how modern ranking works.
  • Hidden text. Text hidden with CSS to feed a crawler is a policy violation and risks a manual penalty.
  • Buying links. Same.
  • A one time task. Content and competitors both change.

Important rules

  • A page must be discoverable, crawlable and indexable before ranking is even relevant.
  • robots.txt controls crawling; noindex controls indexing. They are different stages.
  • Content behind JavaScript can be indexed late or not at all.
  • Indexing is not guaranteed. Google may crawl a page and decide not to list it.
  • Search Console shows what actually happened.

Common mistakes

  • Rewriting content for a page that is blocked from crawling.
  • Assuming publishing is enough, with no internal links or sitemap.
  • Shipping a staging noindex to production.
  • Rendering all content client side and expecting immediate indexing.
  • Never setting up Search Console.
  • Blocking CSS and JavaScript in robots.txt, so Google cannot render the page properly.

Best practices

  • Verify Search Console on launch day.
  • Submit an XML sitemap and keep it accurate.
  • Link every page from at least one other page.
  • Serve the important content in the initial HTML.
  • Audit for stray noindex tags before every release.
  • Diagnose by stage before changing anything.

Practice

  1. Use URL Inspection on one of your pages and read exactly what Google recorded.
  2. Find a page on your site with no internal links pointing at it.
  3. Disable JavaScript and reload a page. What content disappeared?
  4. For each of the five symptoms in the table, name the first thing you would check.

Useful resources

Hand picked references for this topic
Written by Lorens Mishra

Default administrator account created by the installer.

Continue reading

All HTML notes →
HTML

Image SEO

Image search sends real traffic, and images are usually the heaviest thing on a page. File names, alt text, formats and dimensions are the whole job.

Read more

Discussion

0 comments
Sign in to join the discussion.

No comments yet. Be the first to say something.