AI SEO glossary
Every check Rubric runs, explained one page at a time: what it is, why it matters for SEO and for AI search such as ChatGPT and Google's AI Overviews, how Rubric scores it, how common the problem is on real sites, how accurate the check is, and how to fix it on WordPress, Shopify, Webflow or custom code.
122 checks
Bot access
- How do I stop or allow Copilot using my content?With Bing's robots rules.
ai-grounding:disallowin a Bingbot meta tag orX-Robots-Tagheader keeps a page out of Copilot answers. Until 1 Mar 2027 Bing also readsnoarchivethat way, andnocachelimits Copilot to the URL, title and snippet. Rubric flags any of these. - Is my firewall blocking Googlebot or AI crawlers?It might be. Rubric fetches your pages as a search or AI bot and as a browser. When the bot gets a block, a challenge page or a different page from the one people see, Rubric flags it.
- Does noarchive stop Copilot using my page?Yes, for now. Bing's Webmaster Guidelines say NOARCHIVE keeps a page out of Copilot answers, and NOCACHE lets Copilot use only its URL, title and snippet. Rubric flags pages that send either one.
Directives
- What does noindex mean?Noindex is a rule, in a robots meta tag or an
X-Robots-Tagheader, that tells search engines not to show a page in their results. Rubric flags pages you've asked it to score that carry noindex. - Does nosnippet stop AI Overviews using my page?Yes. Google names nosnippet, data-nosnippet and max-snippet as the controls that limit what it shows from your page, in results and in AI Overviews and AI Mode. Rubric flags pages where these rules stop AI answers quoting the main content.
- Should you use noindex and a canonical tag together?No. Noindex says "leave this page out", and a canonical to another page says "this is a copy, use that one". Google calls the pair a mixed signal, so use one or the other.
- What happens if a page has two robots meta tags?Google follows the most restrictive one. If a page says
indexin one place andnoindexin another, it can drop out of search even though one of its tags says it shouldn't. - What does a nofollow meta tag do?A nofollow robots meta tag tells search engines not to follow any of the links on that page. Pages that are only linked from there are found and crawled less often.
- What do noarchive and noimageindex do?These are less common robots rules.
noimageindexkeeps a page's images out of Google Images,unavailable_afterasks Google to drop the page after a date, andnoarchivestops cached copies. Rubric lists them so you can check each one is deliberate. - How do I find which pages are set to noindex?Rubric lists every crawled page that carries noindex, a robots rule in a meta tag or an
X-Robots-Tagheader that tells search engines not to show the page in their results, so you can check each one is deliberate. This list isn't scored. - Does nosnippet stop AI Overviews using your page?Yes.
nosnippetstops Google quoting any text from the page, andmax-snippetcaps how much it can quote. Google applies the same limits to AI Overviews and AI Mode. Rubric lists these pages here, and scores them once, through its snippet gate.
Robots.txt
- How do I allow AI bots in robots.txt?Don't disallow the bots that fetch pages for AI answers and search: OAI-SearchBot (ChatGPT search), Claude-SearchBot and Claude-User (Claude), PerplexityBot, Googlebot and Bingbot. Rubric flags your robots.txt when it blocks any of them.
- Is there a size limit for robots.txt?Yes. Google reads only the first 500 KiB of a robots.txt file and ignores everything after that. If yours is bigger, the rules near the end don't apply to Googlebot at all.
- Does Google respect crawl-delay in robots.txt?No. Google ignores
Crawl-delayin robots.txt. Bing obeys it, so Bingbot fetches your pages no faster than the delay allows. Rubric flags it as a note when the rule applies to Googlebot, Bingbot or all bots.
Business facts
- Why do AI shopping answers show the wrong price?Often because your own site gives two prices. Rubric flags a product (same SKU or same name) whose structured data shows one price on one page and a different price on another.
Canonicals
- Can a canonical point to a noindex page?It shouldn't. Rubric flags pages whose canonical names another page as the main version when that page is set to noindex. The two signals cancel out, and neither page may end up in search.
- Should my website use www or non-www?Either works for search. What matters is that only one loads: Rubric flags a site when both
www.example.comandexample.comserve the full page and neither redirects to the other. - What happens if a page has two canonical tags?Google may ignore all of them. Rubric flags a page that declares more than one different canonical: two
rel="canonical"tags, an HTML tag that disagrees with the HTTP header, or a canonical that JavaScript changes. - What is a canonical chain?A canonical chain is when page A names page B as its canonical, but page B names page C. Search engines may not follow the chain, so they can pick a different page from the one you meant.
- Can a canonical tag point to another domain?Yes, but it hands the page to the other domain. A canonical on another host tells search engines to index that site's copy and not yours, so your page may drop out of results.
- Should a canonical URL be absolute or relative?A canonical URL should be absolute, such as
https://example.com/shoes/. Google accepts a relative one like/shoes/, but recommends full URLs, because a relative link is resolved against whatever address the page was fetched from.
HTML validity
- Why is Google ignoring my meta tags?Often because something in the
<head>broke it. An element that isn't allowed there, such as an<img>,<iframe>or<div>, ends the head early. Rubric flags it when your title, canonical, hreflang or robots tags come after that break. - Can a page have two head tags?No. A page should have exactly one
<head>and one<body>. When there are two, browsers and crawlers merge them, and tags from the second head often end up in the body, where search engines ignore them. - What happens if you don't close a title tag?Everything after it is read as part of the title. An unclosed
<title>,<script>,<style>,<textarea>or comment swallows the rest of the page, so crawlers can't see the content that follows. - What is Googlebot's HTML size limit?Googlebot reads the first 2 MB of an HTML file, measured uncompressed. Google indexes only what it downloaded, so text, links and markup after that point are invisible to Google.
- Where should the Google Tag Manager noscript code go?Put it just after the opening
<body>tag, not in the head. A<noscript>in the head may only holdlink,metaandstyle. When it holds animgoriframe, as Tag Manager and Facebook pixel fallbacks do, readers that don't run JavaScript end the head there. - Which HTML elements are allowed in the head?The head should hold only
title,meta,link,script,style,base,noscriptandtemplate. Your page has another element there. Nothing important comes after it yet, but Google ignores every head tag that follows an invalid one.
Internal links
- What is an orphan page in SEO?An orphan page is a page that no other page on your site links to. Rubric also flags pages that are linked but missing from your XML sitemap, because both routes help search engines and AI tools find a page.
- What is click depth in SEO?Click depth is the number of clicks it takes to reach a page from your home page. Rubric flags pages that are four or more clicks deep, because search engines crawl buried pages less often.
- Why does every page need internal links?Internal links let readers and crawlers move from one page to the rest of your site. Rubric flags a page with no internal links at all, counting menus and footers, because it's a dead end.
- Why do Shopify product URLs include /collections/?Shopify shows the same product at
/products/bootsand at/collections/sale/products/boots. Rubric flags stores that link to the collection version when that address doesn't name/products/bootsas its canonical. Search engines then see two copies of each product page. - Which of my pages have broken outbound links?Rubric's page-level summary lists each page with dead links to other sites (404, 410 or server errors). A few don't hurt rankings much, but they do let readers down. It's shown without scoring, because the same links are scored once in the broken outbound links check.
JavaScript rendering
- Can search engines see a title set with JavaScript?Google can, after it renders the page. Bing often doesn't, and most AI crawlers never do, so they see a page with no title. Rubric flags pages where the server HTML has no title and one appears only after JavaScript runs.
- Can AI crawlers read JavaScript content?Mostly, no. The crawlers behind ChatGPT, Claude and Perplexity read the HTML your server sends and don't run JavaScript, so text that only appears after scripts run is invisible to them. Rubric compares that raw HTML with the rendered page to find content that needs JavaScript.
- Should the canonical tag be added with JavaScript?No. Put the canonical in the HTML your server sends. If it only appears after JavaScript runs, crawlers that don't run scripts never see it, and Google may read it late or not at all.
- Can Google see an H1 added by JavaScript?Google usually can, once it renders the page. Most AI crawlers can't: they read the HTML your server sends, and an H1 added by script isn't there, so to them the page has no main heading.
- Can JavaScript change the canonical tag?It can, but it shouldn't. When the canonical in the server HTML differs from the one after JavaScript runs, search engines get two answers, and Google may pick either one.
- Does Google see titles changed by JavaScript?Google usually does, because it renders JavaScript. Crawlers that don't run scripts, including most AI crawlers, see only the title in the server HTML, so they may show a different or outdated title.
- Can crawlers read a JavaScript meta description?Google can, because it runs JavaScript before indexing. Many other crawlers, including most AI crawlers, can't. Rubric flags pages whose meta description only appears after scripts run, so the HTML your server sends has none.
Product data for AI shopping
- Why doesn't my schema price match the page?Usually because the markup and the page read the price from different places. Rubric flags product pages where the price in the structured data doesn't match any price shown on the page.
- How do I add a return policy to product schema?A missing return policy means neither your product markup nor your Organization markup says how returns work. Google and AI shopping tools can't show your returns window or whether returns are free.
- Why does my schema say in stock when it's sold out?A stock status mismatch is a product page whose structured data and visible page disagree. For example, the markup says
InStockwhile the buy button is disabled and says "Sold out". - Why does my product schema show a price of 0?A zero price in product data means a product you sell has Offer markup with a price of 0, or no price at all. Google's merchant listings need a price above zero, so the product can't appear there with its price.
- How do I add shipping details to product schema?Missing shipping details means neither your products nor your Organization markup give a shipping cost or delivery time. Google and AI shopping tools can't show what delivery costs or how long it takes.
- How do I add stock status to product schema?Missing stock status means your product pages have Product markup but no
availabilityvalue. Search engines and AI shopping tools can't tell from your data whether the product is in stock. - Do my products need a GTIN?Not always, but you should add one if the product has a barcode. Rubric flags products whose data has no GTIN, MPN or SKU and no brand. Identifiers let Google and AI shopping tools match your product to the same item sold elsewhere.
- How do I add return policy structured data?Add
hasMerchantReturnPolicyand shipping details to your Organization or Product structured data, or set the policies in Google Merchant Center or Search Console. Rubric flags shops where it finds no policy markup and no link to a policy page. - Does my online shop need a returns policy page?Yes. Publish clear shipping and returns pages and link them from the footer of every page. Rubric flags shops where none of the pages it crawled links to a shipping, delivery, returns or refunds page.
Security
- Why does my website say Not Secure?Because the page loads over plain
http://instead ofhttps://. Rubric flags pages that are served over http rather than redirecting to an https address. - What is an insecure form?An insecure form is a form whose
actionsends what people type to a plainhttp://address. Anyone on the network can read it, and browsers warn people before they submit. - What is mixed content?Mixed content is an https page that loads some of its files, such as scripts, images or frames, over plain http. Browsers block the risky ones and may drop the padlock, so parts of the page can break.
- What is a missing CSP header?A missing CSP header means your home page sends no Content-Security-Policy, the rule that tells browsers which scripts, styles and frames the page may load. It is a security step, not a search factor, so Rubric counts it as minor. Response headers are set once for the whole site, in your server, host or CDN settings (on Cloudflare, a Transform Rule), so this is one site-wide fix, not a page-by-page one.
- What is an HSTS header and do you need one?An HSTS header (
Strict-Transport-Security) tells browsers to use https for your site every time, even if someone types or clicks anhttp://link. Without it, a visitor's first request can still go out over plain http. It's a security step, not a search factor. Response headers are set once for the whole site, in your server, host or CDN settings (on Cloudflare, a Transform Rule), so this is one site-wide fix, not a page-by-page one. - What is a missing nosniff header?A missing nosniff header means your home page doesn't send
X-Content-Type-Options: nosniff, which tells browsers not to guess what kind of file they're loading. It's a one-line security step, not a search factor, so Rubric counts it as minor. Response headers are set once for the whole site, in your server, host or CDN settings (on Cloudflare, a Transform Rule), so this is one site-wide fix, not a page-by-page one. - What is clickjacking protection?Clickjacking protection is a header that stops other sites showing your page inside a hidden frame to trick people into clicking. Rubric flags it when your home page sends neither
X-Frame-Optionsnor a CSPframe-ancestorsrule. It is a security step, not a search factor. Response headers are set once for the whole site, in your server, host or CDN settings (on Cloudflare, a Transform Rule), so this is one site-wide fix, not a page-by-page one.
Status codes and redirects
- What is a broken key page?A broken key page is one of your most important URLs (the home page, your most-linked pages or pages in your sitemap) that returns an error instead of loading. Search engines and AI tools can't read, rank or cite a page that doesn't load.
- Is a meta refresh redirect bad for SEO?It's a weak redirect. A meta refresh sends visitors on from inside the page instead of from the server. Google follows it, but it's slower for people and some crawlers and AI tools don't follow it at all.
- What is a redirect chain?A redirect chain is a URL that redirects to another URL that redirects again before it reaches the real page. Rubric flags chains, loops and redirects that end on an error on your key pages.
- What is an HTTP Refresh header redirect?A Refresh header redirect is a page that answers with a
RefreshHTTP header, which tells the browser to load another address. Google reads it, but most other crawlers and AI fetchers don't, so they stop at the old page.
Titles, descriptions and headings
- Are duplicate title tags bad for SEO?Yes, usually. A duplicate title tag means two or more indexable pages on your site use exactly the same
<title>, so nothing tells a search engine or AI tool which page answers which question. - What is an H1 tag?A missing H1 tag means the page has no main heading in its HTML. Readers, search engines and AI tools lose the clearest signal of what the page is about.
- What is a title tag?A title tag is the
<title>in a page's HTML: the name of the page that search results show as the blue link. Rubric flags a page that has no title tag, or an empty one. - Can a page have two title tags?It shouldn't. A page with two
<title>tags leaves search engines to pick one, and it may not be the one you wrote. The extra one is usually added by a theme or plugin. - How long should an H1 be?Short enough to name the topic at a glance. Google sets no length limit for an H1, but Rubric notes headings over 70 characters, because a long H1 usually mixes the topic with detail that belongs in the opening sentence.
- How long should a meta description be?A long meta description is a page summary of more than 160 characters. Search results usually cut it short, so the part people and AI tools see may stop mid-sentence.
- What is a meta description?A meta description is a short summary of a page in its HTML, which search engines may show under the title in results. Rubric checks each page has one and that it runs between 50 and 160 characters, long enough to say something and short enough not to be cut.
- What is more than one H1 on a page?"More than one H1 on this page" means the page shows more than one top-level heading. Google says that's fine, and in Rubric it never costs points. It's a tidy-up for accessibility and clarity: worth doing when the extra H1 is a logo or menu label rather than a real heading.
- Can a page have two meta descriptions?It shouldn't. A page needs one meta description tag. Rubric flags pages with more than one, which usually means a theme and an SEO plugin, or two plugins, are both adding a tag.
- What is a long title tag?A long title tag is a page title of more than 65 characters. Google sets no limit, but it usually cuts titles around 60 characters, so the end of a long title is often hidden.
Blocked resources
- Why are my CSS or JavaScript files returning 404?A missing resource is a style sheet, script or image the page asks for that answers "not found" or a server error. The page may then look broken or lose content, for visitors and for crawlers.
- Should you block CSS and JavaScript in robots.txt?No. If robots.txt blocks the CSS or JavaScript a page needs, Google can't render it properly and may see a broken or empty page. Rubric flags pages whose style sheets or scripts are blocked for Googlebot or Bingbot.
Content AI engines can quote
- How should you format a best-of list for AI search?Format it as a numbered list in rank order, with each item's name in a heading. Rubric flags pages with a "best" or "top N" title that present the items as prose instead of a ranked list.
- Where should the answer go on a page for AI search?The answer should go in the first third of the page. Rubric flags pages where the first real paragraph of answer text sits below background, history or author bios, where readers and AI tools may not reach it.
- Do comparison pages help you get cited by AI?Often, yes. Rubric flags shops and software sites that have no comparison or best-of page, such as "X vs Y" or "best X for Y", with the options laid out in a table or list. Those are the questions people ask before they buy.
Dates
- Why does Google show the wrong date for my page?Often because the date on the page and the date in your structured data disagree. Rubric flags pages where the visible date differs from
datePublishedordateModifiedin the markup. - Should a blog post show its publish date?Yes. An article should show when it was published (and when it was last updated, after a real change), and that date should match the one in its structured data. Rubric flags articles with no date, and pages where the date people see disagrees with the structured data.
- Can dateModified be before datePublished?No. A page can't be updated before it was published, so
dateModifiedshould be on or afterdatePublished. When it's earlier, the dates contradict each other and search engines can't trust either. - Why does my page show a date in the future?Usually a time zone setting, a scheduled post, or a template that prints the wrong date field. Rubric flags pages whose publish or update date is later than today, in the visible text or the structured data.
- What happens if datePublished is empty in schema?An empty date in structured data is a
datePublishedordateModifiedfield that is declared but left blank. Search engines treat it as an invalid date, so the page's markup tells them nothing about when it was written or updated.
Fan-out coverage
- What is query fan-out in AI search?Query fan-out is how AI search answers a question: it splits it into smaller sub-questions and searches for each one. Rubric flags questions, from your Search Console or your AI tool, that no page on your site answers.
hreflang
- Can hreflang point to a redirect or a 404?No. Each hreflang link must point at the live, final address of that language version. Google ignores alternates that redirect, return an error or are noindexed, so that version loses its place.
- What are valid hreflang codes?A valid hreflang code is an ISO 639-1 language code, with an optional ISO 3166-1 region after it, such as
en,en-GBorfr-CA, or the special valuex-default. Google ignores codes that don't follow this, such asen-UK. - Do you need hreflang x-default?It's optional, but useful.
hreflang="x-default"names the page to show visitors whose language or region isn't in your hreflang set, such as a language chooser or your main version. Without it, search engines pick a fallback for you. - Does hreflang need a self-referencing tag?Yes. Google says each language version must list itself as well as every other version. Rubric flags pages that list their alternates but leave themselves out.
- Can two pages use the same hreflang code?No. Each language or region code in a page's hreflang set should point to exactly one address. Rubric flags pages that list the same code, such as
en-gb, for more than one URL, so Google can't tell which version to show.
XML sitemap
- Does my website need an XML sitemap?Most sites should have one. Rubric flags a site when it finds no XML sitemap at
/sitemap.xmland none declared in robots.txt, or when the declared address doesn't return a sitemap. - Why does my sitemap list broken pages?Your sitemap lists addresses that return an error, such as a 404 or a 500. A sitemap should list only live pages you want indexed, so each broken entry is a wasted crawl.
- Can a sitemap lastmod date be in the future?No.
<lastmod>should say when a page last changed, so a date in the future is wrong by definition. Search engines may then ignore your sitemap's dates, and those dates are how they spot updated pages quickly. - Why does my sitemap list the wrong domain?Usually because the site's address setting still points at a staging, old or hosting domain. Rubric flags sitemaps that list addresses on a different host from your site. Search engines ignore those entries, so the sitemap does nothing for your real pages.
- Do you need an XML sitemap?Most sites should have one. An XML sitemap lists the pages you want search engines to find. Rubric found none in your robots.txt or at the usual addresses, so search engines have to find every page through links. This note isn't scored on its own, because the sitemap check covers it.
Accessibility
- Why do duplicate IDs break accessibility labels?A duplicate ID used by a label means two or more elements share an
id, and a<label for>,aria-labelledbyoraria-describedbypoints at it. Browsers use the first match, so a screen reader may read out the wrong label or description. - Does an iframe need a title attribute?Yes. An iframe needs a
titleattribute that says what it holds, such as "Video: how to fit a bike helmet". Without one, screen readers announce it as "frame" and users can't tell a video from a map or an ad. - How do you add alt text to an image button?Give every
<input type="image">button analtattribute that says what the button does, such as "Search" or "Subscribe". Without it, screen readers announce a nameless button, so some visitors can't tell what clicking it will do. - What is an unlabelled form field?An unlabelled form field is an input, such as an email box or a search box, with no label,
aria-labelor placeholder. Screen-reader users hear "edit text" and can't tell what to type, so they may give up on the form. - Should you disable zoom on mobile?No. A viewport tag with
user-scalable=noor amaximum-scalebelow 2 stops people pinching to zoom on a phone. People with low vision rely on zoom to read, so Rubric flags it.
Images
- What is missing image alt text?Missing image alt text means an image has no text description in its HTML. Screen-reader users can't tell what it shows, and search engines and AI tools have less to go on when they read the page.
Mobile usability
- How big should tap targets be on mobile?Make links and buttons at least 24 by 24 px, or leave enough space around smaller ones, and aim for 44 by 44 px on the main controls. Rubric flags pages where the lab test finds targets smaller than the WCAG 2.2 minimum without that space.
- What font size is best for mobile?Use a base font size of at least 16 px for body text. Rubric flags pages where, on a phone-sized screen, more than two fifths of the text is smaller than 12 px, which most people have to zoom to read.
New-page discovery
- How do I get new pages found by Google faster?List every new page in your XML sitemap and link to it from a page that gets crawled often, such as the home page or the blog index. Rubric flags sites whose newest pages are missing from the sitemap, or that no crawled page links to.
Outbound links
- Do broken outbound links hurt SEO?Not directly. Google doesn't list broken outbound links as a ranking signal. They do send readers to a dead end, and a page full of them looks unmaintained. Rubric flags links to other sites that return a 404, 410 or server error.
- Should you update outbound links that redirect?Yes, when you can. A link to another site that passes through a redirect still works, but the address is out of date. Some redirects now end on a home page or a different topic, so readers no longer reach the source you meant.
Page speed (lab)
- What is Total Blocking Time?Total Blocking Time (TBT) is how long scripts keep a page too busy to respond while it loads. Under 200 ms is good. Rubric flags pages over that in a lab test. TBT stands in for INP, the responsiveness measure Google takes from real visits.
- What is a good server response time?Under 0.8 seconds is good, by Google's guidance for time to first byte (TTFB). Server response time is how long your server takes to start sending the page after it's asked. Over 1.8 seconds is slow.
- What is a good CLS score?A good Cumulative Layout Shift (CLS) is 0.1 or less. CLS measures how much the page jumps around while it loads, for example when an image or banner pushes the text down. Rubric flags pages over 0.1 in a lab test.
- What is a good LCP score?A good Largest Contentful Paint (LCP) is 2.5 seconds or less. LCP is how long the biggest element in view, usually the hero image or headline, takes to appear. Rubric flags pages over 2.5 seconds in a lab test on a mid-range phone over slow 4G.
Pagination
- Should you noindex paginated pages?Usually not. Rubric flags lists where page 2 and later carry a noindex tag. Over time, search engines may stop following the links on noindexed pages too, so older products or posts listed only there get found less.
- Should paginated pages canonical to page 1?No. Each page in a paginated list (page 2, page 3 and so on) should name itself as canonical. Rubric flags lists where later pages point their canonical at page 1, which tells search engines to ignore them.
Site focus
- Do off-topic pages hurt SEO?A few don't. Rubric flags pages whose words have little in common with the rest of your site. One off-topic page is harmless, but many of them blur what the site is about, for search engines and for readers.
Structured data
- What format should prices and dates use in schema?Use plain numbers for prices (
19.99), three-letter currency codes (GBP), schema.org URLs for availability (https://schema.org/InStock) and ISO 8601 dates (2026-11-16). Rubric flags markup where values aren't in a format Google can use. - How do you get star ratings in Google search results?Star ratings in Google come from
RevieworAggregateRatingstructured data that matches real reviews shown on the page. Rubric looks for this markup, mostly on product pages, where it can still earn stars in search. - Which structured data properties are required?It depends on the type. Product markup for merchant listings, for example, needs a
name, animageandofferswith a price and currency. Rubric flags pages that mark up a type Google uses for a rich result but leave out a property Google lists as required. - Why is my JSON-LD structured data not working?Most likely a syntax error. A trailing comma, a stray or unescaped quote, or a line break inside a value makes the JSON invalid, and search engines then ignore the whole block. Rubric flags pages with a JSON-LD block that fails to parse.
- How do I check if my structured data is valid?Paste the page's URL into Google's Rich Results Test or the Schema Markup Validator. Rubric checks that every JSON-LD block on a page parses as valid JSON and names its
@type. A block that fails to parse is ignored entirely. - Does FAQ schema help with AI Overviews?There's no evidence it does. Google stopped showing FAQ rich results on 7 May 2026, and no Google document says FAQ markup helps AI Overviews. Rubric notes whether pages carry
FAQPageorHowTomarkup, but shows it without scoring it. - Is FAQ schema still worth using?Mostly not. Google stopped showing FAQ rich results on 7 May 2026, and HowTo rich results earlier.
FAQPageandHowTomarkup is harmless, but it no longer earns a feature in Google Search.
URLs
- Are double slashes in a URL bad for SEO?Yes, a little. A URL like
example.com/shop//bootsis a different address fromexample.com/shop/boots, even if both show the same page. That gives search engines two copies to choose from, and it's usually a sign of a template bug. - How long should a URL be for SEO?As short as it can be while still describing the page. Google sets no practical length limit, but Rubric notes addresses over 115 characters, because long URLs are hard to read, share and remember.
- Why does my URL repeat the same folder?Usually a broken relative link. An address like
/blog/blog/postappears when a link written asblog/post(with no leading slash) is followed from a page already inside/blog/. Each click can add another folder, creating endless addresses. - Can URLs have spaces or special characters?They can, but they shouldn't. Spaces, curly quotes and long dashes have to be encoded (a space becomes
%20), and those links often break when people copy or share them. Letters in your own language, such as é or ü, are fine. - Are URL parameters bad for SEO?Not on their own. Parameters such as
?sort=priceor?utm_source=are normal, but each one makes a new address. Rubric notes parameter addresses it crawled, because they can create many copies of the same page for search engines to sort through. - Should you use hyphens or underscores in URLs?Use hyphens. Google recommends hyphens to separate words in URLs, as in
blue-running-shoes, rather than underscores. Rubric notes addresses with underscores, but you don't need to change old ones just for this. - Are URLs case sensitive?Yes, the path part is. To Google,
example.com/Bootsandexample.com/bootsare different addresses. Rubric notes URLs with capitals, because if both versions open the same page, search engines see two copies.
A-Z of Rubric checks
B
C
H
I
M
- markup for retired rich results
- meta description that needs JavaScript
- meta refresh redirect
- missing CSP header
- missing CSS, JavaScript or image file
- missing H1 tag
- missing hreflang x-default
- missing HSTS header
- missing image alt text
- missing nosniff header
- missing or badly sized meta description
- missing or mismatched article date
- missing return policy in product data
- missing review or rating schema
- missing shipping details in product data
- missing stock status in product data
- missing title tag
- missing XML sitemap
- mixed content
- modified date before the published date
- more than one H1 on a page
- more than one meta description
N
P
- page AI answers can't quote
- page blocked or challenged for a search or AI bot
- page Copilot can't use in its answers
- page far from the site's main topics
- page four or more clicks from the home page
- page HTML over Googlebot's 2 MB limit
- page with broken outbound links
- page with conflicting robots directives
- page with more than one head or body
- page with more than one title tag
- page with no internal links
- pages served over http
- paginated pages that are noindexed
- paginated pages that canonical to page 1
- product links inside collection URLs
- product with different prices on different pages
- product with no identifiers
- publish or update date in the future
R
S
- site served on both www and non-www
- sitemap lastmod in the future
- sitemap that lists another host
- sitemap that lists broken pages
- slow main content on phones (LCP)
- slow server response
- snippet limit (nosnippet or max-snippet)
- space or special character in a URL
- stock status mismatch
- structured data missing required properties
- structured data price not shown on the page
- structured data that can't be read
- structured data that doesn't parse