Advanced SEO Glossary: 350 Terms, Defined in Plain English
Every term a business owner or marketer runs into, written by the person who uses them on client sites, and corrected when Google changes the rules.
Reviewed August 14, 2026. Includes the 2026 changes most glossaries have not caught up with.
The short version
- This glossary defines 350 SEO terms across seventeen categories, from crawling and indexing through entities, local search, technical SEO, e-commerce and AI answers.
- You need about twenty of them to hire well. The rest are for the person doing the work.
- Google publishes very few numbers. Clicks, impressions, position, Core Web Vitals and index status are real. Domain Authority, keyword difficulty and search volume are vendor estimates.
- Core Web Vitals thresholds are 2.5 seconds for LCP, 200 milliseconds for INP and 0.1 for CLS, measured at the 75th percentile of real Chrome user data over a rolling 28 day window.
- AI search did not create a new discipline. Google’s own May 2026 guidance says answer engine optimization and generative engine optimization are still SEO, and that no special schema, llms.txt file or content chunking is required.
- The last section lists terms that are retired or misleading, so you can date any proposal that still uses them.
Most SEO glossaries are copied from other SEO glossaries. That is why the same errors survive for a decade, and why so many of them still explain a metric Google retired two years ago as though it were current.
This one is written and maintained by one person, checked against Google’s own documentation rather than against other glossaries, and dated so you can see when it was last verified. Where a term is contested, obsolete or invented by a software vendor, it says so.
Who wrote this
I am David Cragg. I have worked in search since 1990. I built and sold two internet marketing companies, Lotus411.com and MSD2D.com, started Temecula SEO in 2007, and moved the practice to Tucson in 2022. I have a BA in economics from UCLA and an MBA from the University of Washington.
I work with a small number of local businesses, one at a time, and I do the work myself. Every definition on this page describes something I have had to explain to a client, fix on a live site, or undo after somebody else sold it.
Written and maintained by David Cragg. Last reviewed August 14, 2026. How this page is written and checked.
What is in this glossary
- Core search concepts (30)
- Google's named ranking systems (12)
- Keywords and search intent (24)
- On-page SEO (30)
- Content quality and E-E-A-T (25)
- Entities and the Knowledge Graph (10)
- Links and off-page signals (25)
- Local SEO (28)
- Technical SEO (51)
- Images, video and other search surfaces (6)
- International and multilingual SEO (6)
- E-commerce and product SEO (8)
- Measurement and reporting (28)
- AI search: the vocabulary that is actually current (26)
- Spam tactics and black-hat vocabulary (10)
- How SEO work is bought and sold (8)
- Terms that are retired, wrong, or a warning sign (23)
Core search concepts
These are the words that show up in every audit, every proposal and every argument about rankings. Most people learn them roughly. Learning them exactly is what stops you from paying for the wrong work.
Search engine optimization (SEO)
The work of making a website easier for search engines to find, understand and trust, so its pages appear for the searches its customers actually type. It covers content, site structure, code and off-site reputation.
Organic result
A listing a search engine shows because it judged the page relevant, not because someone paid for placement. Organic clicks cost nothing per click, which is why they compound in value over time.
Paid result
An ad placed at the top or bottom of a results page through Google Ads or a similar system. Paid results are labeled as sponsored. Ad spend has no effect on organic rankings.
SERP
Short for search engine results page. Today a SERP is rarely a plain list of ten links. It can hold ads, an AI answer, a map pack, images, videos, People Also Ask boxes and only a handful of organic listings.
Query
The words a person actually types or speaks. A query is not the same thing as a keyword. Keywords are what you target; queries are what real people send, and they are messier, longer and more revealing.
Ranking position
Where a page sits in the organic listings for one query, at one moment, for one type of searcher. Position moves constantly by location, device and search history, so a single screenshot proves very little.
Impression
One instance of your page being shown to a searcher. Search Console counts an impression even when the listing sits far down the page, which is why impressions can rise while clicks stay flat.
Snippet
The short block of text Google shows under your title in the results. Google writes it from your page, using your meta description only when it fits the query well.
Featured snippet
An answer Google lifts out of a page and displays above the normal results, with a link back to the source. You earn one by answering a question directly and early on the page, in a clean paragraph, list or table.
SERP feature
Any result that is not a plain blue link: map packs, image rows, video carousels, People Also Ask, knowledge panels, AI answers. Which features appear tells you what kind of page Google wants for that query.
Algorithm
The set of ranking systems Google uses to order results. There is no single formula, and no one outside Google has seen it. What is public is the documented guidance and the pattern of what wins over time.
Ranking signal
Any input a search engine uses to judge a page. Google has confirmed relatively few by name. Treat any list of "200 ranking factors" as marketing, not documentation.
Core update
A broad, announced change to Google's ranking systems, usually rolled out over one to three weeks. Core updates rebalance how quality is judged across the whole index. Recent ones ran March 27 to April 8, 2026, and May 21 to early June, 2026.
Spam update
A separate announced rollout aimed at content and links that break Google's spam policies. Unlike a core update, a spam update is about enforcement, not recalibration.
Google Search spam policies
The published rules that define what will get a site demoted or removed, including cloaking, scaled content abuse, link schemes and site reputation abuse. In 2026 Google clarified that these policies also cover its AI answers.
Manual action
A human reviewer at Google penalizing a site for breaking policy. It appears by name in Search Console, and it must be fixed and then submitted for reconsideration. Most ranking drops are not manual actions.
Search Quality Rater Guidelines
Google's public handbook for the contractors who score sample search results. Raters do not change rankings directly, but the document is the clearest written statement of what Google means by quality.
Sitelinks
The extra sub-links Google sometimes shows beneath a listing for a brand or a strong page. They are generated automatically from your site structure and internal links. You cannot request them.
Index (search index)
The database of pages Google has crawled, analyzed and stored. A page has to be in the index before it can rank at all, and being in the index is not the same as Google considering it worth showing.
Rich result
A listing enhanced with extra detail pulled from structured data, such as star ratings, event dates or breadcrumbs. Correct markup makes a rich result possible. It never makes one certain, and Google retires the types it decides are not worth showing.
People Also Ask
The expanding list of related questions Google shows inside the results. Each answer is lifted from a page the same way a featured snippet is, which makes the box a free list of the questions your customers are actually asking.
Google Discover
The feed of articles Google shows on mobile home screens and in the Google app, chosen by a person's interests rather than by a query. Discover traffic arrives without anyone searching for you, and it leaves just as abruptly.
Top Stories
The news carousel Google shows for queries with a current-events angle. It draws on news publishers, and AMP has not been a requirement for inclusion since 2021.
Site diversity
Google's system for limiting how many results from one domain appear for a single query, generally no more than two. It is why publishing a fourth page on the same topic rarely buys a fourth listing.
Personalization
The adjustments Google makes to results based on location, device, language and, to a limited extent, past activity. It is the reason your own screenshot of a ranking is not evidence of what anybody else sees.
Search operator
A modifier that changes how a query is handled, such as site:, inurl: or a quoted phrase. Operators are useful for spot checks, but they run against a sampled view of the index and the result counts they return are estimates.
Google Search Essentials
Google's baseline documentation, renamed from the Webmaster Guidelines in 2022. It covers the technical requirements, the spam policies and the key best practices, and it is the shortest honest answer to what Google actually asks of a site.
Autocomplete
The predictions Google offers while you type. They reflect real, common queries, which makes them a decent free source of customer phrasing, but they are shaped by popularity and location rather than by commercial value.
SERP volatility
How much rankings are moving across the web on a given day, tracked by third-party sensors. High volatility suggests an update is rolling out. It does not tell you what changed, or whether you were affected.
Reconsideration request
The submission you make in Search Console after fixing a manual action, explaining what was wrong and what you did about it. It applies only to manual actions, and a vague one gets rejected.
When a prospect calls me and says a page "is not ranking," the first question I ask is whether it is even indexed. About a third of the time it is not, and no amount of content work would have fixed that. Precise words save money.
David CraggTucson SEO consultant, in search since 1990Google's named ranking systems
Google publishes a short list of the ranking systems it is willing to name, and a second list of names it has retired because the system was folded into the core. Knowing which list a name sits on matters, because a great deal of SEO diagnosis is still written in the vocabulary of systems that stopped existing separately years ago.
Ranking system
Google's own word for a component of search that runs continuously, as distinct from an update, which is a one-time change to those components. Google publishes the list of systems it will name, and it is far shorter than any ranking factor list you have been shown.
Panda
The 2011 system that demoted thin and low-value pages. Google retired the name once the system was folded into the core, so a quality problem in 2026 should not be described as a Panda hit.
Penguin
The 2012 system aimed at manipulative links. Since 2016 it has run in real time inside the core systems, and it generally devalues the offending links rather than penalizing the whole site.
Hummingbird
The 2013 rewrite of how Google interpreted queries, shifting the emphasis from matching words to understanding meaning. Google lists it among retired system names because it is now simply how search works.
RankBrain
The machine learning system introduced in 2015 that helps Google relate words to concepts, particularly for queries it has never seen before. There is nothing to optimize for directly.
Neural matching
A system that connects a query to a page by concept rather than by wording. It is why a page can rank well for phrases that appear nowhere on it.
BERT
A language model introduced in 2019 that improved Google's handling of word order, prepositions and nuance in longer queries. It rewards writing that reads naturally and penalizes writing assembled to satisfy a keyword tool.
MUM
Multitask Unified Model, announced in 2021. A much larger language model Google applies to specific features rather than as a general replacement for ranking.
Passage ranking
Google's ability to rank one section of a long page for a query even when the page as a whole is about something broader. It rewards clear headings and self-contained sections.
Helpful content system
The system introduced in 2022 to demote sites that read as though they were written for search engines rather than people. Google folded it into the core ranking systems in March 2024, so there is no separate switch left to flip back.
Reviews system
Google's system for judging review content, which expects first-hand testing, evidence and real comparison rather than a rewrite of the manufacturer's page. It was extended beyond products to services and other subjects in 2023.
Freshness systems
The systems that favor recent content for queries where recency genuinely matters, such as news, scores, prices and this year's version of something. Changing a date without changing the content does not qualify.
Keywords and search intent
Keyword work is where most small business SEO goes wrong, and it usually goes wrong in the same way: picking words by volume instead of by intent. These terms are the vocabulary for getting that decision right.
Keyword
A word or phrase you decide a page should compete for. One page, one primary keyword. When two pages chase the same one, both usually lose.
Head term
A short, high-volume, high-competition phrase such as "dentist" or "SEO." Head terms are vague about intent, so they are hard to convert and hard to win without real authority behind the site.
Long-tail keyword
A longer, more specific phrase such as "emergency dentist open Saturday in Murrieta." Lower volume, far clearer intent, and usually where a small business gets its first wins.
Keyword research
The process of finding the phrases real customers use, estimating demand, judging difficulty, and deciding which ones deserve a page. It ends with a decision, not a spreadsheet.
Search volume
An estimate of how many times a phrase is searched per month. Every tool's number is modeled, not measured, and local volumes are especially rough. Use it to compare phrases, not to forecast revenue.
Keyword difficulty
A tool's score, usually 0 to 100, estimating how hard it is to rank for a phrase. It is a vendor invention based mostly on the links pointing at the pages that currently rank. Useful for triage, meaningless as a promise.
Search intent
What the person actually wants when they type a query. Intent is the single most important thing to get right, because a page that answers the wrong need will not rank no matter how well it is built.
Informational intent
The searcher wants to learn something: what a term means, how something works, whether a problem is serious. Guides, explainers and glossaries serve this intent.
Navigational intent
The searcher already knows where they want to go and is using search as a shortcut, as in "Tucson SEO contact" or a brand name. Do not build content to compete for someone else's brand.
Commercial investigation
The searcher is comparing options before buying: best, versus, reviews, cost, alternatives. These queries convert well and are usually underserved by small business websites.
Transactional intent
The searcher is ready to act: hire, buy, book, call. Service pages, city pages and product pages are built for this intent, and they should make the next step obvious.
Local intent
A query where the searcher expects nearby results, whether or not they typed a place name. Google infers location from the device, so "plumber" and "plumber near me" return the same kind of local result.
Keyword map
A document that assigns exactly one primary keyword and one job to every page on the site. It is the cheapest possible protection against cannibalization and duplicate pages.
Keyword cannibalization
Two or more pages on the same site competing for the same query, splitting the links and engagement that should have concentrated on one. Fix it by merging the weaker page into the stronger one with a 301 redirect, or by giving each page a genuinely different job.
Branded and non-branded queries
Branded queries contain your business name; non-branded queries do not. Separating them in reporting matters, because branded traffic mostly measures your reputation while non-branded traffic measures your SEO.
SERP analysis
Reading the current results for a query before writing anything, to see what format, depth and angle Google is already rewarding. It is the closest thing SEO has to a briefing document.
Seed keyword
The starting phrase you feed a research tool to generate everything else. Take it from how customers describe the problem, not from how your industry describes the service.
Keyword clustering
Grouping phrases that return substantially the same results so one page can serve all of them. Clustering by overlap in the actual results, rather than by similar wording, is what stops a site publishing five pages that compete with each other.
Keyword gap analysis
Comparing the phrases competitors rank for against the phrases you rank for, to find demand you are not serving. It is a source of ideas, not a to-do list, because most gaps are not worth closing.
Zero-volume keyword
A phrase a tool reports as having little or no monthly search volume, which for a local business often means only that the tool cannot measure it. Some of the most profitable queries a small business receives never appear in a volume estimate.
Modifier
A word added to a base keyword that changes the intent behind it: near me, cost, best, emergency, same day, reviews, open now. Modifiers are usually where a small site can actually compete.
Question keyword
A query phrased as a question, usually informational. These are the natural material for FAQ sections, People Also Ask and AI answers, and each one belongs on the single page that owns that topic rather than being repeated across the site.
Buyer journey stage
Where a searcher sits between noticing a problem and hiring somebody: awareness, consideration or decision. Matching the page to the stage is the difference between traffic and phone calls.
Intent mismatch
Publishing a page whose purpose does not match what searchers want from that query, such as a sales page aimed at a research query. It shows up either as rankings with no conversions or as a page that will not rank at all.
The most expensive keyword mistake I see is a service page built for a research query. The business ranks, gets traffic, and gets no calls, because everybody landing on it wanted to read, not to buy. Check what Google already ranks before you write a word.
David CraggTucson SEO consultant, in search since 1990On-page SEO
On-page SEO is everything inside a page that tells a search engine what the page is for and why it deserves the click. It is the part of SEO you control completely, which also makes it the part with no excuses. This is the daily work of SEO content creation.
Title tag
The HTML title of a page, and the strongest single on-page signal of what it covers. Put the important words first, keep it near 60 characters so it is not truncated, and never repeat it on another page.
Meta description
The summary you propose for the search listing. It is not a ranking factor, but it heavily influences whether people click. Google rewrites it whenever it thinks a different sentence fits the query better.
H1
The main visible headline of the page. One per page. It should say plainly what the page is, in the language a customer would use, not in internal jargon.
Heading hierarchy
The H2 to H6 structure under the H1. Correct nesting turns a page into an outline that both a screen reader and a language model can follow. Skipping levels for visual size is a common and avoidable mistake.
URL slug
The readable part of the address after the domain. Short, lowercase, hyphenated, and descriptive. Changing a slug on a page that already ranks costs you unless you add a 301 redirect.
Internal link
A link from one page of your site to another. Internal links tell Google which of your pages matter, how they relate, and how to reach them. They are the most underused lever in small business SEO.
Anchor text
The visible words in a link. They describe the destination to both readers and crawlers, so "Tucson technical SEO services" teaches Google something that "click here" does not.
Alt text
The written description of an image, read aloud by screen readers and used by search engines to understand the picture. Describe the image honestly. Do not stuff keywords into it.
Image optimization
Serving pictures at the size they display, in a modern format such as WebP or AVIF, with width and height set in the HTML. It is usually the fastest route to a better Largest Contentful Paint score.
Above the fold
What a visitor sees before scrolling. On a phone that is a small space, so the first screen has to establish what the business does, where it is, and what to do next.
Keyword placement
Using the target phrase where it carries weight and reads naturally: the title, the H1, the opening sentences, one or two subheadings, and the URL. Beyond that, repetition adds nothing.
Semantic SEO
Writing about the whole subject rather than one phrase, using the related terms a knowledgeable person would naturally use. Search engines match meaning, not just strings.
Topic cluster
A group of closely related pages that link to each other and to a central page on the subject. Clusters are how a small site demonstrates depth instead of publishing scattered one-offs.
Pillar page
The main, comprehensive page at the center of a cluster, which links out to the supporting pages and receives links back from them.
Jump links
In-page anchor links, often collected in a table of contents, that let a reader move to a specific section. On a long reference page they improve usability and give AI systems a precise place to point.
Breadcrumb navigation
A visible trail such as Home > Content Creation > Advanced SEO Glossary, ideally matched by BreadcrumbList structured data. It tells both people and crawlers where a page sits in the site.
Word count
The number of words on a page. It is not a ranking factor. Length is a consequence of covering a subject properly, and padding a page to hit a target actively hurts it.
Readability
How easily a page can be read by the person it is written for. Short sentences, plain verbs, one idea per paragraph. Readability scores are a rough guide, not a target to game.
Call to action
The instruction that tells a reader what to do next: call, book, request a quote. Every page needs one, and it should match how far along the buying decision that page's visitor is.
Boilerplate content
Blocks of identical text repeated across many pages, most often on templated city pages. Google can tell the difference between a genuinely local page and the same page with the town name swapped.
Title link
Google's own name for the clickable headline in a result. Google generates it from your title tag most of the time, but it will substitute your H1 or other on-page text when it judges those a better match for the query.
Open Graph and social preview tags
The og: and twitter: tags that control how a link looks when it is shared in social platforms and messaging apps. They are not ranking signals. They decide whether a shared link looks deliberate or looks broken.
Image file name
The name of the image file itself. A descriptive file name is a small free signal about the subject of the picture, and it costs nothing to set before uploading.
Contextual link
A link placed inside body copy with relevant text around it. It carries more weight than the same link repeated in a menu or a footer, because the surrounding words describe the destination.
Publish and updated dates
The dates shown on a page and carried in its structured data. They should be accurate and should reflect real revisions, because changing a date to look current without changing the content is the oldest trick in publishing and it fools nobody worth fooling.
Byline
The visible credit naming who wrote the page, ideally linked to a bio. It is the human-readable half of an author entity, the structured data is the machine-readable half, and a serious page carries both.
Content formatting
How a page is broken up for reading: subheadings, short paragraphs, lists, tables and emphasis. Formatting is not decoration. It decides whether a page can be skimmed by a person and lifted cleanly by an answer engine.
Intrusive interstitial
A popup or overlay that blocks the content a visitor arrived to read, especially on a phone. Google has treated this as a problem since 2017, and legal notices such as consent and age gates are treated more leniently than marketing popups.
Ad density
How much of the visible page is taken up by advertising and promotional units before the content starts. A heavy ad load above the fold is a documented quality problem and a reliable way to lose the visitor before they read a word.
Content parity
Serving the same content, links and structured data on mobile as on desktop. Because indexing is mobile-first, anything that exists only in the desktop version is effectively not on the site.
I still find sites where every page shares one title tag, usually the business name. That single fix, done across a small site in an afternoon, has produced more measurable movement for my clients than most link campaigns.
David CraggTucson SEO consultant, in search since 1990Content quality and E-E-A-T
This is the section that has changed most since 2023. Google's quality systems are now aimed squarely at whether a real, identifiable person with real experience stood behind the page. Both 2026 core updates pushed further in that direction.
E-E-A-T
Experience, Expertise, Authoritativeness and Trustworthiness. It is not a score and not a ranking factor. It is the framework Google's quality raters use to describe what good looks like, and it maps closely to what the ranking systems reward.
Experience
The first E, added in December 2022. It asks whether the writer has actually done the thing: used the product, treated the patient, run the campaign. First-hand detail is the hardest signal to fake and the easiest to demonstrate.
Expertise
Demonstrated depth in the subject. For most topics this means visible knowledge and a track record. For medical, legal and financial subjects, formal credentials carry much more weight.
Authoritativeness
Whether other credible sources treat you as a reference: citations, links, press, listings and mentions you did not create yourself.
Trustworthiness
The one Google calls most important. Accurate information, a real business identity, working contact details, secure pages, honest claims, and no gap between what the page promises and what the business delivers.
YMYL
Your Money or Your Life: topics that can affect health, finances, safety or legal standing. Google applies a much higher quality bar to these pages, which is why thin medical or financial content struggles.
People-first content
Google's term for content made to help the reader, rather than content made to rank. The published self-assessment questions are worth reading before you publish anything.
Thin content
A page that exists but does not satisfy anyone: a few generic paragraphs, no detail, nothing a reader could not get in one sentence elsewhere. Thin pages drag down the site around them.
Scaled content abuse
Google's spam policy covering mass-produced pages made mainly to rank, whether written by people, by AI, or by both. The policy is about intent and value, not about which tool typed the words.
Site reputation abuse
Sometimes called parasite SEO: renting out sections of a trusted domain to third-party content that has nothing to do with the host site. Google added it as a named spam policy in 2024 and has enforced it since.
Expired domain abuse
Buying a domain that already has links and history and repurposing it for unrelated content in order to inherit its authority. Another named spam policy.
Content decay
The slow loss of rankings and traffic as a page ages, competitors publish better, and facts go stale. Reference content decays fastest, which is why a glossary needs a review date and a maintainer.
Content refresh
Rewriting and re-verifying an existing page instead of publishing a new one. Usually a better investment than new content, because the existing page already has links, history and internal support.
Author entity
A named, consistent, verifiable person credited on the page, with a bio page, a photo, credentials and Person structured data. It is what allows a reputation to attach to the writing.
Original research
Information a reader cannot get anywhere else: your own data, tests, photographs, case results or field observations. In 2026 this is the clearest way to be worth citing, both to journalists and to AI systems.
Content pruning
Deliberately removing, merging or noindexing pages that are thin, duplicated or obsolete. Deleting the weakest quarter of a small site often lifts the rest of it.
Topical authority
The accumulated evidence that a site covers a subject thoroughly rather than occasionally. It comes from covering a topic completely and linking those pages together, and it is why a focused small site can outrank a large unfocused one.
Information gain
What a page adds that the pages already ranking do not have. If a draft could be assembled entirely from the current top five results, it has no information gain and no reason to exist.
Content audit
A full inventory of every page on a site with its traffic, rankings, purpose and condition, ending in one decision per page: keep, rewrite, merge, redirect or delete.
Content brief
The document telling a writer what a page has to accomplish: the query, the intent, the questions to answer, the sources, the internal links and the call to action. Most bad content is a briefing failure rather than a writing failure.
Programmatic SEO
Generating many pages from a data set and a template, such as one page per city or per model number. It is legitimate when each page carries genuinely different information, and it becomes scaled content abuse the moment it does not.
Supplementary content
The parts of a page that are not the main content: related links, navigation, calls to action, author blocks. Google's rater guidelines count helpful supplementary content as a positive and distracting supplementary content as a negative.
Editorial standards
The stated rules for how a site's content is sourced, reviewed, dated and corrected. Publishing them is one of the cheapest trust signals available, and honoring them is what makes the claim worth anything.
Evergreen content
Content whose value does not depend on when it was published. Nothing is permanently evergreen. A reference page stays evergreen only because somebody keeps reviewing it.
Content velocity
How often new content is published. There is no correct rate, and publishing weekly is an advantage only if each piece is worth reading. Volume without value is the exact pattern the spam policies were written for.
Adding a named author with a real bio, a photo and a verifiable history is the cheapest E-E-A-T fix there is, and most small sites still publish under "admin." If nobody on the page has a name, nothing on the page has a reputation.
David CraggTucson SEO consultant, in search since 1990Entities and the Knowledge Graph
Search stopped being about strings and started being about things more than a decade ago. This vocabulary describes how a search engine, and now an AI system, decides that a business is a specific known thing rather than a name that happens to appear on a page. It is also the layer where schema markup earns its keep.
Entity
A thing a search engine can identify and hold facts about: a person, a business, a place, a product, a concept. Search moved from matching strings to recognizing entities years ago, which is why being consistent about who you are matters as much as keywords.
Knowledge Graph
Google's database of entities and the relationships between them. Being represented in it is what lets Google treat your business as a known thing rather than as a string of characters on a page.
Knowledge panel
The box of facts about an entity shown beside or above the results. It is generated from the Knowledge Graph rather than from your website, and it is claimed and corrected rather than created.
sameAs
The schema property that links your entity to its other verified profiles: a Google Business Profile, a LinkedIn page, a professional register, a Wikipedia entry. It is how you tell a machine that these accounts are all the same organization or person. Only list profiles that are real and that you control.
Entity home
The single page that best represents an entity, such as an About page for a person or a home page for a business. Everything else should point at it, and the structured data should name it consistently.
Brand SERP
What appears when somebody searches your business name. It is the results page you have the most influence over and the one a prospect checks before calling you.
Disambiguation
Distinguishing your entity from others with a similar name. A consistent address, category, credentials and sameAs list are what stop a machine merging you with somebody else.
Wikidata
The open, structured database of entities that many systems, including AI models, read as a reference. Most small businesses do not qualify for an entry, and paying somebody to create one is not a strategy.
Semantic triple
The subject, predicate, object structure underneath structured data: this business employs this person, this person holds this degree. Schema markup is a way of writing triples that a machine can read without guessing.
Entity consistency
Saying the same things about your business everywhere: name, address, phone, categories, founding date, credentials. Contradictions between your site, your listings and your markup are the most common reason a machine is unsure who you are.
Links and off-page signals
Off-page SEO is everything that happens away from your website: who links to you, who mentions you, and what the rest of the web says you are. Links still matter, but the quality bar has risen and the shortcuts have all been closed.
Backlink
A link from another website to yours. Treated as a signal of trust, weighted by how relevant and credible the linking page is. Ten links from respected sites in your field beat a thousand from directories.
Referring domain
The unique website a backlink comes from. Fifty links from one site count for far less than fifty links from fifty sites, so referring domains is the more honest number to track.
Link equity
The ranking value passed along a link, informally called link juice. It flows through internal links as well as external ones, which is why your own site structure is a link strategy.
Follow link
A normal link with no attribute limiting it, which passes ranking value to the destination. There is no rel="dofollow" attribute; the absence of a restriction is what makes it a follow link.
nofollow
A link attribute telling search engines you do not vouch for the destination. Since 2019 Google treats it as a hint rather than an instruction. Nofollow links still bring traffic and still build recognition.
rel="sponsored"
The attribute for paid or affiliate links. Using it correctly is how you accept paid placement without breaking the link scheme policy.
rel="ugc"
The attribute for links inside user-generated content such as comments and forum posts.
Anchor text distribution
The mix of wording in the links pointing at a page. A natural profile is mostly brand names and page titles. A profile stuffed with exact-match commercial phrases looks bought, because it usually was.
Link building
Earning links deliberately, through work worth citing, useful data, local sponsorships, partnerships, memberships and press. If the tactic could be automated, it has probably already been devalued.
Digital PR
Earning coverage and links from publications by giving journalists something genuinely newsworthy: original data, expert comment, or a local story. The highest quality links most businesses can realistically get.
Link scheme
Any arrangement to manipulate links, including buying them, excessive reciprocal linking, and automated exchanges. It is a published spam policy and a common cause of manual actions.
Private blog network (PBN)
A network of sites built or bought purely to link to a client's site. It is a link scheme by definition, it is detectable, and the damage outlives the agency that sold it.
Guest post
An article published on someone else's site. Legitimate when the audience is real and the piece is genuinely written for them. A spam policy violation when it is bulk placement bought for the link.
Unlinked brand mention
Your business named somewhere without a link. It still contributes to how search engines and AI systems recognize your business as an entity, and it is often an easy link to request.
Disavow file
A file submitted in Search Console telling Google to ignore specific links to your site. It is a last resort for genuinely toxic profiles. Most sites should never use one, and misusing it can remove links that were helping.
Domain Authority and Domain Rating
Third-party scores from Moz and Ahrefs estimating link strength on a 0 to 100 scale. They are vendor metrics. Google does not use them, does not see them, and does not rank by them.
PageRank
The original Google algorithm for scoring pages by the links pointing at them, named after Larry Page. The public toolbar score is long gone, but link analysis built on the same idea is still one of Google's named ranking systems.
Link velocity
The rate at which new links appear. Steady accumulation looks like a business earning attention. A sudden spike with no news event behind it looks like a purchase, because it usually is one.
Editorial link
A link somebody chose to give you because the page was worth citing, with no payment and no exchange. It is the only kind Google's policies fully endorse.
Link reclamation
Finding places that already mention you, once linked to you, or link to a page that has moved, and getting the link added, restored or repointed. It is the cheapest link work available because the relationship already exists.
Co-citation
Being mentioned alongside the same companies, topics or places repeatedly, with or without a link. It contributes to how search systems classify what your business is and who it competes with.
Sitewide link
A link appearing on every page of another site, usually in a footer or a sidebar. Google generally treats it as a single link, and a paid sitewide link is a link scheme with an unusually wide footprint.
Link gap analysis
Listing the sites that link to several competitors but not to you. It produces realistic prospects rather than a checklist, and most of the list will still not be worth pursuing.
Toxic link
A vendor label for a link a tool's model dislikes. Google devalues most bad links automatically, so the label sells cleanup services far more often than it identifies a real problem.
Trust Flow, Citation Flow and Authority Score
More third-party link metrics, from Majestic and Semrush. Like Domain Authority they are models built from a vendor's own crawl, useful for comparing prospects to each other and meaningless as a target or a promise.
I have never bought a link, in all the years I have been doing this, and I have cleaned up after plenty of people who did. The recoveries take longer than the rankings ever lasted. Earn the link or skip it.
David CraggTucson SEO consultant, in search since 1990Local SEO
For a business with a service area, local search is the highest-value ground on the internet. It is also a separate ranking system with its own signals, which is why a site can rank well organically and be invisible on the map. This is the core of Tucson SEO services.
Local SEO
Optimizing so a business appears for searches from and about a specific area. It combines a website, a Google Business Profile, listings across the web, and reviews.
Google Business Profile
The free Google listing that controls how a business appears in Maps and the local pack. For most local businesses it is the single highest-return asset in search, and it needs maintenance, not just setup.
Local pack
The map and three business listings shown at the top of local results. Placement here usually gets more clicks than the first organic listing, and it is governed by different signals.
Local finder
The expanded list of businesses you reach by clicking "More places" under a local pack. It ranks more businesses and is worth tracking separately from the top three.
NAP
Name, address and phone number. The three facts that identify a business across the web. They must match exactly everywhere, down to punctuation and suite numbers.
Citation
Any listing of your business details on another site: directories, chambers of commerce, industry bodies, data aggregators. Quality and consistency matter far more than quantity.
Citation consistency
The state of every listing carrying identical details. Conflicting phone numbers or old addresses are one of the most common and most fixable causes of weak local rankings.
Primary category
The main category on a Google Business Profile, and one of the strongest local ranking signals available. Choosing it precisely is worth more than most on-page work.
Service area business
A business that travels to customers rather than serving them at an address. It can hide its address on Google and define a service area instead, but it still needs a real, verifiable location.
Proximity, relevance and prominence
The three factors Google names for local ranking. Proximity is distance from the searcher, relevance is how well the listing matches the query, prominence is how well known the business is. You can influence two of the three.
City page
A page built for one city or neighborhood you serve, with genuinely local content: real projects, local landmarks, local pricing, local questions. Only build one for a place you actually serve.
Review velocity
The steady pace at which new reviews arrive. A slow, continuous stream reads as normal. Twenty reviews in one week after two silent years reads as a campaign.
Review response
Replying to reviews, good and bad. It is visible to customers, encourages more reviews, and is a public demonstration of how the business behaves. This is the day-to-day of reputation management.
Google Business Profile post
A short update published directly to the profile. Posts keep the listing active and can carry offers, events and news into the local result.
Suspension and reinstatement
Google can suspend a Business Profile for policy problems such as a mismatched address, keyword stuffing in the business name, or a virtual office. Reinstatement means proving the business is real at the address claimed.
Local link
A link from another organization in your area: a chamber, a supplier, a sponsored team, a local news site. These are worth more for local ranking than national links of similar strength.
Geo-modifier
A place name added to a keyword, as in "SEO consultant Tucson." Useful in titles and headings, damaging when stuffed into every sentence.
Secondary map platforms
Apple Business Connect and Bing Places. Smaller than Google, quick to claim, and increasingly used as source data by voice assistants and AI tools.
Near me search
A query containing the words near me, which Google answers using the searcher's location rather than the phrase itself. You do not need near me in your copy. You need to be genuinely close, relevant and prominent.
Secondary categories
The additional categories on a Google Business Profile beyond the primary one. They broaden the queries you can appear for, and adding categories you do not actually serve dilutes the profile rather than expanding it.
Profile attributes
The structured facts on a Google Business Profile: wheelchair access, veteran owned, free estimates, appointment required, online estimates. They appear as filters and labels in Maps, so they can decide whether you are shown at all.
Justifications
The line Google shows under a local listing explaining why it matched, such as a phrase from your website or a quote from a review. Pages and reviews that use the customer's own words are what produce them.
Questions and answers on a profile
The public Q and A section on a Google Business Profile, which anybody can answer, including people who have never worked for you. Posting and answering your own common questions is basic hygiene.
Data aggregator
A company that distributes business listing data to many directories at once. Correcting a bad record at the aggregator is usually faster than fixing the same error on fifty sites by hand.
Location page
A page for a physical location you operate, with its own address, hours, staff, photos and directions. Distinct from a city page, which covers an area you serve without having a location in it.
Geo-grid rank tracking
Checking local rankings from many simulated points across a map instead of from one. It shows how far your visibility actually extends, which a single average position hides completely.
Duplicate listing
A second Google listing for the same business, usually left behind by an old address, a franchise system or a data aggregator. Duplicates split reviews and confuse Google, and they should be merged or removed rather than ignored.
Review gating
Filtering customers so that only the happy ones are asked for a public review. It violates Google's policy, it is visible in the review pattern, and it suppresses the only feedback that would have improved the business.
Most of the local wins I get in the first ninety days are not clever. They are a complete Google Business Profile, the right primary category, consistent address formatting everywhere, and a real page for each city served. Boring work, reliable results.
David CraggTucson SEO consultant, in search since 1990Technical SEO
Technical SEO is the plumbing: whether search engines can reach your pages, render them, understand them and store them. None of it wins on its own, and all of it can quietly cap everything else. These are the terms behind a Tucson technical SEO engagement.
Crawling
Search engines following links and sitemaps to discover pages. If nothing links to a page and it is not in your sitemap, assume it does not exist as far as Google is concerned.
Googlebot
Google's crawler. It now visits almost every site as a smartphone user agent. You can confirm what it saw for any URL with the URL Inspection tool in Search Console.
Indexing
Storing and analyzing a crawled page so it can be returned in results. Crawled is not indexed, and indexed is not ranking. These are three separate hurdles.
Crawl budget
Roughly how much crawling Google will spend on a site. It matters for very large sites, and almost never for a small business site. If someone raises it for your fifty-page site, ask why.
robots.txt
A text file at the root of a domain telling crawlers where not to go. One careless line can block an entire site. It controls crawling, not indexing, so it is the wrong tool for keeping a page out of results.
noindex
A meta tag or HTTP header telling search engines not to store a page. The correct way to keep a page out of results. It only works if the page is crawlable, so never combine it with a robots.txt block.
Canonical tag
A tag naming the preferred URL when the same content is reachable at more than one address. Google treats it as a strong hint, not a command, and will override it if other signals disagree.
Duplicate content
The same or near-identical content at multiple URLs, on your site or across sites. It is not a penalty. It is a dilution problem: signals split between versions and Google picks one, sometimes the wrong one.
301 redirect
A permanent redirect that sends visitors and ranking value from an old URL to a new one. Mandatory whenever a URL changes. Without it, everything that page had earned is gone.
302 redirect
A temporary redirect. Correct for genuinely short-term moves, wrong for permanent ones. Leaving a 302 in place for years is a common and quiet source of lost authority.
Redirect chain
A URL that redirects to another that redirects again. Each hop adds delay and risk. Point every old URL directly at its final destination.
404
The response for a page that does not exist. Perfectly normal in small numbers. A problem when links from other sites, or your own menu, still point at it.
Soft 404
A page that returns a success code while showing nothing useful, such as an empty category or a "no results" page. Google flags these because they waste crawling and disappoint searchers.
XML sitemap
A machine-readable list of the URLs you want indexed, submitted through Search Console. It supplements internal linking; it does not replace it.
HTTPS
The encrypted version of HTTP, shown by the padlock in the browser. A confirmed but lightweight ranking signal since 2014, and a baseline expectation now. Serving a site over plain HTTP costs trust before it costs rankings.
Core Web Vitals
Google's three field-measured page experience metrics: Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift. They are assessed at the 75th percentile of real Chrome user data over a rolling 28-day window.
Largest Contentful Paint (LCP)
How long the biggest visible element takes to render. The good threshold is 2.5 seconds. For most small business sites the culprit is one oversized hero image.
Interaction to Next Paint (INP)
How quickly the page visibly responds to a tap or click. The good threshold is 200 milliseconds. INP replaced First Input Delay in March 2024.
Cumulative Layout Shift (CLS)
How much the layout jumps around while loading. The good threshold is 0.1. Usually caused by images without width and height, or by ads and banners injected after render.
Field data and lab data
Field data is what real Chrome users experienced, and it is what counts for Core Web Vitals. Lab data is a simulated test run in a controlled environment. A perfect lab score with failing field data means real visitors are having a worse time than your test did.
Time to first byte (TTFB)
How long the server takes to start responding. Slow TTFB puts a floor under every other speed metric, and it is usually a hosting or caching problem rather than a design problem.
Render-blocking resource
A CSS or JavaScript file the browser must fetch before it can display anything. Reducing these is normally the second biggest speed win after images.
JavaScript rendering
Content that only exists after JavaScript runs. Google can render it, but rendering is queued and can be delayed or fail. Anything that must rank is safer in the initial HTML.
Mobile-first indexing
Google indexes and ranks using the mobile version of a page. Rollout finished for all sites in 2024. If content or links are hidden on mobile, treat them as missing.
Structured data
Code that labels what is on a page in a vocabulary search engines and AI systems already understand: a business, a person, an article, a defined term. It does not raise rankings by itself; it removes ambiguity.
JSON-LD
The script format Google recommends for structured data. It sits in the page's code separately from the visible content, which makes it far easier to maintain than markup woven into the HTML.
Orphan page
A page with no internal links pointing to it. It may be in the sitemap and still be treated as unimportant, because nothing on the site vouches for it.
Crawl depth
How many clicks it takes to reach a page from the home page. Anything important should be reachable in three clicks or fewer.
HTTP status code
The three digit code a server returns with every response: 200 for success, 3xx for a redirect, 4xx for a client error such as a missing page, 5xx for a server failure. Diagnosing most SEO problems starts by reading the code rather than the page.
410 Gone
The response for a page you removed deliberately and are not bringing back. It tells Google to drop the URL faster than a 404 does, which makes it the right code for pages you pruned on purpose.
5xx server error
A failure on your server rather than in the request. Occasional 5xx responses are ignored. Sustained ones cause Google to slow its crawling and eventually drop pages, which is how a hosting problem becomes an indexing problem.
X-Robots-Tag
An HTTP header carrying the same directives as the robots meta tag. It is the only way to apply noindex to files that are not HTML, such as PDFs and images.
Snippet controls
The directives limiting what Google may display from a page: nosnippet, max-snippet, max-image-preview and the data-nosnippet attribute for specific passages. Every one of them reduces visibility, so use them deliberately rather than defensively.
Google-selected canonical
The URL Google actually chose to index, which can differ from the one you declared. Search Console reports both, and the gap between them explains a great many pages that seem to have vanished.
URL parameters
The query strings appended after a question mark for filtering, sorting, sessions and tracking. Left unmanaged they turn a handful of real pages into thousands of near-duplicate URLs.
Faceted navigation
Filter systems on listing pages that generate a new URL for every combination of choices. It is the most common cause of index bloat on larger sites, and it is controlled with canonicals, robots rules and link handling rather than left to chance.
Pagination
Splitting a long list across numbered pages. Google stopped supporting rel=next and rel=prev in 2019, so what matters now is that every paginated URL is a real, crawlable, self-canonical page linked from the others.
Index bloat
Having far more URLs indexed than you have real pages, usually from parameters, filters, tag archives or auto-generated stubs. It dilutes the site and buries the pages that were supposed to matter.
Log file analysis
Reading the server's own record of what Googlebot actually requested and when. It is the only source that shows real crawler behavior rather than a tool's estimate of it.
Server-side and client-side rendering
Whether the HTML arrives complete from the server or is assembled in the browser by JavaScript. Server-rendered content is available to every crawler immediately. Client-rendered content has to wait for a second pass that may be delayed.
Dynamic rendering
Serving a pre-rendered version to crawlers and the JavaScript version to people. Google described it as a workaround rather than a recommendation and has advised against relying on it since 2022.
Preload, preconnect and fetchpriority
Instructions telling the browser what to fetch first. Preloading the hero image and marking it high priority is often the single cheapest Largest Contentful Paint improvement available on a small business site.
First Contentful Paint (FCP)
When the first piece of content appears on screen. It is not a Core Web Vital, but it is a useful early warning that a slow server or render-blocking files are holding everything else up.
Total Blocking Time (TBT)
A lab measurement of how long the main thread was too busy to respond to input. It is the closest a lab test can get to Interaction to Next Paint, which can only be measured from real visitors.
Chrome User Experience Report (CrUX)
Google's public dataset of measurements from real Chrome users, and the source of the field data behind Core Web Vitals. A site with too little traffic will have no CrUX data at all, which is a gap rather than a failure.
Site architecture
How pages are grouped and linked: what sits under what, what links to what, and how deeply anything is buried. Architecture decides how authority moves through a site, and it is far harder to change later than copy is.
Site migration
Any change that alters URLs at scale: a new domain, a new platform, a restructure, a move to HTTPS. It is the highest-risk work in SEO and the most common cause of a sudden permanent traffic loss.
Redirect map
The sheet pairing every old URL with its single best new URL, written before a migration goes live. Pointing everything at the home page instead is the classic shortcut and the classic disaster.
IndexNow
An open protocol for notifying search engines the moment a URL is added, changed or removed. Bing, Yandex, Naver and Seznam support it. Google does not, so it supplements a sitemap rather than replacing one.
Rich Results Test and Schema Markup Validator
The two tools worth running on markup. Google's Rich Results Test shows what Google can currently use. The Schema.org validator shows whether the markup is valid at all. A page can pass one and fail the other.
Site crawler
Software that crawls your site the way a search engine would and lists status codes, titles, canonicals, redirects, orphan pages and broken links. Running one is the first step of any technical audit.
The worst technical problem I have ever found took one line to fix. A developer had left a noindex tag on a staging site, pushed it live, and nobody looked for eight months. The business assumed its content was failing. Its content was invisible.
David CraggTucson SEO consultant, in search since 1990Images, video and other search surfaces
Search is not only ten links and a map. Images, video and platform search each have their own ranking behavior, and each rewards the same unglamorous work: real media of your own, described accurately, reachable in the HTML rather than injected by a script.
Image SEO
Making images findable in their own right: descriptive file names, honest alt text, correct width and height, a modern format, and the image present in the HTML rather than loaded by script.
Image and video sitemaps
Extensions to the XML sitemap format that list media a crawler would otherwise struggle to discover, particularly images loaded by JavaScript and videos inside players.
Video SEO
Getting video found, whether it is hosted on your own site or on a platform. The page around the video does most of the work: a real title, a written summary, a transcript, and markup describing the file.
Video structured data and key moments
Markup describing a video's title, thumbnail, duration and upload date, optionally with timestamped sections. Key moments let a searcher jump straight to the part that answers their question.
Transcripts and captions
The written version of spoken content. Captions serve viewers who cannot use sound, transcripts give search engines and AI systems something to read, and both come out of the same file.
YouTube search
Search inside YouTube, which ranks on its own signals: title, description, engagement and watch time. A video can dominate YouTube and be invisible in Google, and the reverse is just as common.
International and multilingual SEO
Most local businesses never need this section. It starts to matter the moment a site serves more than one language or country, because the failure mode is not lower rankings. It is Google showing the wrong version of the page to the wrong person.
hreflang
Annotations telling Google which language and regional version of a page to show to whom. They have to be reciprocal: every version points at every other version, including itself, or Google ignores the set.
x-default
The hreflang value naming the fallback page for anyone whose language or region you do not have a specific version for.
ccTLD
A country-code domain such as .ca or .mx. It is the strongest geographic signal available and the most expensive to maintain, because each one is effectively a separate site with separate authority.
Subdirectories and subdomains for languages
The two practical alternatives to separate country domains: example.com/es/ or es.example.com. Subdirectories keep everything on one domain and one pool of authority, which is usually the right answer for a small business.
Localization and translation
Translation converts the words. Localization changes the currency, units, examples, phone formats, spelling and the phrases people in that market actually search. Machine translation published without review is a scaled content problem.
Geotargeting
Telling a search engine which country a site or a section is aimed at. The signals that do the work are the domain, the hreflang annotations, the language of the content and the details on the page such as address and currency.
E-commerce and product SEO
Selling products online adds a layer of vocabulary a service business never meets: feeds, variants, inventory states and a second set of Google systems that read your prices. These are the terms that come up in every store audit.
Product page
The page for one purchasable item. It has to carry the specifics a buyer compares on, including price, availability, dimensions, materials and shipping, because a page carrying only the manufacturer's description has nothing of its own to rank with.
Category page
A listing page grouping related products. On most stores these hold the real commercial search demand, and on most stores they carry no unique content at all.
Product structured data and merchant listings
The markup describing price, availability and condition, which makes a product eligible for shopping-related results. It has to match what the page actually shows, or the listing is rejected.
GTIN
The global trade item number identifying a manufactured product, such as a UPC or an EAN. Supplying it correctly is what lets a marketplace match your listing to the same product sold elsewhere.
Variant handling
Deciding whether every size and color gets its own indexable URL or whether they consolidate onto one page. Both approaches work. What fails is generating a separate URL for every combination and canonicalizing none of them.
Out-of-stock and discontinued products
What happens to a URL when the item goes away. Keep and update the page if the product returns or has a successor, redirect it to the closest equivalent if it does not, and never mass-delete a catalog into 404s.
Review snippets and the self-serving review policy
Star ratings shown in results, generated from review markup. Google does not allow ratings you collected about your own business, on your own site, to be marked up this way. The reviews have to be about a product or come from an independent source.
Product feed
The structured file of your inventory sent to Google Merchant Center or another marketplace. It is separate from your site's markup and can contradict it, which is a common and invisible cause of rejected listings.
Measurement and reporting
SEO reporting is where a lot of money quietly disappears. These terms are the difference between a report that shows whether the phone rang and a report that shows a chart going up.
Google Search Console
Google's free reporting tool for how a site performs in search. It is the only source of real query data and the first place to check indexing problems. Every site should have it verified.
Performance report
The Search Console view showing clicks, impressions, click-through rate and average position by query, page, country and device. Filter by page before drawing conclusions.
Clicks and impressions
Clicks are visits from search; impressions are appearances in results. Rising impressions with flat clicks usually means you are ranking on page two, or an AI answer is absorbing the click.
Average position
A weighted average of where your listings appeared. It hides more than it shows: one query moving from 40 to 8 can look identical to everything sliding slightly. Read it per query.
Click-through rate (CTR)
Clicks divided by impressions. Low CTR at a good position usually points at a weak title or a mismatch between the listing and what the searcher wanted.
Google Analytics 4 (GA4)
The current version of Google Analytics, built around events rather than pageviews. It answers what visitors did; Search Console answers how they found you. You need both.
Engaged session
A GA4 visit that lasted at least ten seconds, triggered a key event, or included two or more pageviews. It replaced the old bounce rate as the default quality measure.
Bounce rate
The share of visits with no meaningful engagement. It has never been a direct ranking factor, and on a page that answers a question completely, a high bounce rate can mean the page worked.
Key event
GA4's name for a conversion: a call, a form, a booking, a purchase. If key events are not configured, no amount of traffic reporting will tell you whether the work paid.
Attribution
Deciding which channel gets credit for a conversion. Search often does the introducing and something else gets the last click, which is why organic is routinely undervalued in default reports.
Rank tracking
Recording positions for chosen keywords over time. Useful as a trend, unreliable as a fact, because results vary by location, device and personalization.
Visibility index
A third-party estimate of how much of a market a site captures, from tools such as Sistrix or Semrush. Directional only. Nobody outside Google measures this properly.
Cost per lead
Total SEO spend divided by leads produced. The one number that decides whether the work continues, and the reason a flat monthly fee is easier to evaluate than an hourly one. See what my engagements cost and include.
Seasonality
The predictable annual pattern in demand. Compare year over year, not month over month, or you will diagnose an algorithm problem every winter.
Change log
A dated record of what was changed on the site and when. Without it, no ranking movement can be honestly attributed to anything.
Generative AI performance report
A Search Console report introduced in June 2026 giving a dedicated view of impressions from Google's AI features. It is rolling out gradually, and it is the first official window into AI-driven visibility.
Domain property and URL prefix property
The two ways to verify a site in Search Console. A domain property covers every subdomain and protocol. A URL prefix property covers only what matches exactly, which is how sites end up reporting on a fraction of their own traffic.
Page indexing report
The Search Console report showing which URLs are indexed and, for the rest, why not. It is the first place to look when a page is not ranking, because not indexed and ranking badly are different problems with different fixes.
Anonymized queries
Search Console hides queries made by very few people, to protect privacy. Your query totals will therefore never add up to your click totals, and the gap is largest on small local sites.
Regex filters
Regular expression filtering in Search Console, which lets you group queries by pattern: branded against non-branded, questions, city names, service words. It turns a flat query list into a report about intent.
Looker Studio
Google's free tool for combining Search Console, Analytics and other sources into one dashboard. It makes reporting prettier. It does not make the underlying numbers mean more than they did.
Search Console bulk data export
A daily export of unsampled Search Console data into BigQuery. It removes the row limits and the sixteen month history cap, and it is worth setting up early because it cannot backfill what it did not collect.
Organic search channel
The bucket in Analytics that attributes a visit to unpaid search. Broken tracking, redirects and some AI referrers routinely land in direct instead, which quietly understates what search produced.
UTM parameters
Tags added to a link so Analytics can attribute the visit to a specific campaign or placement. Never use them on internal links, and always use them on the website link in your Google Business Profile.
Call tracking
Recording which calls came from which source, using a tracked number, a dynamic number swap, or simply asking every caller. For most local businesses the phone is the conversion, and untracked calls make good SEO look like nothing happened.
SEO split testing
Changing one variable across a group of similar pages while holding a comparable group unchanged, then comparing the two. It is the only honest way to attribute a change, and it needs enough similar pages to be possible at all.
Share of voice
An estimate of how much of a keyword set's total search traffic a site captures. Useful for watching a market over time, and entirely dependent on which keywords the vendor put in the set.
Bing Webmaster Tools
Microsoft's equivalent of Search Console. It is free, it reports on a real if smaller share of searches, and Microsoft's index also sits behind Copilot, which makes it worth verifying even if Bing traffic is small.
| Metric | Source | What it can honestly tell you |
|---|---|---|
| Clicks and impressions | Google, via Search Console | Exactly how often you appeared and how often someone chose you. |
| Average position | Google, via Search Console | A weighted average that hides detail. Read it query by query or not at all. |
| Core Web Vitals | Google, from real Chrome user data | What real visitors experienced, at the 75th percentile, over the last 28 days. |
| Index coverage and manual actions | Google, via Search Console | Whether Google stored your pages, and whether a human penalized you. |
| Domain Authority, Domain Rating | Moz and Ahrefs models | One vendor's guess at link strength. Google does not use it. |
| Keyword difficulty | Vendor model | A rough triage score built mostly from links to the current top results. |
| Search volume | Vendor model | A comparison between phrases. Never a forecast, especially locally. |
| Visibility index | Vendor model | Direction of travel across a market. Not a measurement of your traffic. |
I send clients the same two screens I look at myself: Search Console clicks by page, and the calls and forms that resulted. If a report has thirty widgets and none of them is a lead, it was built to be admired, not read.
David CraggTucson SEO consultant, in search since 1990AI search: the vocabulary that is actually current
This is where most glossaries are out of date. In May 2026 Google published its first official guidance on optimizing for generative AI features, and much of what the industry had been selling as AI optimization was named in it as unnecessary. These definitions follow Google's own documentation.
AI Overviews
The AI-generated summary Google places above the results for many queries, with links to sources. It appears automatically and is not something you opt into.
AI Mode
Google's conversational search interface, where a searcher asks follow-up questions in a session rather than running separate queries. Google reported it passed one billion monthly users in May 2026.
Generative AI features
Google's own umbrella term for AI Overviews, AI Mode and related surfaces. Google states these run on the same ranking and spam systems as the rest of Search, not on a separate stack.
Generative Engine Optimization (GEO)
The industry term for trying to appear in AI-generated answers. Google's May 2026 guidance says plainly that GEO is still SEO, and that no separate technique set is required.
Answer Engine Optimization (AEO)
A near-synonym for GEO, focused on being the source an AI system quotes. Practically it means answering a specific question directly, early, in language a machine can lift without distortion.
Large language model (LLM)
The kind of AI system behind ChatGPT, Claude, Gemini and Perplexity. It generates text by predicting language, which is why it needs grounding in retrieved sources to be reliable about facts.
Grounding
Attaching an AI answer to retrieved documents rather than to the model's memory. Grounded answers carry citations, and being one of those citations is the realistic goal of AI-era SEO.
AI citation
A link or source credit inside an AI answer. It is the AI equivalent of a ranking position, and it tends to favor pages that are specific, attributed and internally consistent.
Zero-click search
A search resolved on the results page itself, through an AI answer, a featured snippet or a map pack, with no visit to any website. Zero-click results still build recognition, but they do not fill a calendar.
Query fan-out
The technique where an AI search system breaks one question into several related sub-questions, runs them, and assembles an answer from the results. It rewards pages that cover a subject completely rather than one phrase narrowly.
llms.txt
A proposed file for telling AI systems how to use a site. Google has stated explicitly that it is not needed for its generative AI features. Treat it as an experiment, not a requirement.
Content chunking
Restructuring content into machine-sized blocks specifically for AI retrieval. Google lists this among the tactics that are not necessary for its AI features. Clear headings do the same job for readers too.
Conversational query
A longer, spoken-style question typed into an AI interface, often with context carried from a previous question. These queries are more specific than traditional keywords and less predictable.
AI crawler controls
Directives such as Google-Extended, and robots.txt rules for crawlers like GPTBot, ClaudeBot and PerplexityBot, that let a site opt out of AI training or retrieval. Blocking them removes you from those answers, so decide deliberately.
Hallucination
An AI system stating something false with confidence. It is the reason AI output must never be published without a person checking the facts, and the reason clear, verifiable pages get cited over vague ones.
Non-commodity content
Google's own phrase from its 2026 AI guidance for content nobody else could have written: your results, your process, your evidence. It is the sharpest published statement of what now separates cited pages from ignored ones.
Retrieval augmented generation (RAG)
The architecture behind most AI answers: retrieve documents first, then write the answer from them. It is why being retrievable and quotable matters more than being long.
Embeddings and vector search
Representing text as numbers so that passages with similar meaning sit close together, which allows retrieval by meaning instead of by wording. It is the mechanism behind semantic matching in both search and AI systems.
Prompt
The instruction a person gives an AI system. Prompts are longer, more specific and more contextual than keywords, which is why AI answers surface pages that handle narrow questions properly.
Knowledge cutoff
The date after which a model has no training data. Anything later has to be retrieved at the moment of the question, which is why clearly dated, currently accurate pages get cited and stale ones do not.
Training data and retrieval
Two different routes your content can take into an AI answer: absorbed during training, or fetched when somebody asks. Blocking a training crawler and blocking a retrieval crawler have very different consequences for whether you are ever cited.
Common Crawl
A free public crawl of the web that has been a standard ingredient in AI training sets. Its crawler is CCBot, and it can be blocked in robots.txt like any other.
AI visibility tracking
Tools that ask AI systems the same set of questions repeatedly and record which brands and sources get named. The idea is sound and the numbers are unstable, because the same question asked twice can produce different sources.
AI agent
An AI system that browses, clicks and completes tasks on a person's behalf instead of returning links. Agents read the page a machine sees, which turns clean HTML, working forms and accurate business details into a commercial issue rather than a technical one.
Synthetic content
Content generated by a model. Google's stated position is that this is neither good nor bad in itself. What matters is whether a person is accountable for its accuracy and whether it was made to help anybody.
Multimodal search
Searching with an image, or with an image and text together, as in Google Lens. It rewards businesses with genuine, well-described photographs of their own work rather than stock library images.
My honest read after working through Google's AI optimization guide: there is no new discipline here. The sites getting cited in AI answers are the ones that were already specific, credited to a named person, and easy to parse. The novelty is in the reporting, not the strategy.
David CraggTucson SEO consultant, in search since 1990Spam tactics and black-hat vocabulary
You do not need this vocabulary to do SEO. You need it to recognize what somebody is proposing to do to your site, and to understand a Search Console message if one ever arrives. Most of what follows is named in Google's published spam policies. The rest is a documented way to waste money.
Keyword stuffing
Repeating a phrase unnaturally in copy, headings, alt text or a business name in order to rank for it. It is a named spam policy, it reads badly to customers, and in a Google Business Profile name it is grounds for suspension.
Cloaking
Showing search engines different content than you show people. It is one of the oldest violations, it is detectable, and Google treats it as deliberate because it is.
Hidden text
Text placed where a visitor can never see it: white on white, behind an image, positioned off screen, or inside an element with no way to open it. Content in a genuinely expandable section is fine. Content nobody can reach is not.
Doorway page
A page built only to catch a query and funnel the visitor somewhere else, most often a set of near-identical city pages. The test is whether the page would still be useful if the funnel were removed.
Sneaky redirect
Sending a visitor somewhere other than the page the search engine was shown, often based on device or referrer. It is cloaking with an extra step and it is handled the same way.
Link farm
A network of sites that exist to link to each other. It is the oldest form of link scheme and the easiest to detect, because the pattern is the entire point of it.
Hacked content
Pages or links injected into a site by an attacker, usually pointing somewhere else entirely. It gets flagged in Search Console, and it damages the site that was hacked rather than the attacker.
Negative SEO
Attempting to damage a competitor's rankings, usually by pointing spam links at them. Google devalues most of it automatically, and the phrase is used to sell link cleanup far more often than the attack is actually encountered.
Spun content
Text run through a rewriting tool to produce many near-identical versions. It predates AI writing by fifteen years, it has always been scaled content abuse, and it reads exactly as badly now as it did then.
CTR manipulation
Paying for automated or crowdsourced clicks on your listing in the belief that it lifts rankings. Google's systems are built to discount manufactured behavior, and the traffic it buys never converts.
How SEO work is bought and sold
The last set of terms is not about search engines at all. It is the vocabulary of the transaction, which is where most small business SEO actually goes wrong. What an engagement here costs and includes is a separate page. These are the words you need to read anybody else's proposal.
SEO audit
A structured review of a site's technical health, content and off-site profile, ending in a prioritized list of changes. An audit that produces a document and no sequence of work is a document, not a plan.
Retainer
A recurring fee for ongoing work. It suits SEO because the work is continuous rather than a project, and it is only fair to both sides when what it buys each month is stated plainly.
Scope of work
The written statement of what will be done, by whom, how often, and what is excluded. Most disputes in this industry are scope disputes that nobody wrote down at the start.
White-label SEO
An arrangement where the company you hired subcontracts the work and puts its own name on it. It is legal and common. The thing to establish before signing is who is actually going to touch your site.
Guaranteed rankings
A promise of a specific position. Nobody outside Google controls the results, so the guarantee is either meaningless, limited to phrases nobody searches, or hedged into nothing in the contract.
Ownership of your assets
Who holds the domain registration, the hosting, the website, the Google Business Profile and the analytics accounts. If it is not you, changing providers means starting over, and some companies depend on that.
Reporting cadence
How often you receive results and in what form. Monthly is normal. What matters is whether the report shows leads and names what changed, or only shows charts.
Churn and burn
A sales model built on signing many clients, doing shallow templated work, and replacing the ones who leave. It is recognizable by long contracts, no named practitioner, and no record of what changed on your site last month.
Terms that are retired, wrong, or a warning sign
Half of SEO literacy is knowing which words no longer mean anything. If a proposal you are reading leans on the terms in this section, it was written from an old playbook, and you should ask when it was last revised.
Keyword density
The percentage of a page made up of a target phrase. It has not been a useful ranking concept since the early 2000s, and optimizing toward a target percentage produces text nobody wants to read.
LSI keywords
A term borrowed from an unrelated 1980s information retrieval patent and sold as an SEO technique. Google has said directly that there is no such thing as LSI keywords. Use related terms because they help readers, not because of this.
Meta keywords tag
A list of keywords in the page code. Google stopped using it in 2009. Its only remaining function is telling competitors what you are targeting.
PageRank sculpting
Using nofollow on internal links to funnel authority to chosen pages. Google closed this off in 2009. Internal links still matter enormously; this particular trick does not.
Google Authorship
The rel=author markup that once put a photo next to results. Google retired it in 2014. Author identity still matters, but it is expressed now through Person structured data and a real bio, not that tag.
Exact-match domain strategy
Buying a domain such as bestplumbertucson.net expecting the name alone to rank it. Google specifically devalued this in 2012. The domain is now a branding decision.
Mass directory submission
Submitting a site to hundreds of directories for links. It has ranged from useless to harmful for over a decade. A handful of relevant, real local and industry listings still helps.
First Input Delay (FID)
The old responsiveness metric in Core Web Vitals. Replaced by Interaction to Next Paint in March 2024. Any guide still listing FID as current predates that change.
FAQ rich results
The expandable question dropdowns that used to appear under listings. Google restricted them to authoritative government and health sites in August 2023, and retired them entirely on May 7, 2026. FAQPage structured data is still a valid type and is still parsed, but it no longer produces that display.
HowTo rich results
Step-by-step rich results, deprecated by Google in 2023. Step-by-step content still helps readers; the search feature is gone.
Sitelinks search box markup
Structured data for a search box inside your Google listing. Google retired the feature in late 2024, and the markup no longer does anything.
"Google penalty"
Used loosely to mean any ranking drop. A penalty is specifically a manual action, visible by name in Search Console. Most drops are algorithmic recalibration or a competitor doing better, and the fix is completely different.
Cached page and the cache: operator
The stored copy of a page Google used to show from its results. Google removed the cached links in 2024 and retired the cache: operator with them, so the old habit of checking the cache to see what Google indexed is gone. Use the URL Inspection tool in Search Console instead.
Toolbar PageRank
The public 0 to 10 score Google once displayed in its browser toolbar. It stopped updating in 2013 and was removed in 2016. Link analysis still exists. The public number does not.
AMP
Accelerated Mobile Pages, a stripped-down page format Google once required for the Top Stories carousel. That requirement ended in 2021, and for almost every site building one fast normal page is now the better answer.
rel=next and rel=prev
Markup that once told Google how a paginated series fitted together. Google announced in 2019 that it had not used it for years.
Mobile usability report
The Search Console report on mobile rendering problems, retired in December 2023. Mobile experience still decides rankings. That particular report simply no longer exists.
Retired rich result types
Google has pruned its rich result formats in batches, dropping HowTo, then a group including Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement and Vehicle Listing, then FAQ. In each case the schema type stayed valid and the visible feature disappeared. Check Google's current structured data gallery before building anything on a rich result.
Position zero
An old nickname for the featured snippet, from when it sat directly above the first organic result. The results page no longer has a stable top, so the phrase describes a layout that has not existed for years.
Search engine submission
Paying to have a site submitted to search engines. Google finds pages by following links and reading sitemaps, and there are effectively two engines to be found in rather than hundreds.
Dwell time
The time between a searcher clicking a result and returning to the results page. It is quoted constantly as a ranking factor, it has never been confirmed as one, and no tool you own can measure it.
The Google sandbox
The belief that new sites are held back for a fixed probation period. Google has denied running a deliberate sandbox. New sites do take time, because they have no history, no links and no track record yet.
Duplicate content penalty
A penalty that does not exist. Duplicate content causes consolidation and dilution, not punishment. The exception is content copied at scale from other sites, which is a separate spam policy with real consequences.
I have been in search since 1990, and I have watched every one of these go from best practice to punchline. That is the actual lesson of a glossary: the vocabulary is a snapshot, and anybody who tells you SEO is settled is selling something.
David CraggTucson SEO consultant, in search since 1990Knowing the words is not the same as fixing the site
If you run a business in Tucson or Southern Arizona and you want to know which of these terms actually apply to your site, I will tell you. One consultant, $1,000 a month, month to month, no contract.
Get a free look at your site or call (520) 207-6000.
Not sure what you need yet? Start with an SEO audit, or read what working with me costs and includes.
Questions about SEO terminology
How many SEO terms do I actually need to know?
About twenty. If you understand crawling, indexing, ranking, search intent, title tag, internal link, backlink, citation, NAP, canonical tag, 301 redirect, schema markup, Core Web Vitals, E-E-A-T, keyword cannibalization, thin content, impressions, clicks, conversions and manual action, you can follow any competent SEO conversation and challenge any incompetent one.
The other three hundred and thirty terms on this page are for the person doing the work. You do not need them to hire well. You need enough vocabulary to ask what a recommendation will change, how it will be measured, and what happens if it does not work.
Which SEO metrics come from Google and which are invented by tool vendors?
Google publishes these: impressions, clicks, click-through rate and average position in Search Console; Core Web Vitals thresholds of 2.5 seconds for LCP, 200 milliseconds for INP and 0.1 for CLS; index coverage; and manual actions. These are measurements of what actually happened.
These are vendor estimates, not Google data: Domain Authority, Domain Rating, Trust Flow, spam score, keyword difficulty, search volume, and every visibility index. They are useful for comparing options and useless as promises. Google does not see them and does not rank by them.
The test is simple. Ask where a number comes from. If the answer is a tool rather than Google or your own site, treat it as an opinion with a decimal point.
Which SEO terms are obsolete in 2026?
Keyword density, LSI keywords, the meta keywords tag, PageRank sculpting, Google Authorship, exact-match domain strategy and mass directory submission are all long dead. First Input Delay was replaced by Interaction to Next Paint in March 2024.
Two more retired recently. Google retired the sitelinks search box in late 2024, and it retired FAQ rich results in Google Search on May 7, 2026, with the reporting and testing support removed through June and August 2026. FAQPage structured data is still a valid schema type and Google still parses it, but the expandable dropdowns are gone.
If a proposal you are reading still promises FAQ dropdowns in the search results, or quotes a keyword density target, it was written from a playbook that has not been revised in years.
What is the difference between SEO, AEO and GEO?
SEO is optimizing to be found in search. AEO, answer engine optimization, is optimizing to be the source an AI answer quotes. GEO, generative engine optimization, is the same idea aimed at generative systems such as AI Overviews, AI Mode, ChatGPT and Perplexity.
In May 2026 Google published its first official guidance on optimizing for generative AI features and stated that AEO and GEO are still SEO. The same guidance says you do not need an llms.txt file, AI-specific rewriting, content chunking, or any special schema markup to appear in AI Overviews or AI Mode.
So the practical difference is smaller than the marketing suggests. What earns an AI citation is what earned a good ranking already: a specific, accurate page, written by a named person, easy to parse, and consistent with what the rest of the web says about that business.
What does it actually mean when someone calls something a ranking factor?
It means the person believes Google's systems use that input. Google has confirmed relatively few by name, including HTTPS, mobile-friendliness, page experience signals, and the obvious ones like relevance and links.
Everything else on a list of two hundred ranking factors is inference from testing and correlation. Some of it is well evidenced. Some of it is folklore that has been repeated so long it sounds official.
The useful question is not whether something is a ranking factor. It is whether doing it makes the page better for the person reading it. That standard has survived every algorithm update; the factor lists have not.
What SEO jargon should make a business owner suspicious?
Guaranteed number one rankings, because nobody controls Google's results. Secret algorithms or proprietary ranking technology, because the working methods are public. Submission to hundreds of search engines, because there are effectively two. High DA link packages, because Domain Authority is a vendor score Google does not use.
Also be careful with anyone who reports only rankings and traffic and never leads, anyone who cannot name what changed on your site last month, and anyone who owns your website, domain or Google Business Profile rather than you.
Vagueness is the real warning sign. Every legitimate recommendation can be stated as a specific change, on a specific page, with a specific expected effect.
Do I need to understand this vocabulary to hire an SEO consultant?
No, but you need enough of it to tell a plan from a pitch. The point of a glossary is not to turn a business owner into a practitioner. It is to make the conversation honest.
A good consultant will explain any of these terms in a sentence, without irritation, and will connect each one to something that happens on your site. If a term cannot be translated into a specific change and a way to measure it, it does not belong in your proposal.
How is this glossary kept accurate?
It is maintained by one person, David Cragg, and it carries a review date at the top. Every definition that touches a Google behavior is checked against Google's own published documentation rather than against other glossaries, because most SEO glossaries copy each other and errors survive for years that way.
Definitions that were wrong get corrected rather than quietly deleted, and terms that have been retired stay on the page in the retired section, dated, so you can recognize outdated advice when you meet it.
The last full review was August 14, 2026, after Google's May 2026 core update, the retirement of FAQ rich results, and the publication of Google's generative AI optimization guidance.
Where to go next
The work these terms describe
Related reading
About the practice
Primary sources
Definitions that describe a Google behavior are checked against Google’s own documentation. These are the pages worth bookmarking yourself.
- Google Search Central: SEO starter guide
- Google Search Central: optimizing for generative AI features
- Google Search Central: AI features and your website
- Google Search Central: spam policies
- Google Search Central: introduction to structured data
- web.dev: Core Web Vitals
- Google Search Central: guidance on AI generated content
- Schema.org: DefinedTermSet
How this page is written and checked
I use AI openly as a drafting and research tool, the same way I use a keyword tool or a crawler. It helps me draft, organize and pressure test my own explanations, and it is good at catching a definition I have written sloppily.
What it is not allowed to do here: supply a fact about Google, invent a statistic or a source, decide what a page is for, or reach this site without me reading it. Every definition that describes a Google behavior is checked against Google’s published documentation, linked above. Where a claim has no primary source, it is either labeled as my own judgment or it is not on the page.
That is the line Google draws too. Its spam policies target scaled content abuse, meaning pages produced at volume with nobody accountable for them, not the use of a tool. See also Google Search and AI content.
Published April 25, 2026. Expanded from 176 terms to 350 and last fully reviewed August 14, 2026, covering the May 2026 core update, the retirement of FAQ rich results on May 7, 2026, and Google’s generative AI optimization guidance. Corrections are welcome: (520) 207-6000.
