Skip to content
NessFlow
Menu

Glossary

The terms we use on this site, defined in a sentence or two. Where a word means two different things depending on who says it, we say so.

AI crawler

A program that walks the web on behalf of a model operator. Three very different purposes coexist: training a model, feeding an answer engine, or fetching one page a user has just asked for.

See also: robots.txt, User agent, Answer engine

Where you can see it: AI crawler directory

Anchor text

The clickable text of a link. It tells an engine what sits on the other side; an unnaturally uniform anchor distribution across inbound links reads as manipulation.

See also: Backlink, Internal linking

Where you can see it: Backlink analysis

Answer engine

An interface that answers a natural-language question instead of returning a list of links: ChatGPT, Gemini, Perplexity, Copilot. The answer usually cites sources, and being among them is the new visibility question.

See also: GEO (generative engine optimization), Share of citations, AI crawler

Where you can see it: AI visibility audit

Canonical URL

A declaration in a page stating which address should be treated as the reference version when several URLs serve the same content. It is a hint, not a directive: an engine may ignore it if it looks inconsistent.

See also: Duplicate content, hreflang, Indexing

Click depth

The smallest number of links you have to follow from the home page to reach a given page. Greater depth correlates with rarer crawling, provided it is measured on the site’s real links rather than on an injected sitemap.

See also: Internal linking, Orphan page, Crawl budget

Where you can see it: Technical SEO audit

CLS (cumulative layout shift)

How much a page moves under the reader while it loads. The published reference threshold is 0.1.

See also: Core Web Vitals, LCP (largest contentful paint), INP (interaction to next paint)

Contrast ratio

The luminance ratio between text and its background. WCAG 2.2 level AA asks for 4.5:1 on body text and 3:1 on large text, those are standards thresholds, not product measurements.

See also: WCAG, RGAA

Where you can see it: Accessibility audit

Core Web Vitals

Three loading-experience metrics published by Google: LCP for how fast the main content appears, CLS for visual stability, INP for responsiveness to interaction.

See also: LCP (largest contentful paint), CLS (cumulative layout shift), INP (interaction to next paint)

Crawl budget

How many pages an engine is willing to crawl on a site in a given period. It is not contractual: you infer it from behaviour observed in server logs, never from an announced figure.

See also: Server log, Click depth, Ghost page

Where you can see it: SEO log file analysis

Duplicate content

The same content served under several addresses. The cost is not a penalty but dilution: signals split across the URLs, and the engine picks one itself.

See also: Canonical URL, hreflang

EAA (European Accessibility Act)

An EU directive making accessibility mandatory across a broad set of digital products and services. It applies through national transposition, so deadlines and penalties differ by member state.

See also: WCAG, RGAA, VPAT

Where you can see it: Accessibility audit

Gap analysis

Cross-referencing your link or keyword profile against competitors to isolate what cites or ranks them and not you. It produces a target list, not a diagnosis.

See also: Backlink, Referring domain

Where you can see it: Backlink analysis

GEO (generative engine optimization)

The practices aimed at being cited by answer engines. The vocabulary has not settled: some people say "AEO" or "AI SEO" for roughly the same thing.

See also: Answer engine, llms.txt, Share of citations

Where you can see it: AI visibility audit

Ghost page

A page your server logs show being crawled although it appears in no crawl of the site. Usually a removed URL, a forgotten parameter, or an inbound link from a third party.

See also: Server log, Orphan page, Crawl budget

Where you can see it: SEO log file analysis

hreflang

A declaration linking the language versions of the same page. It has to be reciprocal: page A declaring B without B declaring A is ignored.

See also: Canonical URL, Duplicate content

Indexing

Whether a crawled page is kept in an engine’s index and therefore eligible to be shown. Crawled is not indexed: an engine can read a page and decide not to keep it.

See also: Search Console, Canonical URL, Sitemap

Where you can see it: Technical SEO audit

INP (interaction to next paint)

The delay between a visitor’s interaction and the next visible paint. It replaced FID as the Core Web Vitals responsiveness metric; the published reference threshold is 200 ms.

See also: Core Web Vitals, LCP (largest contentful paint), CLS (cumulative layout shift)

Internal linking

The links a site points at its own pages. They decide what an engine reaches, how often, and how value spreads between pages.

See also: Click depth, Orphan page, Internal PageRank

Where you can see it: NessFlow vs Screaming Frog

Internal PageRank

A model of how value flows from a page to the pages it links. Recomputed on a site’s internal graph alone, it ranks pages against each other, it says nothing about how an engine values them.

See also: Internal linking, Click depth

Where you can see it: NessFlow vs Screaming Frog

JavaScript rendering

Executing a page’s JavaScript before analysing its content. Without it a tool reads the initial HTML (often nearly empty on a modern application), and wrongly concludes the page has no content.

See also: SPA (single-page application), Crawl budget

Where you can see it: Technical SEO audit

LCP (largest contentful paint)

How long before the largest visible element of the page is painted. It is the perceived-speed metric of Core Web Vitals; the published reference threshold is 2.5 s.

See also: Core Web Vitals, CLS (cumulative layout shift), INP (interaction to next paint)

llms.txt

A text file at the root of a domain that introduces the site to answer engines in a model-readable form. It is not an official standard: no operator commits to reading it.

See also: GEO (generative engine optimization), robots.txt, Answer engine

Where you can see it: robots.txt for AI crawlers: the complete configuration, crawler by crawler

Mixed content

An HTTPS page loading a resource over HTTP. Browsers block or downgrade those resources, so what the visitor sees is a broken page rather than a warning.

See also: Structured data

Orphan page

A page with no inbound link from the site itself. Engines may still know it from a sitemap or an external link, but nothing in the navigation leads there.

See also: Internal linking, Click depth, Ghost page

Where you can see it: Server log analysis: what your servers know and Search Console does not

Referring domain

A distinct domain pointing at least one link at your site. This is the unit that counts: a hundred links from one domain are not a hundred domains.

See also: Backlink, Spam score, Gap analysis

Where you can see it: Backlink analysis

RGAA

The French accessibility framework, a national derivation of the WCAG criteria. It adds a test method and a conformance declaration format that carry legal weight in France.

See also: WCAG, EAA (European Accessibility Act), VPAT

Where you can see it: Accessibility audit

robots.txt

A file at the root of a domain declaring what each crawler may fetch. Groups do not stack: a crawler applies only the most specific group that names it.

See also: AI crawler, User agent, Crawl budget

Where you can see it: robots.txt for AI crawlers: the complete configuration, crawler by crawler

Search Console

Google’s interface giving impressions, clicks and indexing status for a verified property. Its numbers are sampled and rounded: treat them as orders of magnitude, not accounting.

See also: Indexing, SERP, Server log

Where you can see it: NessFlow vs Botify & Oncrawl

Semantic coverage

A measure of how many of the notions expected on a topic actually appear in a text. It shows what a piece is missing; it does not say whether the piece is good.

See also: Structured data

Where you can see it: Semantic optimisation

SERP

A search engine results page for a given query. It varies with location, language and device: a ranking only means something alongside those three parameters.

See also: SERP feature, Search Console

Where you can see it: Rank tracking

SERP feature

Any block of a results page that is not a blue link: featured snippet, people-also-ask, image carousel, AI-generated overview. Their presence moves clicks far more than rank itself.

See also: SERP, Answer engine

Where you can see it: Rank tracking

Server log

The file where a server records every request it receives. It is the only source that says what engines actually asked for, as opposed to what a tool assumes they do.

See also: Crawl budget, Ghost page, Search Console

Where you can see it: SEO log file analysis

Share of citations

The proportion of an answer engine’s responses, over a given set of questions, that cite your domain. The figure only means something with its denominator: 30% of ten answers and 30% of a hundred are different claims.

See also: Answer engine, GEO (generative engine optimization)

Where you can see it: AI visibility audit

Sitemap

An XML file listing the URLs a site wants crawled. It is a suggestion: being listed guarantees neither crawling nor indexing, and being absent forbids neither.

See also: Indexing, Orphan page, Click depth

SPA (single-page application)

A web application that navigates without full page reloads, with the browser building the content. This is where JavaScript rendering stops being a convenience and becomes a condition for being crawled at all.

See also: JavaScript rendering, Click depth

Where you can see it: JavaScript SEO: getting a single-page app crawled

Spam score

A vendor-supplied index estimating how likely a referring domain is to be artificial. It is a proprietary estimate, not an engine’s verdict.

See also: Backlink, Referring domain

Where you can see it: Backlink analysis

Structured data

Markup describing what a page’s content means in a format engines interpret. Markup describing something other than what the page shows loses the rich result it existed to win.

See also: SERP feature, Indexing

User agent

The string a client sends to identify itself to a server. It is declarative and trivially forged: blocking on it alone stops the honest crawlers and nobody else.

See also: AI crawler, robots.txt, Server log

Where you can see it: AI crawler directory

VPAT

A standardised US-originated document in which a vendor states how far its product conforms to accessibility criteria. It is a supplier declaration, not an independent certification.

See also: WCAG, RGAA, EAA (European Accessibility Act)

Where you can see it: Accessibility audit

WCAG

The international Web Content Accessibility Guidelines, organised into testable criteria and three conformance levels (A, AA, AAA). Version 2.2 is the current reference.

See also: RGAA, VPAT, Contrast ratio

Missing a term?

This glossary follows what the product measures. If a term is missing, a demo is the best place to ask, we answer on your site.

Access opens in waves: we email you when yours is ready.