Google ranks pages with automated systems. They weigh relevance, helpful and original content, links from other sites, page experience and how people interact with results, then demote spam. Google documents most of this itself. The 2024 API leak and the US antitrust trial added detail on clicks, site-wide quality scores and Chrome data. Neither revealed any weights. For a B2B site, the order that pays is simple. Get the selling pages indexed, match each one to a real buyer search, and make it the best page on that search. Then earn links.
I'm Brenden House, and I run Searchmaxxed, an SEO and AI SEO agency. On 27 October 2026 (AEST) I read Google's ranking documentation, the leak write-ups and the trial exhibits, then sorted every claim by its evidence.
In summary
Google confirms its ranking factors in its own documentation. Everything else is evidence of varying strength, and you should weight it that way.
| Ranking factor | How strong the evidence is |
|---|---|
| Relevance to the search (words, intent, headings) | Google documents it |
| Helpful, people-first, original content | Google documents it |
| Links from other sites (PageRank) | Google documents it |
| Page experience and Core Web Vitals | Google documents it, as a tie-breaker |
| Interaction data from searchers | Google confirms; trial evidence and the leak add detail |
| Site-wide authority and topic focus | The leak named the attributes; weights unknown |
| Spam policies (scaled content, link spam) | Google documents and enforces them |
- There's no official list of 200 factors. A 2025 trial exhibit says Google uses "over 100 raw signals", rolled up into a smaller set of top-level signals.
- Clicks matter. Google's own How Search Works page now says it uses "aggregated and anonymized interaction data" to judge relevance.
- The leak showed what Google stores, not what it weighs. Treat its attribute names as hints.
- On a B2B site, indexing and intent match come first, because nothing else counts on a page Google won't show.
- The same work feeds Google's AI Overviews and AI Mode. Google says they need nothing extra.
What are Google's ranking factors?
Google's ranking factors are the signals its automated systems use to order results for a search. Google groups them into five areas on its How Search Works page: the meaning of your query, relevance, quality, usability and context.
A ranking signal is one measurable input, such as how well a title matches a query. A ranking system combines them into an order. Google's guide to its ranking systems names the notable ones. How Search Works adds that "the weight applied to each factor varies depending on the nature of your query". So news searches lean hard on freshness, while a search for a definition barely cares how old the best page is.
Here's what that means for a marketing team. No single factor wins on its own. That's why we plan SEO page by page at Searchmaxxed, starting from the searches that bring pipeline, never from a checklist.
What does Google itself confirm?
Google confirms relevance, helpful content, links, page experience, interaction data and spam detection as ranking inputs, all on its own pages. Build your plan on these.
| Factor | What Google says | Source |
|---|---|---|
| Keywords and relevance | "The most basic signal that information is relevant is when content contains the same keywords as your search query." | How Search Works |
| Helpful content | Its systems "prioritize helpful, reliable information that's created to benefit people" | Creating helpful content |
| Links | PageRank "continues to be part of our core ranking systems" | Ranking systems guide |
| Mentions from known sites | One quality factor is whether "other prominent websites link or refer to the content" | How Search Works |
| Core Web Vitals | "Core Web Vitals are used by our ranking systems." | Page experience |
| Interaction data | Google uses "aggregated and anonymized interaction data to assess whether search results are relevant" | How Search Works |
| Freshness | "Query deserves freshness" systems favour newer pages where searchers expect them | Ranking systems guide |
| Spam | SpamBrain and other systems act on its spam policies | Spam policies |
I read each page on 27 October 2026 (AEST); the ranking systems guide itself was last updated by Google on 10 December 2025.
The guide also names the AI systems behind relevance. RankBrain links words to concepts, BERT reads how word order changes meaning, and neural matching connects queries to pages that use different words. Passage ranking scores single sections of a long page. So the old habit of repeating the exact phrase matters less than covering the topic in the words your buyers use.
The helpful content system no longer runs on its own. Google says it "became part of" the core ranking systems in March 2024, the same path Panda and Penguin took years earlier.
What did the 2024 Google API leak reveal?
The leak exposed internal Google documentation describing 14,014 attributes across 2,596 modules, according to the Wikipedia summary of the analyses. It showed what Google stores about pages and sites. Which attributes affect ranking, and by how much, it never said.
Rand Fishkin broke the story on SparkToro on 27 May 2024, and Mike King published his technical read on iPullRank a day later. Fishkin says the files sat on GitHub from 27 March to 7 May 2024. Two days after his post, on 29 May 2024, Google told The Verge the documents were real. Its spokesperson added a warning: "We would caution against making inaccurate assumptions about Search based on out-of-context, outdated, or incomplete information."
That warning deserves weight. The Verge noted that some data "could be out of date, used strictly for training purposes, or collected but not used for Search specifically".
The attributes that drew the most attention, as the analysts reported them:
| Attribute named in the leak | What the documentation describes | Fair reading |
|---|---|---|
| goodClicks, badClicks, lastLongestClicks | Click features tied to a system called Navboost | Google records whether searchers stay or bounce back |
| siteAuthority | A site-level score "applied in Qstar" | Google keeps some site-wide quality measure |
| siteFocusScore, siteRadius | How tightly a site sticks to one topic, and how far each page strays | Topical focus gets measured |
| chromeInTotal | "Site-level Chrome views" | Chrome data exists in quality modules |
| hostAge | Used "to sandbox fresh spam in serving time" | New domains can face extra caution |
| titlematchScore | How well a title matches the query | Titles still count for relevance |
| OriginalContentScore | Originality scoring for short content | Thin can mean unoriginal, not only short |
| exactMatchDomainDemotion | A demotion for keyword domains | Matches Google's documented exact match domain system |
Notice the pattern. Most of the leak lines up with something Google already documents, or with trial evidence. The surprise was in the detail. Google spokespeople had denied a site authority score and the use of clicks for years, as Fishkin and King both set out.
So I treat the leak as confirmation of direction. Never a recipe. An attribute name tells you Google measures something, but it can't tell you how hard to push or what a change is worth.
Does Google use clicks to rank pages?
Yes. Google says on its How Search Works page that it turns "aggregated and anonymized interaction data" into signals for its machine-learned systems. Trial evidence names one system that does this: Navboost.
A US Department of Justice exhibit records a 31 January 2025 call with Pandu Nayak, a Google search engineer. The notes call Navboost "a measure of how frequently users (subset by location and device type) click on a particular document" for a query. They add that it "uses most recent 13 months of data", and the exhibit itself sits on the DOJ site for anyone to read.
The same notes carry a caution worth quoting. "Google avoids simply 'predicting clicks' because clicks are easily manipulated and are a poor proxy for enhancing user experience."
What does that mean in practice? Click-through rate tricks won't help. A page that earns the click and then ends the search should, because that's the behaviour Google can see across thousands of searches. For a B2B buyer, ending the search means answering the real question. Does this product do what I need, at my scale, with my stack? And what will it cost to find out?
That's how we write every page at Searchmaxxed. The answer goes first. The title promises exactly what the page delivers, so the click and the read agree.
How many ranking factors does Google use?
Google doesn't publish a number. The best primary evidence is the January 2025 DOJ exhibit. It says Google "uses over 100 raw signals", which it rolls into top-level signals and then into one score for each result.
The popular "200 ranking factors" lists are compilations. They mix confirmed factors, patents, old Google statements and guesses into one page. Some items, like keyword density, now count as spam.
Why can't any list rank the factors for you? The exhibit says top-level signals are "a linear combination of log of individual raw signals". Each signal's effect differs by search, and Google keeps every weight private, which is why any ranked list of factors is a guess.
My advice to marketing leads is short. Spend nothing on factors you can't trace to Google or to trial evidence, since the confirmed ones already hold a year of work.
Which factors matter most on a B2B website?
On a B2B site, the factors that matter most are indexing, intent match, useful content, links from your market and the click. Order matters. Content quality can't help a page Google never shows or never matches to a search.
Here's the order I'd work in, and what each step looks like on a real B2B site.
| Priority | What to check | Ranking factor behind it |
|---|---|---|
| 1. Indexing and rendering | Product, solution, integration and pricing pages are crawlable, render without JavaScript failures and carry the right canonical | Pages must be indexed to rank at all |
| 2. One page per buyer search | Each comparison, alternative, use-case and industry search has its own page and title | Relevance, title match, intent |
| 3. Content worth ranking | Specs, pricing logic, implementation detail, real customer numbers | Helpful content, originality |
| 4. Links and mentions | Coverage in trade press, analyst lists, partner directories, review sites | PageRank, prominence |
| 5. The click | Titles and snippets that match what the page delivers | Interaction data, Navboost |
| 6. Topic focus | Content clusters around what you sell, joined by descriptive internal links | Site focus, internal links |
| 7. Page experience | Good Core Web Vitals on the templates that carry traffic | Page experience, as a tie-breaker |
Why indexing comes first
Indexed and ranked are different stages. Google's guide to how Search works describes three: crawling, indexing and serving results. A page can pass the first two and still earn almost nothing, while a dashboard full of indexed pages makes the site look healthy.
Here's the gap on a site we own. We launched a new site with 161 guides, each built for one search from the live results and People Also Ask. We submitted the sitemap and pinged IndexNow. No paid indexer. About 25 hours after Google's first crawl, 203 of 228 URLs were indexed. Indexed isn't ranked, though: 5,198 impressions in ten days and 5 clicks.
That split matters. Indexing problems are technical, and a good developer can usually fix them in days. Ranking takes relevance, links and time. Our technical SEO service covers the first half: crawl, rendering, canonicals, sitemaps, redirects and Core Web Vitals on the templates that matter.
A worked example: one SaaS company, three fixes
Take a B2B SaaS company that sells expense-management software. It's an illustration.
Its product pages rank on page two for "expense management software". It has no page for searches that pair a rival's name with "alternatives", no page per integration, and a blog full of general finance tips.
Fix one is indexing: its integration pages render through JavaScript and Google sees empty templates. Fix two is intent: it builds an alternatives page, a comparison page and one page per integration, each titled for its search. Fix three is focus. The blog moves toward the problems buyers search for, and every post links to the product page that solves them.
None of those fixes is clever. Each one maps to a factor Google documents. That's the whole point.
Do backlinks still matter in 2026?
Yes. Google says PageRank "continues to be part of our core ranking systems". Its How Search Works page also counts links and references from "other prominent websites" as a quality factor. The trial exhibit says PageRank still feeds the quality signals for pages.
What's changed is which links count. Google's spam policies list buying links, excessive link exchanges, automated links and "low-quality directory or bookmark site links" as link spam. Google designs its systems to ignore or demote them.
For a B2B company, the links that move rankings come from places your buyers already read. Think trade publications, industry associations, partner and marketplace listings, conference pages and analyst roundups. Original data earns the best of them, because journalists need numbers to quote and most B2B companies sit on numbers nobody has published.
Our plans include this work. Growth System includes one qualified link action a month, while Agentic Search Engineering adds a monthly press release, continuous digital PR and qualified link campaigns. If you want the full case for links, read are backlinks still important.
What is E-E-A-T, and is it a ranking factor?
E-E-A-T stands for experience, expertise, authoritativeness and trustworthiness. Google describes it as a quality concept its systems approximate with other signals. In its words, "E-E-A-T itself isn't a specific ranking factor".
Trust matters most of the four. Google's human quality raters score results against E-E-A-T, using the search quality rater guidelines, and Google says that feedback helps it check its systems. Its helpful content page explains the link. Google also gives "even more weight" to strong E-E-A-T on topics that affect health, money or safety. It calls these Your Money or Your Life (YMYL) topics.
B2B pages often touch YMYL ground: finance software, security, compliance, medical devices, infrastructure. On those pages, show who wrote each one, what they've done and where every number comes from. We hold our own research to that rule.
What does Google mean by helpful content?
Helpful content is content made for a real audience first, with original information, depth and first-hand expertise. Google's self-assessment questions ask whether a page offers "original information, reporting, research, or analysis" and "substantial value when compared to other pages in search results".
Information gain is the shorthand for that second test: what does your page add that the current top results don't? Google's warning signs help too. It flags "lots of content on many different topics" and "extensive automation", plus writing to a word count and changing dates without real changes. On word counts, Google's answer is blunt: "No, we don't" have a preferred one.
Two questions on Google's list line up with the leak. Google asks, "Does your site have a primary purpose or focus?" The leak describes siteFocusScore and siteRadius, which measure exactly that. When Google's public advice and its internal attribute names agree, I treat the factor as real.
Does it matter whether AI wrote the page? Google's spam policy targets pages made "for the primary purpose of manipulating search rankings", "no matter how it's created". My own view, as I put it on a call this year: "It just matters if it's good and if you're giving it information gain. That's literally all that matters."
For B2B companies, the best information gain sits in your own business. Implementation timelines, pricing logic, benchmark data, failure cases and spec sheets are material your competitors can't copy. AI answers like to quote it too. Our guide to B2B content strategy goes further on building pages from that material.
How much do Core Web Vitals and page experience matter?
Core Web Vitals count, but mostly as a tie-breaker. Google says that "when all other signals are relatively equal, content that people will find more accessible may perform better".
Core Web Vitals are three measures of real-user experience. Google's targets are:
| Metric | What it measures | Good score |
|---|---|---|
| Largest Contentful Paint (LCP) | Loading | Within 2.5 seconds |
| Interaction to Next Paint (INP) | Responsiveness | Under 200 milliseconds |
| Cumulative Layout Shift (CLS) | Visual stability | Under 0.1 |
Google adds that other page experience aspects "don't directly help your website rank higher". It still counts HTTPS, mobile display, restraint with ads and no intrusive pop-ups as part of a good experience. And it warns that chasing a perfect score "just for SEO reasons may not be the best use of your time".
So fix the templates that carry traffic: product pages, solution pages, the blog template. A heavy marketing site with a chat widget, three tracking scripts and a hero video can fail INP and LCP on mobile. That's worth fixing for buyers anyway.
What pushes a site down in Google?
Spam and manipulation push sites down. Google's spam policies name the practices it demotes or removes, and its spam updates enforce them several times a year.
The three that hit B2B sites most often:
- Scaled content abuse. Many pages made mainly to rank, "no matter how it's created". Hundreds of near-identical location or "use case" pages with swapped keywords fall here, whoever or whatever wrote them.
- Link spam. Bought links, link swaps, keyword-rich anchor text in guest posts and press releases, and low-quality directory links.
- Site reputation abuse. Third-party pages hosted on a strong site mainly to borrow its ranking signals. Think sponsored "best of" sections on an unrelated domain.
Google's Search Status Dashboard shows how often the rules change. In 2026 so far Google has run core updates in March and May, and spam updates in March, June, August and September. The September 2026 spam update took 13 days and 16 hours to roll out.
A core update reweighs the quality signals across the web. A drop after one rarely has a single cause, so measure the pages and searches that moved before you change anything.
Do the same ranking factors apply to AI Overviews and AI Mode?
Yes. Google says there are "no additional technical requirements" to appear in AI Overviews or AI Mode. Indexed, and eligible for a snippet. That's it.
Google's AI features guidance says that "all existing SEO fundamentals continue to be worthwhile". It lists crawl access, internal links, page experience, text content and structured data that "matches the visible text on the page". It also explains "query fan-out", where AI Mode runs several related searches across subtopics to build one answer.
That changes one thing for B2B teams. One buyer question can trigger searches for pricing, integrations and alternatives at once. A site with a clear page for each gets more chances to appear. Want the AI side in more depth? Our guide on how to rank in AI Overviews covers it.
How do you measure whether ranking work is paying off?
Track rankings and impressions in Search Console, then tie them to enquiries in your analytics. Rankings alone say little about pipeline.
Check the tracking first. Here's what we found on one client account. A client's four enquiry types never counted as conversions. Their event names ran 41 to 50 characters; GA4 cuts at 40. The 40-character "career" event counted every time. Nobody had noticed, because the events still showed up.
That account would have judged its SEO on the wrong numbers for months. So before you report on ranking work, confirm three things:
- Every enquiry, demo and contact form fires a key event in GA4, and every event name stays within GA4's 40-character limit.
- Search Console covers every host and subdomain that carries selling pages.
- You compare matched periods, such as the last 28 days against the 28 before, rather than one noisy week against another.
Where does Searchmaxxed fit?
Searchmaxxed is an SEO and AI SEO agency for enterprise, B2B SaaS, venture-backed, manufacturing and public-sector companies. We do technical SEO, on-page SEO, content, link building, digital PR, GEO and AEO, so you rank in Google and get named in AI answers.
Every factor in this article maps to work we ship each month:
| Ranking factor | What we do about it |
|---|---|
| Indexing and rendering | Fix crawl, rendering, canonicals, sitemaps and redirects |
| Relevance and intent | Build one page per buyer search: comparisons, alternatives, integrations, industries |
| Helpful, original content | Write answer-first pages from your own data, specs and customer numbers |
| Links and mentions | Digital PR and qualified link work, scaled to the package |
| The click | Title and snippet changes where you rank but don't get clicked |
| Page experience | Core Web Vitals work on the templates that carry traffic |
| AI answers | Track what ChatGPT, Gemini, Perplexity and AI Overviews say about you |
Here's how we do it. We build on a website you own, our Agentic Website. AI agents pull your Search Console and GA4 data, read the Google results and check what AI answers say about you and your competitors. That data feeds the Managed Search Loop, which shows which page, fix or link should bring in the most enquiries next. We build it, check the live page and measure what changed. Off your site, the Off-Page Source Layer builds the links, digital PR and mentions that back your pages up.
| Package | What it includes |
|---|---|
| Growth System | 30 page releases a month, landing-page work on up to four existing pages, one qualified link action a month, AI visibility tracking |
| Agentic Search Engineering | 50 page releases a month, comparison, industry and integration page families, monthly press release, continuous digital PR and link campaigns, AI answer tracking |
| Custom & Enterprise | Multiple markets, dedicated engineering, custom reporting and original first-party research |
A page release is an approved new page or a substantial rebuild of one. Every managed package includes ongoing audits, a 90-day plan in weekly releases and a monthly decision memo. You own the site, repository, content and data from day one. The SEO services page sets out the full scope.
Before Searchmaxxed, I built Radiant Roof Repairs from zero to more than A$1 million a year in revenue using SEO as the only marketing channel. Three high-intent Gold Coast roofing searches reached number one within 90 days. I did the pages, research, competitor checks and updates manually.
Searchmaxxed's work has generated more than A$10 million in revenue through search for its clients.
Is Searchmaxxed right for you?
We suit marketing and growth teams at enterprise, B2B SaaS, venture-backed, manufacturing and public-sector companies that want the ranking work built and shipped, not described. If you want advice and a slide deck, we're the wrong call.
| Searchmaxxed is a good fit if | Look elsewhere if |
|---|---|
| You want pages built and shipped each month, counted as page releases | You only want a one-off fix, not monthly work |
| You compete in a crowded category and need links and digital PR | You need search to fill this month's pipeline on its own |
| You want SEO and AI search from one team | You want reports about the work more than the work |
| You want to own your site, repository, content and data | You want strategy advice with no changes to your site |
| You want direct access to the people doing the work | You're shopping for the cheapest retainer |
If you'd rather test the fit first, the AI SEO Audit takes five business days. It maps where buyers search without finding you, what the site is doing to itself, and every fix ranked by what it's worth.
Frequently asked questions
How many ranking factors does Google have?
Google doesn't publish a number. A January 2025 DOJ trial exhibit says Google "uses over 100 raw signals", combined into top-level signals. The "200 factors" lists online compile confirmed factors, patents and guesses.
Is PageRank still used?
Yes. Google's ranking systems guide says PageRank "continues to be part of our core ranking systems", though it has "evolved a lot" since launch. A 2025 trial exhibit says it still feeds Google's quality signals for pages.
What are the three stages of Google ranking?
Crawling, indexing and serving. Googlebot finds pages, Google stores and analyses them in its index, then its ranking systems order indexed pages for each search. A page that isn't indexed can't rank.
Can you provide a list of Google's ranking algorithms?
Google's ranking systems guide names these: BERT, deduplication, exact match domain, freshness, PageRank and link analysis, neural matching, original content and passage ranking. It adds RankBrain, reliable information, reviews, site diversity and spam detection. Helpful content, Panda and Penguin no longer run as separate systems, because Google folded each one into its core ranking systems over the years.
Is 75 a good SEO score?
An SEO score comes from a third-party tool. Google uses no such score, so a 75 means whatever that tool's checklist says it means, and nothing more. Judge pages by indexing, rankings, clicks and enquiries instead.
Does domain authority affect Google rankings?
Domain Authority is a third-party metric. Google doesn't use it. The 2024 leak did describe a Google attribute called siteAuthority. Nobody outside Google knows how it works, so build the links and topic focus it probably reflects.
Does AI-generated content rank on Google?
It can. Google's spam policy targets pages made mainly to manipulate rankings "no matter how it's created". Pages that add original value can rank. Mass-produced pages with nothing new fall under scaled content abuse, whatever tool or person produced them.
How long do Google ranking changes take?
Indexing can happen within a day or two. Ranking takes longer, because links and interaction data build over months, and trial evidence says Navboost uses 13 months of click data. Core updates add their own swings while they roll out over roughly two weeks.
What should you do next?
Open Search Console and check that your product, solution and integration pages are indexed. Then list the ten searches your best buyers make and check whether each one has its own page, matched to what the buyer wants. If you'd like us to find the gaps and build the pages, let's talk pricing.