Find Out What's Holding Your Blog Back

Fail (0) if the page is not served over HTTPS. Otherwise start at full marks and deduct: 0.4 if http:// is served without redirecting to https://, 0.1 if that redirect is temporary (302/307), 0.1 if http:// cannot be reached at all, 0.4 for active mixed content (scripts, iframes or stylesheets over http://), 0.15 for passive mixed content (images or media over http://), 0.15 if our server waited more than 1.8 seconds for the first byte.

Not scored if robots.txt could not be fetched (timeout or connection error). A missing robots.txt (404) passes: crawlers then treat everything as allowed. Fail (0) if robots.txt returns a server error, blocks Googlebot from the entered page, or disallows the whole site. Otherwise deduct 0.3 if any sampled post is blocked, 0.3 if Bingbot is blocked site-wide, 0.3 if CSS/JS needed for rendering is blocked, 0.1 for unreadable lines, 0.1 if the file exceeds Google's 500 KiB limit.

We test every sitemap listed in robots.txt, any <link rel="sitemap">, and /sitemap.xml, /sitemap_index.xml, /wp-sitemap.xml, /sitemap-index.xml, /sitemap.php and /sitemap.txt. Not scored if none of those could be fetched. Fail (0) if none holds a valid sitemap. Otherwise deduct 0.2 if the sitemap is not referenced in robots.txt, 0.3 if robots.txt points to a sitemap URL that does not work, 0.4 if the first child sitemap is broken, 0.5 if it lists no URLs, 0.2 if it lists URLs on another domain.

Fail (0) if the entered page carries noindex (meta robots or X-Robots-Tag). Otherwise deduct 0.3 per sampled post with noindex, 0.2 for a nofollow directive on the entered page, 0.3 if a canonical points to another domain, 0.3 if the entered page canonicalises to a different URL, 0.2 for conflicting multiple canonicals, 0.2 for a canonical pointing to http:// on an https site, and 0.1 per page with no canonical (max 0.3).

We status-check up to 30 internal links found on the analysed pages (main-content links first). Links that time out or refuse automated requests are not counted. Not scored if no link could be verified. Deduct 0.4 for one broken link (4xx or 5xx), 0.6 for two, and all 5 points for three or more or for more than 10% of checked links. Deduct 0.2 if 20% or more of internal links point at a redirect, and 0.2 if the entered URL needs two or more redirects to load.

Each analysed page scores 1 for a title of 15 to 65 characters, 0.7 if longer than 65 (likely cut off in results), 0.5 if shorter than 15 or if it looks broken (raw shortcode, "Untitled", "Home" alone), and 0 if missing. The check score is the average, minus 0.2 if two pages share the same title and 0.1 if a page has more than one <title>.

Each page scores 1 for a description of 70 to 160 characters, 0.8 if longer than 160 (shortened in results), 0.6 if shorter than 70, and 0 if missing or if it contains raw shortcode or template code. The check score is the average, minus 0.2 if two pages share a description. Meta descriptions are not a ranking factor and Google may rewrite them; they affect how your result reads, and so whether people click.

Each page scores 1 with one non-empty H1, 0.8 with several H1s (valid HTML, but less clear), 0.3 if the H1 is empty, and 0 with no H1. Deduct 0.1 from a page that skips heading levels in its content (for example H2 straight to H4) and 0.3 from a post of more than 800 words with fewer than two H2s. The check score is the average.

Covers content images on every analysed page (logos, icons and tracking pixels in headers and footers are excluded). Not scored if there are no content images. Score = 1 minus (images with no alt attribute + half of images whose alt text is just a file name + a quarter of images with empty alt) divided by the number of content images. Empty alt is correct for purely decorative images, so it is only lightly weighted.

Half the points for internal links, half for structured data. Links: each post scores 1 with 3 or more internal links inside its main content, 0.5 with 1 or 2, 0 with none. Structured data: posts score 1 with Article/BlogPosting markup, the homepage scores 1 with WebSite, Organization or Person markup; deduct 0.25 of this half for each JSON-LD block that fails to parse (max 0.5). Structured data is an opportunity, not a ranking requirement: it helps search engines and AI systems understand the page.

Per post: 1 point, minus 0.25 if the average sentence is longer than 25 words, 0.25 if more than a quarter of paragraphs run over 150 words, and 0.2 if the Flesch reading ease is below 30 (English pages only). A post with under 150 words of readable main text scores 0.2. The check score is the average across posts. Readability formulas are rough guides, not rules.

Automated signals only. Per post: 0.4 if at least half the headline's topic words appear in the first 120 words; 0.3 if the opening paragraph is under 80 words and does not open with filler ("In today's...", "Have you ever...", "Welcome to..."); 0.3 if the post offers a summary, key takeaways, table of contents, or question-style subheadings. The check score is the average.

Per post: 0.5 for a visible or marked-up author (byline, rel="author", author meta or Article author markup); 0.25 for an author bio or profile link; 0.125 each for links to an About page and a Contact page anywhere on the page. The check score is the average. This measures whether credibility signals are present, not whether the author is a genuine expert.

Per post with at least 300 words: 1 point with one or more external reference links per 1,000 words, 0.6 with fewer, 0.3 with none, and 0 if the post states three or more statistics but links to no external source. Share buttons and social profile links are not counted. The check score is the average. We count links, not whether the sources are good or the claims are true.

Per post: 1 point, minus 0.5 if no publish or update date can be found in the page or its markup, 0.4 if the headline or title names a past year (for example "Best tools 2023" read in a later year), and 0.4 if the most recent date is more than two years old. The check score is the average. Older is not automatically worse: evergreen posts can be fine. Changing a date without changing the content does not make a post fresher.

Per post, using the raw HTML your server sends (many AI crawlers do not run JavaScript): 0 if under 100 words of main text are present, 0.5 if under 250 words, otherwise 1. Deduct 0.2 if the page has no <main> or <article> landmark. The check score is the average.

0.35 for Organization or Person markup naming the site owner; 0.2 if that markup links to official profiles (sameAs); 0.3 if posts mark up their author as a Person or Organization with a name; 0.15 for a declared site name (WebSite markup or og:site_name).

Reads robots.txt for the crawlers that fetch pages for search and AI answers. Deduct 0.5 each if Googlebot or Bingbot is blocked (Google AI Overviews and AI Mode use Googlebot; Bing's index powers Copilot and other AI services), and 0.25 each for OAI-SearchBot, PerplexityBot and Claude-SearchBot. Blocking AI training crawlers (GPTBot, ClaudeBot, Google-Extended, CCBot and similar) is a legitimate choice and costs nothing here. Not scored if robots.txt could not be fetched.

Per post. If it has question-style subheadings: 0.6 x the share whose first paragraph is a self-contained answer (20 to 90 words, not opening with "This", "It", "As mentioned" or similar), plus 0.2 for lists or tables, plus 0.2 for a summary or a definition near the top. Without question subheadings: 0.6 if the opening paragraph is 20 to 80 words and names the topic, plus the same 0.2 + 0.2. The check score is the average.

Uses source density, the measure from Blogging Titan's page-one blogging census: split the content into paragraphs, list items, table cells and headings, find every percentage and "Nx" multiplier, and count how many sit in a block that also links to another site. Score = 0.2 + 1.33 x source density, capped at 1 (so 60% or more sourced scores full marks). Not scored when the posts contain no such statistics. Original data you collected yourself has nothing to link to; that is fine, but say where it came from.

Fail (0) if the entered page has no viewport meta tag, or one without width=device-width. Otherwise deduct 0.3 if pinch-zoom is disabled (user-scalable=no or maximum-scale below 2), and 0.2 if the pages have three or more content images and none offer responsive sizes (srcset or <picture>).

From Google PageSpeed Insights (mobile). When Google has real-user data (Chrome UX Report) for the page, or failing that for the whole site, each of LCP, INP and CLS scores 1 if Good, 0.5 if Needs improvement and 0 if Poor, averaged. Without real-user data we fall back to the simulated Lighthouse lab score: 1 at 90+, 0.5 at 50 to 89, 0 below 50, and label the result as lab data. Not scored if PageSpeed Insights did not respond.

Covers content images on the analysed pages; not scored if there are none. Start at 1 and deduct 0.25 if there are three or more images and none are WebP or AVIF, 0.2 if more than half lack width and height (which causes layout shift), 0.15 for each weighed image over 300 KB (max 0.45, up to 8 images weighed), 0.15 if the first image (often the largest thing on screen) is lazy-loaded, and 0.1 if more than five images load and none are lazy-loaded.

Automated checks on the entered page: deduct 0.25 if <html> has no lang attribute, 0.2 if any links have no accessible text, 0.2 if any buttons have no accessible name, 0.2 if any form fields have no label, and 0.1 if there is no <main> landmark. Automated tests catch only some accessibility problems; colour contrast and keyboard use need a manual check.

Opportunities rather than faults. 0.35 if any analysed page offers an email sign-up; 0.35 if posts link onward to at least three other pages from their content (or show a related-posts block). If affiliate links are detected, 0.3 more is split between marking them rel="sponsored" (or nofollow) and showing an affiliate disclosure; if none are detected, the first two parts are scaled up to the full score.