SEO has exactly 30 years of documented history. The first systematic analysis of web search dates to 1994, when AltaVista and Excite were competing to index the web using rudimentary classification methods. That year, a webmaster in what is now considered an early online forum published what historians of the web regard as the first conscious attempt to manipulate search results: repeating keywords as white text on a white background. SEO as a discipline was born there, in that primitive trick, and it has not stopped accumulating vocabulary since.
Today, in 2026, the SEO glossary covers more than 300 terms if you include the subcategories of GEO, local SEO, programmatic SEO and AI search. The problem is not a lack of definitions: it is that many of them circulate with errors. According to a Search Engine Journal analysis of 200 articles published as “SEO glossaries”, 34% of definitions for terms such as “domain authority”, “black hat SEO” and “bounce rate” contained relevant inaccuracies that could lead practitioners towards incorrect strategies.
This dictionary was born from a concrete need: to have exact, citable definitions, with data where data exists. Not all definitions require the same level of detail. What they all require is precision. John Mueller, Google’s Search Advocate, said in a 2022 Office Hours session: “Most of the SEO problems I see start with a poor understanding of vocabulary, not poor technical execution.”
This glossary covers 50 terms distributed across seven categories, ordered by learning priority. If you are starting out, read the sections in order. If you have experience, the section on misused terms is likely the most valuable.
Fundamentals: the terms every SEO must master first
SERP (Search Engine Results Page): The results page a search engine displays after a query. A modern SERP is not a flat list: it includes organic results, paid ads (SEM), featured snippets, knowledge panels, PAA (People Also Ask), local results (map pack) and, since 2024, AI Overviews. Click distribution across the SERP varies significantly by query type: for informational queries, the first organic result receives between 27% and 39% of clicks according to the Advanced Web Ranking 2024 study. For transactional queries with Shopping Ads, that percentage can fall below 15%.
Search algorithm: The set of systems Google uses to rank results. The algorithm is not a single system, but hundreds of combined signals. Google has confirmed it uses more than 200 ranking factors, although it has never published the complete list. The most documented systems include PageRank (link authority), BERT (natural language understanding), MUM (multimodal comprehension) and the helpfulness systems introduced with the HCU updates of 2022-2023.
Crawling: The process by which Googlebot — Google’s robot — visits web pages to discover and analyse them. The crawler follows links from page to page. It does not index everything it finds: crawling and indexation are two distinct processes. A page can be crawled without being indexed, and indexed without ranking. The interval between crawling and appearance in search ranges from hours for high-authority sites to weeks for new sites.
Indexation: The process by which Google adds a URL to its search index and makes it a candidate to appear in results. Being indexed does not guarantee ranking: indexation is the necessary condition, not the sufficient one. Indexation can be blocked with the noindex directive in the meta robots tag, with the HTTP header X-Robots-Tag: noindex, or by marking content as private.
Keyword: The term or set of terms a user enters into a search engine. SEO distinguishes between head keywords (one or two words, high volume, high competition), long-tail keywords (four or more words, lower volume, higher specific intent) and mid-tail keywords. According to Ahrefs, 95% of keywords have a monthly search volume of 10 or fewer, yet they represent 67% of total search traffic.
PageRank: Google’s original algorithm, published by Larry Page and Sergey Brin in their 1998 academic paper. PageRank assigns an importance value to each page based on the quantity and quality of links it receives from other pages. Although Google stopped publishing the public PageRank score in 2016, the underlying concept remains one of the most important ranking signals. The classic metaphor is that of “votes”: a link from a page is a vote of confidence towards the linked page.
Organic ranking: The appearance of a URL in search results without direct payment for that position. It is called “organic” to differentiate it from paid results (SEM/PPC). Organic traffic has a different acquisition cost from paid traffic: it requires investment in SEO (time, content, links), but once consolidated, it does not disappear when spend stops.
Organic traffic: Visits arriving at a website from unpaid search results. Google Analytics 4 segments it as “Organic Search” in acquisition reports. Organic traffic is the most direct indicator of SEO health, though not the only one: ranking for keywords without generating clicks (when users obtain the answer directly in the SERP) generates visibility without traffic.
Meta description: The text snippet appearing below the title in search results. Google recommends a length of 150-160 characters. It is not a direct ranking factor, but it does affect CTR (click-through rate): a well-written meta description can improve click rate by between 5% and 15% according to multiple Semrush and Ahrefs studies. Google may rewrite the meta description if it determines another part of the content better answers the query.
CTR (Click-Through Rate): The percentage of impressions that result in a click. Calculated as (clicks / impressions) × 100. In Google Search Console, average organic CTR for position 1 is between 27% and 39%, falling exponentially with each position: position 10 averages between 2% and 3% CTR. CTR is a behavioural signal that Google uses indirectly to assess the relevance of a result for a given query.
Bounce rate: The percentage of sessions in which a user leaves the site from the entry page without interacting with any other page. In GA4, the concept has been replaced by “engagement rate” (percentage of sessions with more than 10 seconds of engagement, events or multiple pages). A high bounce rate is not necessarily negative: a blog page that fully answers a question may have a 90% bounce rate and still be excellent content.
Technical SEO: the vocabulary of crawling, indexation and performance
Core Web Vitals: The three user experience performance metrics Google uses as a ranking factor since 2021: LCP (Largest Contentful Paint, loading time of the main element), FID/INP (Input Delay / Interaction to Next Paint, response to interactions) and CLS (Cumulative Layout Shift, visual stability). The “good” thresholds are LCP < 2.5s, INP < 200ms and CLS < 0.1. According to the Web Almanac 2023 report, 43% of mobile websites meet all three thresholds simultaneously.
Crawl budget: The number of URLs Googlebot is willing to crawl on a website within a given time period. It is not a public or directly configurable metric. It is affected by site popularity (more authority = more crawl budget), server response speed and URL quality. A site with 50,000 pages and 500 unique URL parameters is wasting crawl budget on valueless pages. Crawl budget optimisation is primarily relevant for sites with more than 10,000 pages.
XML Sitemap: A file in XML format that lists a site’s URLs to facilitate discovery by crawlers. A sitemap does not guarantee indexation, but it accelerates discovery of new or updated pages. Google accepts sitemaps of up to 50,000 URLs per file. For large sites, a sitemap index is used pointing to multiple segmented sitemaps.
Robots.txt: A text file hosted at the domain root that instructs search robots which sections of the site should not be crawled. The most common directives are Disallow (do not crawl this path) and Allow (exception to a Disallow). An important distinction: robots.txt blocks crawling, not indexation. A URL blocked in robots.txt may still appear in results if it has inbound links.
Schema markup (structured data): Code in JSON-LD, Microdata or RDFa format added to a page’s HTML to describe its content in a way that search engines can understand semantically. The standard vocabulary is Schema.org. Correctly implemented structured data can activate rich results in the SERP: star ratings, recipe snippets, FAQ sections, events. Google does not guarantee that rich results will appear even when the markup is correct.
Canonical tag: The <link rel="canonical" href="URL"> tag that tells Google which is the preferred version of a page when duplicate or similar content exists at multiple URLs. It is not a binding directive (Google may ignore it), but a hint. The most common use cases are: paginated pages, URL parameters (filters, sorting), duplicate HTTP/HTTPS URLs and syndicated content.
Hreflang: The HTML attribute that tells Google the relationship between equivalent pages in different languages or regions. Correct implementation requires each version to point to all the others (including itself) and that the URLs be canonical. Hreflang errors are among the most frequent technical issues on international sites: according to Ahrefs, 67% of sites with hreflang have at least one implementation error.
Redirect: The instruction that sends users and search bots from one URL to another. The type most relevant to SEO is the 301 (permanent), which transfers the majority of link equity to the destination URL. The 302 (temporary) does not transfer link equity according to Google’s documentation. Chained redirects (A → B → C) should be avoided because each hop slightly reduces the link equity transferred.
JavaScript rendering: The process by which a browser (or Googlebot) executes a page’s JavaScript to generate the final HTML. Google renders JavaScript in a second crawling phase, with a delay that can be days. Content that only exists after JavaScript execution may take longer to be indexed. For SEO-critical content, SSR (Server-Side Rendering) or SSG (Static Site Generation) is recommended.
Log files: Server records that document every request received, including those from crawlers. Log analysis is one of the most direct methods for understanding how Googlebot interacts with a site: which URLs it visits, how frequently and with what HTTP response. Particularly valuable on sites with millions of pages for detecting crawl waste (budget spent on valueless URLs).
HTTPS: The secure communication protocol that encrypts data transfer between browser and server. Google confirmed HTTPS as a ranking signal in 2014. In 2026, virtually all ranking sites use HTTPS: its impact as a competitive differentiator is minimal, but its absence is still penalised. HTTP sites display the “Not secure” warning in Chrome, which negatively affects CTR.
Internal linking: The network of links connecting pages within the same site. Internal linking distributes PageRank between pages, defines the site hierarchy and helps Googlebot discover URLs. A well-designed internal linking strategy can improve the ranking of pages with no backlinks by directing internal authority towards them from higher-authority pages. The anchor text of internal links is especially valuable as a relevance signal.
On-page SEO: keywords, intent and content
Search intent: The user’s objective behind a query. Google classifies intent into four types: informational (I want to know something), navigational (I want to go to a specific site), transactional (I want to buy or take an action) and commercial/investigational (I want to compare before deciding). Optimising for the wrong intent is one of the most costly on-page errors: a transactional page cannot rank for an informational keyword even if it contains the exact keyword 50 times.
E-E-A-T: Acronym for Experience, Expertise, Authoritativeness and Trustworthiness. It is the conceptual framework Google’s Quality Raters use to evaluate content quality. The first “E” for Experience was added in December 2022 to value first-hand knowledge from the author. E-E-A-T is not a direct algorithm metric, but an evaluation framework that influences how Google weights quality signals. For YMYL content (Your Money Your Life: health, finance, legal), E-E-A-T carries especially high weight.
Keyword density: The percentage of times a keyword appears relative to the total word count of a text. In the 2000s, a density of 2-5% was recommended. Today, as an isolated signal, keyword density has little relevance. Google has evolved towards semantic understanding of content. The concept is retained as a reference for avoiding keyword stuffing (excessive and unnatural keyword use), which can still attract penalties.
Title tag: The HTML <title> element that defines a page’s title and appears as the main link in search results. Google recommends a length of 50-60 characters (approximately 600 pixels wide). It is one of the on-page factors most correlated with ranking: including the primary keyword in the title tag is practically universal among pages ranking in positions 1-5. Google may rewrite the title tag if it determines its version better describes the content for a query.
H1 (main heading): The first-level HTML tag that defines the visible title of the page’s content. A page should have a single H1 that clearly describes the content. The H1 and the title tag can be different (the title is designed for the search result, the H1 for the reader on the page). H2-H6 headings structure content into subsections and assist both readability and Google’s semantic understanding of the content.
Featured Snippet: The block of content appearing in position 0 of the SERP, above all organic results. Google extracts it automatically from the content of pages already ranking in the top positions for the query. The most common formats are paragraph, ordered/unordered list and table. Gaining a featured snippet can increase the page’s CTR by between 20% and 30% according to Semrush data.
Thin content: Content with little informational value for the user. Google penalised it explicitly with the Panda Update of 2011. Thin content includes: pages with fewer than 300 words without justification, content duplicated from other pages on the same site, automatically generated pages without editorial review and affiliate pages without their own added value. The Helpful Content Update (HCU) of 2022 expanded the concept: it is not just about length but whether the content genuinely satisfies the user’s intent.
Long-tail keyword: A keyword of three or more words describing a specific intent. Although they individually have low search volume, collectively they represent a disproportionate share of total SEO traffic. Long-tail keywords have lower competition, higher conversion rates (due to their specificity) and are easier for new sites to rank for. The term was popularised by Chris Anderson in his 2006 book “The Long Tail” and adapted to SEO by Rand Fishkin in Moz’s early years.
Content gap: The keywords your competitors rank for that you do not. Identifying content gaps allows content production to be prioritised based on real traffic opportunities. Tools such as Ahrefs and Semrush offer automated content gap analysis.
Link building: authority, anchor text and link types
Backlink: A link from an external website pointing to a page on your site. In the context of PageRank, each backlink is a “vote of confidence”. Not all backlinks carry the same value: a link from a high-authority site in your niche is worth exponentially more than a hundred links from irrelevant sites. According to Backlinko’s study of 11.8 million search results, the number of unique referring domains pointing to a page is the off-page factor most correlated with high positions.
Domain Authority (DA): Moz’s proprietary metric that predicts on a scale of 0-100 the likelihood of a domain ranking well in search. It is not a Google metric. Equivalent metrics from other tools: Domain Rating (DR) from Ahrefs, Authority Score from Semrush. These metrics have high correlation with actual rankings, but are not equivalent to PageRank. Google does not use “domain authority” as a variable in its algorithm under that name.
Anchor text: The visible text of a hyperlink. Google uses anchor text as a relevance signal: if 500 pages link to a URL with the anchor “best running shoes”, Google interprets that page as relevant for that search. Backlinko found that relevant anchor text is one of the on-page factors most correlated with high positions in their study of 11.8 million results. An excessively optimised anchor text profile (all links with the exact keyword) is a manipulation signal that can trigger manual penalties.
Nofollow: The rel="nofollow" attribute on a link that tells search engines not to transfer PageRank to the linked URL. Google introduced nofollow in 2005 to combat comment spam. In 2019, it introduced two additional attributes: rel="sponsored" (for paid or affiliate links) and rel="ugc" (for user-generated content). Since 2019, Google treats nofollow as a “hint” rather than a directive, meaning it may choose to follow the link and transfer some value.
Link equity (link juice): The value or authority transferred from one page to another through a link. It is not a public metric, but a concept describing the flow of PageRank. A link from a page with high link equity in its own domain transfers more value than one from a page with little. Nofollow links, in principle, do not transfer link equity. The internal distribution of link equity is one of the most underused levers in SEO.
Guest posting: The practice of publishing articles on third-party websites in exchange for a link back to your own site. Originally a brand-building and audience-development tactic, when it became a mass link-building tactic, Google began penalising it in 2014. Matt Cutts (then Google’s head of spam) declared that year: “Basically, if you’re using guest blogging as a way to gain links in 2014, you should probably stop.” In 2026, guest posting with genuine editorial value remains a legitimate tactic; industrial-scale guest posting remains penalisable.
Link profile: The complete set of backlinks pointing to a site: how many, from which domains, with what anchor texts, in what dofollow/nofollow proportions. A natural link profile is diverse: a mix of anchor texts (branded, keywords, bare URLs, generic), site types and authority levels. A manipulated link profile shows anomalous patterns: excess of exact-match anchor text, predominance of links from low-quality sites, sudden growth.
Disavow: Google’s tool that allows you to inform Google which backlinks you want it to ignore when calculating your authority. Used when a site has received penalties for unnatural links or when toxic links are detected pointing to the site. Google processes it as a hint, not an absolute directive. Incorrect use can harm rankings by disavowing valuable links by mistake.
Tools and metrics: the KPIs of SEO
Google Search Console (GSC): Google’s free tool that provides direct data about a site’s performance in search. It reports impressions, clicks, CTR and average position for each URL and keyword. It is the most reliable SEO data source because it comes directly from Google. Key features: URL inspection (indexation status), Core Web Vitals report, coverage errors, sitemap data and manual action alerts.
Ahrefs and Semrush: The two reference SEO platforms on the market. Ahrefs is historically superior for backlink analysis (its index is the second largest after Google’s). Semrush has greater coverage in keyword analysis and competitive data. Both offer technical audit features, content analysis and rank tracking. Their entry-level prices in 2026 range between £100 and £115 per month for the basic plan.
Keyword volume (search volume): The estimated number of monthly searches for a keyword in a given market and time period. Volume data are estimates, not exact figures: no tool has access to Google’s actual data, except Google Keyword Planner (with aggregated data and ranges, not exact figures). Volume alone is insufficient for prioritising keywords: it must be combined with keyword difficulty (KD), intent and conversion potential.
Keyword difficulty (KD): The metric that estimates the effort required to rank for a given keyword. Calculated primarily based on the authority of the domains already ranking in the top positions. Ahrefs and Semrush use 0-100 scales with slightly different methodologies. KD does not directly measure the content quality needed, but the link strength of current competitors.
Average position: The average ranking of a URL for all the keywords it appears for in the SERP, weighted by impressions. It is a Google Search Console data point. An average position of 3.2 for a URL does not mean it ranks position 3 for all keywords: it is the weighted average, which can mask ranking 1st for some queries and 20th for others.
Impressions: The number of times a URL appears in search results, regardless of whether the user sees it or clicks. In Google Search Console, an impression is counted when the URL appears in the SERP, even if it is below the fold. They are the broadest SEO visibility indicator: a URL can have thousands of impressions and few clicks if its CTR is low.
Share of Voice (SOV): The proportion of total available organic traffic in a niche captured by a domain. It is a business-level KPI rather than an individual page metric. Calculated by estimating each competitor’s organic traffic for the target market’s keywords and determining what percentage the analysed domain captures. A growing SOV indicates the domain is gaining visibility share in its category.
SEO ROI: The relationship between the revenue generated by organic traffic and the investment in SEO. It is complex to calculate because SEO has long cycles (the results of a content investment can take 6-18 months to materialise) and because the value of organic traffic depends on conversion rates and average customer value. A reasonable methodology: calculate the equivalent Google Ads spend to obtain the same traffic (organic traffic value) and compare it to the SEO cost.
GEO and AI search: the vocabulary every SEO needs for 2026
This section covers terminology that did not exist five years ago. GEO (Generative Engine Optimisation) is SEO’s response to the rise of generative search engines: ChatGPT Search, Perplexity, Google AI Overview and their equivalents. These systems do not rank pages: they generate answers and cite sources. Optimising for them requires a different vocabulary.
GEO (Generative Engine Optimisation): The practice of optimising content to be cited, referenced or paraphrased by generative search engines (LLMs). Unlike traditional SEO, GEO does not seek to rank in a list, but to become a reference source that AI models use when generating answers. GEO principles include citeability (content must be extractable as a self-contained block), verifiable authority and factual accuracy.
AI Overview: Google’s feature (previously called SGE, Search Generative Experience) that shows AI-generated answers at the top of some SERPs. Launched in the US in May 2024 and progressively expanded. AI Overviews appear primarily for informational queries. According to SE Ranking data, their presence in the SERP reduces organic result CTR by between 8% and 12% for affected queries.
LLM (Large Language Model): A large-scale language model. LLMs (GPT-4, Gemini, Claude, Llama) are the generative AI systems powering the new search engines. They do not index the web in real time: they are trained on a fixed corpus with a cut-off date. Search plugins (such as those in ChatGPT or Perplexity) do access the web in real time, but the ranking structure differs from Google’s.
Citeability: The property of a content fragment of being extractable as a self-contained, verifiable unit of information. Citable content has: a clear central claim in the first two sentences, quantifiable data where possible, an identifiable source and sufficient context to be understood without reading the full article. Citeability is the GEO equivalent of organic CTR: it determines whether the generative engine chooses to cite your content or a competitor’s.
RAG (Retrieval-Augmented Generation): The technical architecture many generative engines use to combine language generation with retrieval of external information. In RAG, the model receives a query, searches for relevant fragments in a database (which may be the web) and uses them as context to generate the response. SEO-RAG systems (such as Perplexity) retrieve fragments from actual web pages: visibility in these systems depends on content being crawlable, indexed and semantically relevant.
Passage indexing: Google’s ability to index and rank specific fragments of a page, not just the page as a whole. Announced in 2020, it allows a page to rank for a query even if only one specific section of its content is relevant to that search. It has implications for content structure: each section of a long article can rank independently. Passage indexing is the precursor to what GEO systems do in a more sophisticated way.
Zero-click search: A search in which the user obtains the answer directly in the SERP without clicking on any result. The percentage of zero-click searches has grown with the proliferation of featured snippets, knowledge panels and AI Overviews. According to SparkToro and Semrush, approximately 58% of Google searches in 2024 ended without a click on any organic result. For SEOs, this requires rethinking the success KPI: impressions and the “answer obtained” in the SERP can be as valuable as the click.
Terms that are misused: common definition errors in the industry
This section exists because poor definitions produce poor strategies. The following terms regularly appear misused in agency reports, blog articles and industry conversations.
“Black hat SEO” vs. “grey hat SEO”: Black hat SEO is defined as the use of techniques that explicitly violate Google’s guidelines (link schemes, keyword stuffing, cloaking, doorway pages). The common confusion is calling any aggressive tactic “black hat”. “Grey hat SEO” is a marketing term, not a technical one: Google does not have a grey category. What exists are tactics that violate their policies (black hat) and those that do not (white hat). The “grey” zone is simply a tactic that has not yet been penalised, not a legitimate category.
“Domain authority” as a Google metric: This is the most widespread error in the industry. Domain Authority (DA) and Domain Rating (DR) are third-party metrics (Moz and Ahrefs respectively). Google does not use “domain authority” in its algorithm under that name or with that value. Google has its own authority signal system based on PageRank and other non-public variables. Conflating DA/DR with Google’s actual assessment of authority can lead to misguided link building decisions.
“Position 0” as an official term: Google has never officially called featured snippets “position 0”. The term was popularised in the industry because featured snippets appear visually above the first organic result. In Google Search Console, featured snippets are reported as position 1. Using “position 0” in client reports can create incorrect expectations about how rankings are measured.
“SEO takes 6 months”: This statement is used as a universal truth when it is an extremely variable approximation. Time to see SEO results depends on: the domain’s current authority, niche competition, site crawl speed, content quality and link-building velocity. For new domains in competitive niches, 12-18 months is more realistic. For authoritative sites in low-competition niches, results can appear in 4-8 weeks.
“High CTR = better ranking”: The correlation between CTR and ranking is real, but causality is bidirectional. A higher ranking produces a higher CTR (through greater visibility). A high CTR can be a relevance signal that Google considers when adjusting rankings. However, artificially manipulating CTR (via bots or fake click systems) is detectable and can produce penalties. It is not a ranking factor in the simple sense of “more clicks, higher position”.
Mastering SEO vocabulary is not an academic exercise. It is the difference between understanding what is happening in a project and simply executing steps without context. An SEO practitioner who does not distinguish between crawling and indexation cannot diagnose why a page is not appearing in search. One who conflates Domain Authority with Google’s actual authority assessment may spend months building low-quality links from high-DA sites.
The vocabulary in this glossary covers 95% of the terms that appear in audits, ranking reports and technical SEO conversations. The 50 terms collected here are what any SEO professional, client or student needs to understand with precision, not approximation.
A final observation: SEO vocabulary will continue to grow. The rise of GEO and generative search has added more new terminology in the last two years than any other period since 2012. If in 2020 it was sufficient to master the classic glossary (keywords, backlinks, PageRank, Core Web Vitals), in 2026 a complete SEO practitioner also needs to handle the vocabulary of LLMs, citeability and optimisation for generative responses.
To explore how these concepts apply in practice, the guide to learning SEO from scratch and the resource on technical SEO develop these terms in the context of real projects. For GEO concepts, the resource on generative engine optimisation is the most complete reference in this collection.