Title, meta description, canonical, viewport and Open Graph tags: present, sized right, unique across pages and describing what’s actually on the page.
Two lines decide whether anyone visits. The checker shows how yours render and where they get cut, so the part that matters is never the part that disappears.
Search engines cut titles by pixel width rather than character count, and may rewrite either line when the page offers a better match for the query.
Only three of the eight reach Google. The rest matter to a browser or to whoever pastes your link into a chat, and confusing the three groups is how an hour gets spent on a tag that was never going to move a ranking.
affects indexing and how results look
Kalenux <title>The clickable headlinemeta descriptionThe snippet below it, when usedmeta robotsindex / noindex, follow / nofollowaffects how the page renders, not how it ranks
meta viewportMobile layout, required for mobile-friendlycharsetHow the text is decodedaffects the card when someone shares the link
og:title · og:descriptionThe headline on shared linksog:imageThe picture on shared linkstwitter:cardLarge image or summary card on XWithout Open Graph tags, platforms guess, and they usually guess badly.
Google truncates at a rendered width, so the honest answer is a pixel figure and the character counts everyone quotes are approximations of it. These are the widths at which the cut usually falls.
The width is fixed by the result card, so the same title truncates at the same point regardless of how many characters it contains.
Mobile wraps to more, shorter lines rather than cutting at the same width, so a description that fits on desktop usually survives there too.
Fonts are proportional: a capital W is more than twice the width of a lower-case l. A title of capitals and wide letters runs out of room while a longer one in narrow letters still fits, which is why counting characters predicts the cut so poorly.
Google publishes no limit for either element, and the figures above move with the font, the device and whatever Google is testing this month. Front-load both lines and the cut-off stops mattering.
Google rewrites descriptions it finds unhelpful, and it does so for most pages [2]. These are the things that decide whether yours survives, none of which is a character count.
A description copied across a template is the commonest reason Google replaces it. If it would fit any page, it describes none of them.
✕ Acme is a leading supplier of
quality products since 1998.
✓ Waterproof hiking boots in sizes
36–48, shipped free, returns
within 60 days.
The snippet is read by someone deciding between ten results. It should confirm this page has what they searched for.
✕ Learn more about our range of
products and services.
✓ Compare 40 models by weight,
drop and terrain, with the
current price for each.
Truncation takes the end, never the beginning. Whatever makes this page worth clicking has to survive the cut.
✕ At Acme, we believe that every
customer deserves… [cut]
✓ Free returns within 60 days on
every order… [cut]
There is no character limit in the specification. Google truncates on pixel width, which varies by device, so a figure is a guide and never a target.
~155 chars → usually fits desktop
~120 chars → usually fits mobile
→ both are approximations
→ neither is a rule Google states
Google states there is no limit on how long a meta description can be, and that it rewrites snippets when a better match for the query exists on the page [2]. Writing to a number optimises the one thing that is not a rule.
What the search engines document about each check this tool runs, with every claim linked to its source.
Google generates the title link automatically, using the <title> element as a main source alongside headings and other content [1]. A title it judges unhelpful is replaced rather than truncated.
<title>Home</title>Too vague, site name only, or keyword stuffing. All three are commonly rewritten [1].
<title>Acme Ltd</title>
<title>Product | Acme | Shop | Buy</title>
<title>Waterproof Hiking Boots | Acme</title>Describes the page, puts the distinctive part first, and names the site once.
→ specific, then the brand
Google uses the meta description when it is a good match and generates one from the content when it is not [2]. This is not a penalty; the same page can show different snippets for different queries.
• the same text on many pagesEach gives Google a reason to prefer the page text [2].
• no relation to what was searched
• a list of keywords
• empty, or just the site name
length → no limit is specifiedThe thing that keeps a description is usefulness to the searcher, not any of these.
keywords→ not a ranking factor
exact match → not required
Google documents the meta tags it supports, including description, robots, viewport and a handful of others [3]. Tags outside that list are ignored by Search, whatever else they do.
<title>The short list that affects how the page is indexed or shown [3].
<meta name="description">
<meta name="robots">
<meta name="viewport">
<link rel="canonical">
<meta name="keywords"> → nobodyOpen Graph matters for shared links, not for ranking [6].
<meta name="author"> → not Search
<meta property="og:*"> → social
<meta name="twitter:*"> → X
Google confirmed in 2009 that it does not use the keywords meta tag in web ranking [5], and the position has not changed. It is neither helpful nor harmful; it is simply read by nothing.
<meta name="keywords"Usually emitted by a CMS field somebody is still filling in by hand [5].
content="boots, hiking, outdoor,
waterproof, shoes">
→ read by no search engine
the <title> → main sourceThree places the same minute of work actually changes something.
the description → the click
the page content → the ranking
Open Graph is not a Google ranking signal [6], but it is what every platform reads when a link is shared. Without it they guess, and they usually guess badly.
<meta property="og:title">Everything else is optional. These four build the card [6].
<meta property="og:description">
<meta property="og:image">
<meta property="og:url">
<meta property="og:image"Platforms fetch the image from their own servers, so a relative path resolves against the wrong host [6].
content="/images/hero.jpg">
→ relative: platforms fetch nothing
content="https://site.com/images/hero.jpg"
→ absolute: works
Google does not publish a character limit, because the cut-off is a pixel width that differs between desktop and mobile [2]. Character guides are approximations of a measurement nobody states.
WWWWWWWWWWWWWWWWWWWW → 20 wide charsA title of capitals and wide letters truncates far sooner than the same count in narrow ones.
illillillillillillil → 20 narrow chars
→ roughly double the pixels
title → distinctive part firstFront-loading makes the cut-off irrelevant, which is the only robust answer.
desc → the offer in the first clause
→ truncation then costs nothing
The robots meta tag and its X-Robots-Tag header equivalent control indexing, snippets and caching [4]. A stray noindex on a live template is the most expensive thing that can sit in a head.
noindex → remove from the indexSeveral can be combined on one tag, which is how they get set by accident [4].
nofollow → do not follow its links
max-snippet:0 → no description shown
noarchive → no cached copy
<meta name="robots" content="noindex">The checker reports this because nothing else in the head will tell you.
→ shipped from staging to production
Google treats a templated description as a weak signal and generates its own instead [2], and duplicate titles make results from one site indistinguishable from each other.
• one template, no per-page fieldAll structural, which is why they appear in hundreds at a time.
• paginated pages sharing page one
• filters producing near-identical pages
• a CMS default nobody overrode
/boots?page=2Pagination does not need clever copy; it needs to not be identical.
<title>Hiking Boots — page 2 | Acme</title>
→ distinct, and honest about what it is
Checking one URL tells you that page has a title and a description. It cannot tell you that 300 other pages share the same one, which is the failure that actually costs traffic, because it only exists in the comparison between pages. The full audit reads the head of every URL and reports the tags that repeat, the ones missing, and the ones Google is likely to rewrite.
Three hundred pages carrying one templated description is a single template to edit. Listed individually it reads as three hundred problems.
Pages with no description at all, no title, or a title that is only the site name, which is the version Google almost always replaces.
Truncation depends on the characters you used, not how many. Each title and description is measured at the width Google actually renders.