Digital Marketing Glossary

Making your content the source a system quotes when it answers a question directly, rather than one of ten blue links. Covers featured snippets, voice results and AI assistants.

The practical difference from classic SEO is that you are optimising to be extracted, not to be clicked. That changes how you structure an answer: direct first sentence, no preamble, facts stated definitively.

Google's generated summary at the top of some results pages, assembled from multiple sources with citations. It appears more often on informational queries and less on transactional ones.

Being cited in one does not guarantee traffic, because many readers get what they needed without clicking. Measuring its effect means watching impressions and brand search rather than clicks alone.

A change to how a search engine ranks results. Minor adjustments happen constantly; named broad core updates roll out a few times a year and can reorder an entire sector.

Sites rarely fall because of a single fixable error. More often an update re-weights something ... expertise signals, content depth, brand authority ... and a site that was marginal on that dimension drops.

A written description of an image, read aloud by screen readers and used by search engines to understand what the image shows.

Good alt text describes what matters about the image in context. "Chart showing organic sessions rising from 400 to 2,100 between January and August" beats "chart" and also beats a keyword-stuffed string nobody would say out loud.

The visible, clickable words in a link. It tells both readers and search engines what to expect on the other side.

Over-optimising it ... every link to a page using the same exact phrase ... looks manufactured. Natural profiles vary, and a share of links will be branded or generic no matter what you do.

Deciding which marketing touchpoint gets credit for a conversion. Last-click gives it all to the final interaction; data-driven models spread it across the path.

Every model is wrong in a different direction. Last-click systematically undervalues channels that create demand early, which is why organic and social often look weaker than they are.

Loosely, how much a search engine trusts a source on a subject. Google does not publish an authority score; the numbers you see in tools are third-party estimates.

Treat tool scores as a rough comparison between sites, never as a metric to optimise. A site can gain real authority while its Domain Rating stands still.

A link from another website to yours. Still one of the strongest ranking signals, because it is a vote someone else had to choose to cast.

Quality varies enormously. One link from a relevant publication with real readers outweighs hundreds from directories nobody visits, and the cheap ones can actively hurt.

Historically, the share of visits where someone viewed one page and left. GA4 replaced it with engagement rate, which is roughly its inverse.

A high bounce rate is not automatically bad. Someone who lands on your opening hours page, reads them and leaves has been served perfectly.

A navigational trail showing where a page sits in the site hierarchy, usually above the heading. Home → Services → SEO → Technical.

They help users orient and give search engines a clear statement of your structure. With breadcrumb schema, Google often displays them in place of the URL in results.

An instruction telling search engines which version of a page is the original when several URLs show near-identical content.

It is a hint, not a command. Google can and does ignore canonicals it disagrees with, particularly when the pages are not actually duplicates.

Two or more of your own pages competing for the same query, so neither builds enough signal to rank well.

The usual cause is publishing without a keyword map. The fix is usually merging into one stronger page rather than trying to differentiate two weak ones.

The document that tells a writer what to produce: the query network, the headings, the questions to answer, the entities to name, the format, and the internal links.

Most content fails at the brief rather than the writing. A good writer given a vague brief produces a coherent article about the wrong thing.

A group of related pages covering one subject in depth, linked to each other and usually to a central pillar page.

The point is coverage. A search engine judges your standing on a subject partly by how completely you have treated it, not by how good one page is.

An action you have decided counts as valuable ... a form submission, a call, a purchase, a booking.

Define these before you measure anything. Teams that never agreed what counts end up arguing about whether a channel works rather than about what to do next.

A broad change to Google's ranking systems, announced and typically rolling out over one to three weeks.

Do not make large changes mid-rollout. Rankings fluctuate while it settles, and you will misattribute the cause of whatever happens next.

Three measures of page experience: LCP (how fast the main content appears), INP (how quickly the page responds to input), and CLS (how much the layout shifts while loading).

They are a genuine ranking factor but a weak one. Passing them will not rescue thin content; failing them badly will cost you on mobile, where patience is shortest.

How much crawling a search engine is willing to do on your site in a given period.

Irrelevant for most sites under a few thousand pages. It matters for large ecommerce catalogues and anywhere faceted navigation generates millions of parameter URLs.

The process of a bot fetching pages and following links to discover more. The first of three stages, before indexing and ranking.

A page that is not crawled cannot be indexed, and a page that is not indexed cannot rank. Most "why doesn't this rank" questions turn out to be one of these two earlier problems.

Clicks divided by impressions. In Search Console it tells you how compelling your title and description are at a given ranking position.

Compare it against expected CTR for that position, not against an absolute benchmark. A 4% CTR is poor at position two and excellent at position nine.

What you pay each time someone clicks your ad. Set by auction, so it moves with competition rather than staying fixed.

A low CPC is not automatically good. A cheap click on an irrelevant query costs more in wasted budget than an expensive click from someone ready to buy.

Third-party scores from Ahrefs and Moz estimating a site's backlink strength on a 0–100 scale. Neither is used by Google.

Useful for comparing link profiles quickly. Damaging when treated as a target, because it can be raised with links that do nothing for rankings.

Substantially identical content on more than one URL, whether on your site or across sites.

There is no duplicate content penalty, contrary to persistent belief. What happens is Google picks one version to show and ignores the rest, which is a visibility problem rather than a punishment.

How long someone stays on your page after clicking from search before returning to the results. You cannot measure it directly in any analytics tool.

Google has never confirmed using it as a ranking factor. Treat it as a useful concept for thinking about satisfaction rather than a number to chase.

Earning coverage and links from publications by giving journalists something worth writing about ... original data, a useful tool, an informed comment on a live story.

The most durable link building method, because the links are a by-product of genuine coverage rather than the purchase being disguised as one.

Experience, Expertise, Authoritativeness and Trust. Criteria from Google's Search Quality Rater Guidelines, used by human raters who evaluate results.

It is not a ranking factor you can optimise directly ... there is no E-E-A-T score. It describes the qualities Google's systems are built to approximate, which is why author credentials, citations and first-hand experience matter more than they used to.

A distinct thing a search engine recognises and holds facts about ... a person, company, product, place or concept ... rather than just a string of characters.

The shift from keywords to entities is the central change in search over the past decade. Being understood as a thing, with consistent facts across the web, is what makes you eligible for knowledge panels and AI citations.

GA4's replacement for bounce rate. The share of sessions lasting over ten seconds, containing a key event, or including two or more page views.

The ten-second threshold is arbitrary and generous. A session can count as engaged while achieving nothing, so pair it with a real outcome metric.

Filters on a category page ... size, colour, price, brand ... that generate a new URL for each combination.

Left uncontrolled, twelve filters can produce hundreds of thousands of crawlable URLs with near-identical content. It is the most common cause of index bloat in ecommerce.

A block at the top of results answering the query directly, pulled from a page that already ranks on page one.

You win it by answering the question in a self-contained paragraph of roughly forty words, placed directly under a heading that matches the question. Lists and tables win their own formats.

Information you collect directly from your own audience ... email subscribers, customer records, on-site behaviour.

As third-party cookies decline, this becomes the durable asset. It is also the one thing a platform change cannot take away from you overnight.

The current version of Google Analytics, built on events rather than sessions and pageviews.

The model change is the hard part for anyone used to Universal Analytics. Nearly everything is an event now, which is more flexible and considerably less intuitive.

Getting your brand cited inside answers produced by large language models such as ChatGPT, Perplexity, Gemini and Claude.

Overlaps heavily with AEO and with classic SEO. What differs is that citation rather than ranking is the outcome, and brand mentions across the web appear to matter more than backlinks alone.

The free listing that controls how a business appears in Google Maps and the local pack.

For a local business it usually drives more enquiries than the website does. Category selection and review volume are the two levers that move it most.

Google's free tool reporting which queries surfaced your pages, your positions, click-through rates, indexing status and technical errors.

The single most useful free tool in SEO, and the only place you see actual query data. Its numbers will never match GA4's, because the two count different things at different points.

The structure of H1 through H6 on a page. One H1 stating the subject, H2s for major sections, H3s for subsections beneath them.

Skipping levels or choosing a heading tag for its font size breaks the document outline that screen readers and search engines rely on.

Markup telling search engines which language and regional version of a page to show which audience.

Notoriously easy to break. Every version must point at every other version including itself, and one missing return tag invalidates the set.

Large numbers of low-value pages in the search index ... filter combinations, tag archives, thin pagination, internal search results.

It dilutes how a search engine perceives site quality and wastes crawl capacity on pages nobody should land on.

Storing and organising a crawled page so it can be returned in results. Crawled is not the same as indexed.

Google now declines to index a significant share of what it crawls, usually because the page duplicates something else or offers too little to be worth storing.

A link from one page on your site to another. The most underused lever in SEO, because it is entirely under your control.

Internal links distribute authority, establish which pages matter, and tell search engines what a page is about through anchor text. A link placed in body content carries more weight than one in a sitewide footer.

What someone actually wants from a search. Usually grouped as informational, navigational, commercial and transactional.

Intent is settled by the SERP, not by the words in the query. If page one is all product pages, an article will not rank there regardless of quality.

A 0–100 estimate from SEO tools of how hard a term is to rank for, usually derived from the backlink profiles of the pages already ranking.

Each tool calculates it differently and none accounts for topical relevance. A "hard" term in a niche you cover thoroughly is often easier than an "easy" term in one you do not.

Grouping queries that should be served by one page, usually by checking how much the top ten results overlap between them.

This is what prevents you writing five thin articles about what is really one topic. If two queries return largely the same results, they want one page.

Google's database of entities and the relationships between them ... the structure behind knowledge panels and much of what AI systems treat as fact.

Getting your brand into it requires consistent information across authoritative sources, structured data on your own site, and ideally a Wikidata entry.

GA4's term for an event you have marked as important, replacing what Universal Analytics called a goal or conversion.

Marking too many events as key makes the report meaningless. Three to five that represent real business outcomes is usually enough.

Strictly, any page someone arrives on. In paid media, a purpose-built page designed for one campaign with one action.

The reason dedicated landing pages convert better is message match ... the page repeats the promise the ad made, rather than dropping someone on a homepage to find it themselves.

The map and three business listings that appear for searches with local intent.

Ranked on relevance, distance and prominence. Distance you cannot change, which is why service-area businesses need a different strategy from those with a storefront.

Examining your server logs to see exactly what search engine bots fetched, when, and what response they got.

The only way to see real crawler behaviour rather than inferring it. Overkill for small sites, invaluable on large ones.

A longer, more specific query with lower search volume and usually clearer intent.

They convert better because specificity signals a later buying stage. "Waterproof hiking boots wide fit women's" is closer to a purchase than "boots".

An audience built by a platform to resemble people on a list you supplied, usually customers or high-value visitors.

Quality depends entirely on the seed list. A lookalike built from all website visitors is close to broad targeting; one built from repeat purchasers is a genuinely different audience.

The summary beneath your title in search results. Not a ranking factor, but it affects whether anyone clicks.

Google rewrites it more often than not, usually pulling text from the page that better matches the query. Write one anyway ... it is the version shown when yours fits.

Google indexes and ranks based on the mobile version of your page, not the desktop one.

The failure mode is content parity. If your mobile template hides sections or loads less content, that hidden content is effectively what Google sees ... which is to say, it does not.

A term you tell an ad platform not to match against, so your ad never shows for it.

The single cheapest optimisation in paid search. Most wasted spend in a new account comes from broad match picking up queries nobody thought to exclude.

A directive telling search engines not to include a page in results.

Do not combine it with a robots.txt block. If the crawler cannot fetch the page, it never sees the noindex, and the page can remain indexed indefinitely.

A link attribute telling search engines not to pass authority through it. Now treated as a hint rather than a directive.

Related attributes sponsored and ugc exist for paid and user-generated links respectively, and using the right one is a compliance matter as much as a technical one.

Everything affecting your rankings that happens away from your site ... backlinks, brand mentions, reviews, coverage.

Less controllable than on-page work and, increasingly, more decisive. Brand mentions in particular carry weight in AI-generated answers where links do not appear at all.

Everything on the page itself ... title, headings, content, internal links, images, structured data.

The part fully under your control, and where most sites have unclaimed gains sitting untouched.

A page with no internal links pointing to it, reachable only by knowing the URL or finding it in a sitemap.

Search engines struggle to discover them and read the absence of links as a signal the page does not matter. Usually caused by pages created outside the normal publishing flow.

The expandable questions in Google's results, which generate more questions as you open them.

One of the best free sources of real query language. Answering a PAA question in one direct sentence under a matching heading is the standard route to appearing there.

A Google Ads campaign type running across Search, Shopping, Display, YouTube and Gmail from one set of assets, with placement decided automatically.

It performs well when fed good conversion data and poorly when not. The persistent complaint is reporting opacity ... you often cannot see where spend went.

A comprehensive page covering a broad subject, linking out to detailed pages on each sub-topic and receiving links back from them.

It works as a hub that concentrates authority and makes your coverage of a subject legible to a search engine in one place.

Google Ads' 1–10 rating of expected click-through rate, ad relevance and landing page experience.

A higher score lowers what you pay per click for the same position. The landing page component is the one advertisers most often ignore and most easily fix.

What someone actually types or says. Distinct from a "keyword", which is the term you have chosen to target.

The distinction matters because Search Console reports queries, and the real ones are usually longer, messier and more revealing than the keyword you planned around.

An instruction sending a browser and crawler from one URL to another. 301 signals permanent, 302 temporary.

Use 301 unless the move genuinely is temporary. Avoid chains ... each hop adds latency and, over several, can lose signal.

The stage where a search engine executes JavaScript to see the page as a browser would.

It happens, but often later and not always completely. Content that only exists after JavaScript runs is at greater risk of being missed than content in the initial HTML.

A file at your domain root telling crawlers which paths they may fetch.

It controls crawling, not indexing. A blocked URL can still appear in results if other pages link to it, which surprises people who used it hoping to hide something.

Showing ads to people who already visited your site or engaged with your content.

It reports well because it reaches people who were going to return anyway, which means its measured ROAS routinely overstates its real contribution.

Revenue generated divided by advertising cost. A 4x ROAS means four units of revenue for every one spent.

Revenue is not profit. A 4x ROAS on a product with a 20% margin is losing money, which is why ROAS targets set without margin data mislead.

A schema property listing other authoritative URLs that refer to the same entity ... your LinkedIn page, YouTube channel, Wikidata entry.

It is how you tell a machine that five scattered profiles are one organisation rather than five. Central to entity consolidation.

Structured data in your page's code describing what the content is ... an article, a product, an organisation, an FAQ ... in a format machines parse reliably.

It does not raise rankings directly. It makes you eligible for rich results and gives AI systems facts they can use with confidence rather than inference.

Search Engine Results Page. The page returned for a query, including ads, snippets, local pack, images and AI overviews alongside the organic links.

Reading the SERP before planning a page is the most useful ten minutes in SEO. It tells you the intent, the expected format and how much real estate organic results still get.

How pages are organised and linked. Flat architectures keep important pages few clicks from the homepage; deep ones bury them.

It decides crawl efficiency, how authority flows, and whether a visitor can find anything. Worth getting right before publishing three hundred pages rather than after.

A file listing the URLs you want indexed, submitted through Search Console.

It helps discovery; it does not guarantee indexing. Include only canonical, indexable URLs ... a sitemap full of redirects and noindexed pages is a quality signal in the wrong direction.

The work that lets search engines crawl, render and index a site properly ... speed, architecture, indexation control, structured data, migrations.

It rarely creates rankings by itself. It removes the ceilings that stop good content ranking, which is a different and often more urgent job.

The page's title as shown in search results and browser tabs. Still one of the strongest on-page signals.

Front-load the distinctive words. Google rewrites roughly a third of titles, usually when yours is too long, keyword-stuffed, or fails to describe the page.

Being treated as a credible source across a whole subject rather than for individual pages.

Built through comprehensive coverage, consistent publishing and accumulated engagement over time. It is why a well-covered niche site can outrank a much larger generalist one.

A planned structure of every page a site should publish on its subject, derived from the central entity and its attributes rather than from a keyword list.

The difference from a content calendar is that a map is designed for completeness. It tells you not just what to write next but when a subject is finished.

Tags appended to a URL identifying the campaign, source and medium that produced a visit.

They only work if everyone uses the same conventions. Inconsistent capitalisation alone will split one campaign into four rows in your reports.

Your brand named on another site without a link back.

Worth finding and politely asking to convert. They also matter on their own for AI search, where a mention can influence citation even with no link attached.

How much rankings are moving across the web at a given moment, tracked by third-party sensors.

Useful for context. A drop during high volatility is probably an algorithm change; the same drop on a quiet day is more likely something you did.

An open, machine-readable database of entities, used as a source by Google's Knowledge Graph and by language models.

A Wikidata entry is one of the more direct routes to being recognised as an entity, but notability requirements are real and self-created entries without independent sources get removed.

How long a page is. Not a ranking factor, despite decades of belief otherwise.

Longer pages correlate with better rankings because thorough treatment takes words, not because length itself helps. Padding a page to hit a number makes it worse.

A search resolved on the results page itself, through a snippet, knowledge panel or AI overview, without anyone visiting a website.

A growing share of all searches. It forces a measurement change ... if visibility produces value you never see as a click, click-based reporting will understate the channel.