The SEO & AI Search Glossary

An A/B test splits your live traffic between two versions — a control (A) and a variant (B) — and measures which produces more of the action you want. It replaces "I think this is better" with "the data shows this is better".

"Above the fold" is the portion of a web page a visitor sees before scrolling — the term comes from newspapers, where the most important story sat above the physical fold. On the web it is the first screen, and it is where you win or lose most visitors, who decide in seconds whether the page is relevant.

An AI crawler is an automated bot that fetches web pages for artificial-intelligence systems rather than for a classic search index. Some gather content to train models (like GPTBot), and some fetch live pages to answer a user’s question in real time (like OpenAI’s or Perplexity’s user-facing fetchers). They are the plumbing behind AI answers that cite the web.

AI Overviews are the AI-written summaries Google now places at the top of many results pages (they grew out of the Search Generative Experience). Google stitches an answer together from several sources and links to them. For some queries this answers the question outright, so the user never clicks — the "zero-click" effect — while creating a new prize: being one of the few sources cited inside the overview.

Alt text (the alt attribute) is a written description of an image in the HTML. Screen readers read it aloud for visually impaired users, and it displays if an image fails to load. It is also how search engines understand what an image shows, since they cannot fully "see" pictures.

Anchor text is the clickable, visible text of a link. Search engines read it as a description of what the linked page is about, so "crawl budget guide" tells them far more than "click here". This applies to both internal links and backlinks.

AEO is about winning the single answer rather than a ranked list — the response a voice assistant reads aloud or an answer engine displays directly. It reframes the goal from "where do we rank?" to "are we the answer?" It overlaps heavily with featured-snippet optimisation and with GEO, but the framing is answer-first.

MRR (monthly recurring revenue) and ARR (annual recurring revenue) are the predictable, repeating subscription revenue a SaaS business earns each month or year. They are the heartbeat metrics of any subscription company because, unlike one-off sales, recurring revenue compounds and can be forecast.

Attribution is how you assign credit for a conversion to the marketing touchpoints that led to it. Because buyers rarely convert on their first visit, the model you choose — who gets the credit — dramatically changes how each channel looks.

A backlink is simply a link from another website pointing to a page on yours. Search engines treat these as votes of confidence: if many credible sites link to a page, it is probably worth trusting. Backlinks were the original breakthrough behind Google’s PageRank and remain one of the most powerful ranking signals.

Bottom-of-funnel content targets people who are close to a decision: "X vs Y" comparisons, "best tools for Z" roundups, "[competitor] alternatives", pricing explainers, and specific use-case pages. Search volume is lower than for broad awareness topics, but intent and conversion rate are far higher — these are prospects with a wallet out.

Bounce rate is the percentage of visits where someone views a single page and leaves without any further interaction. It sounds like a quality score but it is easy to misread — a high bounce can be perfectly healthy or a serious problem depending entirely on the page’s job.

A call to action is the prompt that tells a visitor exactly what to do next — start a trial, book a demo, download a guide. It is the hinge between content and conversion; without a clear one, even an engaged reader has nowhere to go.

The same content can often be reached at several URLs — with tracking parameters, http vs https, trailing slashes, or print versions. To a search engine those look like separate pages competing with each other. A canonical tag placed in the page head names the one "master" URL you want ranked, so the duplicates hand their signals to it instead of splitting them.

Churn is the rate at which customers cancel or stop paying over a given period. In a subscription business it is the counterweight to growth: even strong new-signup numbers do not add up if customers leave just as fast. It is usually expressed as a percentage of customers (or revenue) lost per month or year.

Click-through rate is the share of people who click your result after seeing it: clicks divided by impressions. In Search Console you can see it for any query or page. It tells you how compelling your listing is relative to how often it shows up.

Content decay is the slow decline in traffic and rankings that most pages experience over time. It is normal: information goes out of date, competitors publish something better, and search intent shifts. A page that ranked well two years ago can quietly fade without anyone noticing.

A content gap analysis compares your site’s ranking footprint against competitors to surface the demand you are missing — keywords they rank for and you do not, or topics where you sit stranded on page three. It turns "what should we write next?" from a guessing game into a prioritised, demand-backed list.

A conversion is any action you have defined as valuable being completed: a purchase, a demo request, a trial signup, a newsletter subscription, a form fill. It is the point where a visitor does the thing you actually wanted them to do, rather than just browsing.

Conversion rate is the share of visitors who complete the action you care about — a signup, a demo request, a purchase — calculated as conversions divided by visitors or sessions. It is where traffic turns into value; traffic that never converts is just cost.

CRO is the disciplined practice of increasing the share of visitors who take a desired action, by improving the page: its copy, layout, offer, friction, and trust signals. Instead of buying more traffic, you get more value from the traffic you already have.

Core Web Vitals are three metrics Google uses to quantify how a page feels to a real user. Largest Contentful Paint (LCP) measures loading — how fast the main content appears. Interaction to Next Paint (INP) measures responsiveness — how quickly the page reacts when you tap or click (it replaced First Input Delay in 2024). Cumulative Layout Shift (CLS) measures stability — how much the layout jumps around as it loads.

Search engines send automated bots (crawlers) to fetch your pages before they can rank them. Crawl budget is simply the practical ceiling on how many of your URLs Googlebot will fetch in a given period. It is governed by two things: how hard Google can hit your server without slowing it down (crawl rate limit), and how much Google actually wants your content (crawl demand, driven by popularity and freshness).

Customer acquisition cost is everything you spend on sales and marketing to win a customer, divided by the number of customers won in a period. It is the metric executives use to judge whether a growth channel is worth funding.

Customer lifetime value (LTV or CLV) is the total revenue a customer generates over the whole time they stay with you. It matters because acquisition cost only makes sense in relation to it: spending money to acquire a customer is smart only if that customer is worth more over their lifetime than they cost to win.

Domain Authority (DA) is a score from 0 to 100, created by Moz, that predicts how likely a website is to rank. Similar scores exist from other tools (Ahrefs’ Domain Rating, Semrush’s Authority Score). Crucially, none of these are used by Google — they are third-party estimates built mostly from backlink data.

Duplicate content is the same, or very similar, content appearing at more than one URL — whether across your own site or copied between sites. It confuses search engines about which version to rank and can split the signals that would otherwise concentrate on one page. There is no single "duplicate content penalty", but it dilutes performance.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trust. It is the framework in Google’s Search Quality Rater Guidelines for judging whether content deserves to be shown — especially for "Your Money or Your Life" topics like health, finance, and safety, where bad information causes real harm. The first E, Experience, was added to reward genuine first-hand knowledge.

Engagement rate is a GA4 metric: the percentage of sessions that were "engaged", meaning they lasted longer than 10 seconds, had a conversion, or included multiple page views. It is the positive mirror image of bounce rate (which GA4 now defines as simply the inverse) and a more useful way to think about whether visits had substance.

An entity is a distinct, uniquely identifiable thing: a person, company, place, product, or concept. Search engines increasingly understand the world as a web of connected entities (a knowledge graph) rather than just strings of text — so "Apple" the company is a different entity from "apple" the fruit, and each has known relationships to others.

Faceted navigation is the system of filters and sorting options on category pages — size, colour, price, brand — that let users narrow down listings. It is great for shoppers, but each combination of filters can create a unique URL, so a handful of facets can silently generate hundreds of thousands of crawlable pages.

A featured snippet is the answer box that sometimes appears at the very top of the results, above the normal listings. Google extracts a concise answer — a paragraph, list, or table — from a page that already ranks and displays it with a link back to the source. It is often called "position zero".

Friction is anything that makes it harder or more effortful for a visitor to do what you want them to do — a long form, a confusing layout, an unexpected cost, a slow page, an unclear next step. Every point of friction is a reason to give up, and reducing it is one of the most reliable ways to lift conversion.

GEO is the practice of getting your content cited and surfaced inside AI-generated answers — the responses from ChatGPT, Perplexity, Google’s AI Overviews, and similar engines. Where classic SEO optimises for a ranking position in a list, GEO optimises for being quoted, referenced, or named as the source inside a synthesised answer.

Google Analytics 4 is Google’s current analytics platform, which replaced Universal Analytics in 2023. It measures what people do once they reach your site — pages viewed, actions taken, conversions completed — using an event-based model where every interaction is an event, rather than the old session-and-pageview model.

Google Search Console is a free tool from Google that shows how your site is doing in search. It reports the queries you appear for, your clicks, impressions, average position, and CTR, plus which pages are indexed and any technical issues Google found. It is first-party data straight from the source, unlike third-party estimates.

Grounding is the practice of anchoring an AI system’s answer in real, retrieved sources rather than letting it rely only on what it memorised during training. A grounded answer is built from documents the system fetched at query time, which is why AI search engines can cite links — those citations are the grounding.

A hallucination is when an AI language model produces information that sounds confident and plausible but is actually false or made up — a fake statistic, a misattributed quote, a nonexistent source. Because models generate fluent text by predicting likely words, they can state wrong things as smoothly as right ones.

Header tags (H1 through H6) are HTML elements that create the heading hierarchy of a page — the H1 is the main title, H2s are major sections, H3s are subsections, and so on. They give content a clear structure that readers skim, screen readers navigate, and search engines use to understand what the page covers.

Hreflang is a snippet of code that tells search engines a page has versions for different languages or regions, and which is which. It helps Google serve the right version to the right user — the Spanish page to Spanish speakers, the UK page to UK searchers — instead of the wrong-language result.

Every time a browser or crawler requests a URL, the server answers with a three-digit HTTP status code. They group into families: 2xx means success (200 = OK), 3xx means redirection (301, 302), 4xx means a client-side problem (404 not found, 410 gone, 403 forbidden), and 5xx means a server-side problem (500 internal error, 503 unavailable).

An ideal customer profile is a clear definition of the type of customer your product serves best — the ones who get the most value, stay longest, and are most profitable. In B2B it describes company attributes (industry, size, tech stack, needs); it is about the account you most want, not a single persona.

An impression is counted each time one of your results appears in the SERP for a search, whether or not anyone clicks it. Impressions map your visibility surface — which queries you show up for, and how often.

There are two separate steps before a page can rank. First a crawler fetches the page (crawling). Then the engine decides whether the page is worth storing and showing (indexing). Only indexed pages are eligible to appear in search results — a crawled page is not automatically an indexed one.

Internal links are links from one page on your site to another. They do three jobs at once: they help crawlers discover pages, they pass ranking signals (often called link equity or PageRank) between pages, and they tell search engines how your topics relate to each other. The clickable anchor text is a strong relevance cue.

JavaScript SEO is the work of making sure content built with JavaScript frameworks (React, Vue, Angular) can actually be seen and indexed by search engines. Because much of a modern site’s content may load via JavaScript after the initial HTML, there is a risk that crawlers see an empty shell instead of your content.

Keyword cannibalization happens when two or more pages on your own site target the same query and end up competing with each other. Instead of one strong page ranking well, you get several weaker pages splitting the signals — and Google unsure which to show.

A keyword ranking is simply where your page appears in the results for a specific query. Because position varies by location, device, personalisation, and time, tools usually report an average position rather than a fixed spot.

The Knowledge Graph is Google’s vast database of entities — people, places, companies, things — and the relationships between them. It is what powers the knowledge panels you see on the right of many results, and it lets Google answer factual questions directly rather than just listing links.

In SEO terms, a landing page is any page where a visitor "lands" from search. In CRO terms, it is a focused page built to convert one audience toward one action. The best organic landing pages are both at once: they satisfy the search that brought the visitor and guide them to the next step.

A large language model is an AI trained on enormous amounts of text to predict the next word (technically, token) in a sequence. That simple objective, at scale, produces systems that can summarise, answer questions, and write fluently. LLMs are the engines behind ChatGPT, Claude, and Gemini, and increasingly behind search itself.

Link building is the practice of getting other websites to link to yours. Because backlinks are a strong trust signal, earning relevant, authoritative links can improve your rankings. It ranges from creating content people naturally want to cite, to actively reaching out and asking for links.

Log file analysis means examining your server’s access logs — the record of every request made to your site — to see exactly how search engine crawlers behave. Unlike estimates or tools, logs show precisely which URLs Googlebot fetched, how often, and what status code each returned. It is ground truth about crawling.

A long-tail keyword is a longer, more specific search phrase — "best project management tool for remote agencies" rather than "project management". Each one has lower search volume, but they are far more specific about what the searcher wants, and collectively the long tail makes up the majority of all searches.

The funnel is a model of the journey a person takes from first becoming aware of a problem to becoming a customer. It is usually split into stages: top of funnel (awareness), middle (consideration), and bottom (decision). It is called a funnel because many enter at the top and fewer reach the bottom, and it helps you match the right message to where someone is.

A marketing qualified lead is a lead that has cleared a defined bar — the right actions taken and the right firmographic fit — signalling enough genuine interest to be passed toward sales. It is the point where a raw lead becomes worth a salesperson’s time.

The meta description is a short HTML summary of a page. Google often shows it as the grey snippet text beneath the title in search results. It is not a ranking factor, but it is prime advertising space: a good one persuades the searcher to click your result over the others.

Mobile-first indexing means Google primarily uses the mobile version of your pages to index and rank your site, rather than the desktop version. Since most searches happen on phones, Google evaluates what mobile users see. If your mobile site is missing content that is on desktop, that content may effectively not count.

A 404 is the HTTP status code a server returns when a requested page does not exist. Seeing 404s is completely normal — the web is full of broken and mistyped links — and having them does not by itself hurt your rankings.

Organic traffic is the visitors who reach your site by clicking an unpaid listing in a search engine — as opposed to paid ads, social posts, email, or typing your URL directly. It is the headline outcome most people mean when they say "SEO is working".

A pillar page is the hub of a topic cluster. It targets a broad, high-volume "head" term and covers the whole topic at a survey level, then links out to the deeper cluster pages for the details. It is the page you would send someone who wants the complete overview of a subject.

A product qualified lead is a user who has already experienced meaningful value from your product — usually through a free trial or freemium plan — and shown behaviour that signals readiness to buy. Unlike a marketing qualified lead (interested based on content) or a sales qualified lead, a PQL has actually used the product and hit an activation moment.

Product-led SEO builds organic growth from the product itself rather than only from the blog. That can mean pages generated from your product data, free tools that rank and convert, integration or template pages, or user-generated content. The product becomes the engine that creates ranking pages at scale.

Programmatic SEO means generating a large set of similar pages from a template plus a structured data source, to capture repeatable long-tail query patterns — "[city] plumbers", "[integration] for [tool]", "[currency] to [currency]". Instead of writing each page by hand, you build one template and let the data populate thousands.

A prompt is the input someone gives an AI system — the question or instruction typed into ChatGPT, Perplexity, or an AI search box. In the AI era, prompts are becoming the new queries: longer, more conversational, and often multi-part compared to the short keyword phrases people type into traditional search.

A 301 is an HTTP status code meaning "this page has moved permanently". When a browser or crawler requests the old URL, the server sends them to the new one. Because it signals permanence, search engines transfer the old page’s ranking signals to the destination and eventually swap the new URL into their index.

A 302 is an HTTP status code meaning "this page has moved temporarily". Unlike a 301, it tells search engines the original URL is coming back, so they should keep indexing the old URL rather than swapping in the destination.

RAG is the architecture behind most AI search answers. Instead of relying only on what a language model memorised during training, the system first retrieves relevant, up-to-date passages from an index, then feeds them to the model as context so the answer is grounded in real, citable sources rather than the model’s memory alone.

Return on investment measures what you get back relative to what you put in. For SEO, it compares the revenue (or pipeline) organic search generates against the cost of the program — content, tools, people, links. It is the number that ultimately decides whether SEO keeps getting funded.

Robots.txt is a plain text file at the root of your domain that gives crawlers instructions about which paths they may fetch. It is a crawl-control tool, not a security or privacy tool — anything you truly need private must be behind authentication, because the file itself is public and only well-behaved bots obey it.

SaaS SEO is search engine optimisation adapted to the specifics of software-as-a-service businesses. Instead of selling a one-off product, you are guiding people through a longer, more considered journey toward a subscription — often via a free trial or demo — which changes what content you need and how you measure success.

Search intent is the reason behind a query — what the person actually wants, not just the words they typed. It usually falls into four buckets: informational (they want to learn), navigational (they want a specific site or page), commercial (they are comparing before a purchase), and transactional (they are ready to act or buy). Matching intent is the single biggest on-page ranking lever.

Semantic search means the engine tries to understand the meaning behind a query rather than matching the exact words. Powered by embeddings and language models, it recognises synonyms, context, and related concepts — so a page can rank for a query it never literally contains, as long as it genuinely covers the underlying topic.

SERP stands for Search Engine Results Page — the page you see after searching. It is no longer just ten blue links: it is a composite of organic results, ads, featured snippets, People Also Ask boxes, image and video packs, local map packs, and increasingly AI-generated overviews. Each of those is a distinct feature you can win or lose.

Social proof is evidence that other people trust and use your product — customer reviews, testimonials, client logos, case studies, ratings, user counts, and endorsements. It works because people look to others’ behaviour to decide what is safe and credible, so it reduces the perceived risk of choosing you.

Structured data is code — usually JSON-LD in the page head — that uses the shared schema.org vocabulary to explicitly label what a page is about: this is an Article, that is a Product with this price, this is a Person, these are FAQ questions. It translates human-readable content into machine-readable facts.

Thin content is any page that offers little or no value to the user — a few lines of generic text, auto-generated filler, doorway pages, or near-empty templated pages. Search engines increasingly judge quality at the page and site level, so thin pages struggle to index and rank, and enough of them can weigh down the whole site.

The title tag is an HTML element that sets the title of a page. It shows up as the clickable blue headline in search results and in the browser tab. It is one of the most important on-page elements: it tells both searchers and search engines what the page is about.

A topic cluster is a way of organising content by subject instead of by isolated keywords. One broad pillar page covers a theme at a high level, and a set of focused cluster pages each go deep on a sub-topic. Every cluster page links back to the pillar and the pillar links out to them, forming a tight, interlinked group.

Topical authority is how much expertise and comprehensive coverage a site demonstrates on a particular subject. Rather than ranking one page for one keyword, the idea is to cover a topic so thoroughly — every subtopic, question, and angle — that search engines come to see you as an authority on the whole subject and rank you more easily across it.

User experience is the overall quality of how it feels to use your site — how fast it loads, how easy it is to navigate, how clear the content is, and how well it works on any device. Good UX helps people find what they need and act on it; bad UX quietly sends them back to the search results.

A value proposition is the clear answer to a visitor’s core question: why should I choose you over the alternatives, including doing nothing? It states who it is for, the specific benefit they get, and what makes you different — ideally understandable in the first few seconds on a page.

An embedding turns a word, sentence, or document into a list of numbers (a vector) arranged so that things with similar meaning sit close together in mathematical space. It is how machines measure "these two pieces of text mean roughly the same thing" — the quiet engine under semantic search and AI retrieval.

An XML sitemap is a structured file listing the URLs you want search engines to know about, optionally with a lastmod date signalling when each page changed. You submit it in Search Console. It does not force anything to rank — it is a discovery and prioritisation aid that helps crawlers find pages efficiently.

A zero-click search is one where the user gets their answer directly on the results page — from a featured snippet, a knowledge panel, an AI Overview, or a calculator — and never clicks through to a website. A large and growing share of all searches now end this way.