Site Audit Example: 67/100 with Errors and Warnings

The number is a vendor heuristic, not a Google metric. Google does not score your page out of 100 and no ranking system consumes this figure. It is a way of comparing a page against itself over time, and a poor way of comparing it against anything else.

This matters because of what the score does to your priorities. Ten issues on this page are worth different amounts, and the score treats them as arithmetic. A page can score 100 and rank for nothing. The checks below are all real, and passing every one of them removes the reasons a good page might fail — it does not make the page good.

Severity is assigned in the abstract. Impact is severity multiplied by what the page is for. This is a blog post, and that changes the ranking of these ten issues considerably:

1. Missing canonical. Genuinely first. Everything else is cosmetic by comparison.
2. 3 images with no alt text. Fast, and it is simultaneously an accessibility failure and an SEO one.
3. Title at 73 characters. Cheap to fix and it affects every impression the page ever gets.
4. og:image. Matters exactly as much as this post gets shared. On a blog, that is usually a lot.
5. Twitter Card tags. Mostly solved by fixing og:image — see below.
6. Meta description "too short". Near the bottom. Read on.
7. hreflang. Ignore it unless this site genuinely serves multiple languages.

The audit cannot know which of these is expensive for you. It flags a missing canonical and a short meta description with similar confidence. One of them can cost you the page; the other very likely costs nothing at all.

Without a canonical tag, Google does not refuse to index the page. It chooses a canonical itself — and it does so by weighing the URLs it has seen, the internal links, the sitemap and the redirects. Frequently it chooses correctly. Sometimes it chooses a URL with a tracking parameter attached, or the http version, or a paginated variant, and consolidates your signals onto that.

Three things worth knowing that the report does not say:

• A canonical is a hint, not a directive. Google can and does override it when the rest of your signals contradict it — internal links pointing elsewhere, a sitemap listing a different URL, a redirect. If your canonical is being ignored, the cause is nearly always that the rest of the page disagrees with it.
• Every page needs a self-referencing canonical, including this one. It is not only for duplicates. It is the page stating which URL it considers itself to be, and it is what protects you when someone links to you with a UTM parameter attached.
• It must be absolute, and it must match exactly — the same protocol, the same host, the same trailing-slash convention as the URL you actually want indexed. A canonical pointing at a URL that redirects is a contradiction you have written down.

You can see the outcome directly: the URL Inspection tool in Search Console shows "Google-selected canonical" alongside "User-declared canonical". When those two disagree, Google is telling you plainly that it did not believe you.

The report lists the missing og:image and the absent Twitter Card as two separate critical errors. In practice they are largely one problem, and understanding the fallback saves you half the work:

Twitter/X falls back to Open Graph. If twitter:title is absent it uses og:title; the same for description and image. The only tag with no Open Graph equivalent — and therefore the only one you genuinely must add — is twitter:card itself, which selects the card type. Set it to summary_large_image and the rest is inherited.

So the fix is: add og:image, add twitter:card, done. Two lines resolve two of the three critical errors.

"Meta description too short — 73 unused characters"

Unused characters are not a defect. The meta description is not a ranking factor, and Google rewrites it most of the time anyway, preferring a passage lifted from the page that matches the specific query. A concise, accurate 87-character description that Google actually uses is better than a padded 155-character one that it discards.

Fill the space if you have something to say. Do not pad it to hit a number — you are optimising a field that is not scored, in a slot that is frequently overridden. If you want to control your snippet, the reliable lever is elsewhere: state the answer plainly in the page's opening paragraph, in the words people search, so that the passage Google chooses to quote is one you wrote deliberately.

"Title tag too long — 73 chars"

Correct to flag, and the character count is a proxy. Google truncates on rendered pixel width, so a title of narrow letters survives longer than the same length in capitals. Two rules beat any character target: titles cut from the right, so put the distinguishing words on the left, and put the brand last. And note that an over-long title is one of the conditions under which Google discards your title entirely and writes its own — usually from your H1.

Every check here answers the question "can this page be crawled and understood?" None of them answers the two questions that decide whether the page succeeds:

• Is it indexed? Crawlable and indexed are different gates. Google can fetch a page perfectly, decide it is not worth keeping, and drop it. Only Search Console's Page Indexing report tells you this, and a clean audit sitting alongside a wall of "Crawled — currently not indexed" is diagnostic: the technical work is done and the problem is the content.
• Does anyone want it? No check has an opinion on whether the 2,847 words say anything that is not already said better elsewhere. Google does.

There is also a set of failures no single-URL audit can see at all: cannibalisation (four of your pages targeting one query, each passing every check individually), intent mismatch (a perfect page competing in a SERP made entirely of a different content type), and orphaning (this page might be linked from nowhere on the site, and a report on one URL cannot know). Fix the ten items above, then go and look at Search Console — that is where the answer lives.

What does the Site Audit check?

The aiwebpageseo Site Audit checks: title tag, meta description, viewport, canonical, Open Graph tags (og:title, og:description, og:image, og:image:width, og:image:height, og:image:alt), Twitter Card tags, H1 presence and uniqueness, heading hierarchy, image alt text coverage, schema markup types, HTTPS, TTFB, page size, render-blocking resources and mobile usability.

How is the site audit score calculated?

The score is calculated from the number and severity of issues found. Critical errors reduce the score significantly. Warnings reduce it slightly. Passed checks contribute positively. The maximum score is 100 — achieved when all 70+ checks pass without any critical issues or warnings.

How often should I run a site audit?

Run a full site audit after any significant content or technical change. For active sites, monthly audits catch regressions early. For high-traffic pages or e-commerce sites, weekly automated audits via the aiwebpageseo scheduled audit feature ensure issues are caught before they impact rankings.

Is the score a Google ranking factor?

No. It is a vendor heuristic — Google does not score pages out of 100 and no ranking system consumes this figure. It is useful for comparing a page against itself over time and poor for comparing it against anything else. A page can score 100 and rank for nothing: passing every check removes the reasons a good page might fail, but it cannot make the page good.

What happens if I have no canonical tag?

Google chooses one itself, weighing the URLs it has seen, your internal links, your sitemap and your redirects. Often it chooses correctly; sometimes it picks a URL with a tracking parameter, the http version, or a paginated variant, and consolidates your signals onto that. Every page should carry a self-referencing canonical — absolute, and matching the protocol, host and trailing-slash convention of the URL you want indexed. Check the result in Search Console's URL Inspection tool, which shows Google-selected canonical alongside user-declared canonical.

Do I need Twitter Card tags as well as Open Graph tags?

Barely. Twitter/X falls back to Open Graph — with no twitter:title it uses og:title, and the same for description and image. The only tag with no Open Graph equivalent, and therefore the only one you genuinely must add, is twitter:card itself, which selects the card type. Set it to summary_large_image, add og:image, and two of the three critical errors here are resolved by two lines.

Should I pad my meta description to 155 characters?

No. Unused characters are not a defect. The meta description is not a ranking factor, and Google rewrites it most of the time anyway, preferring a passage from the page that matches the specific query. A concise, accurate description that Google uses beats a padded one that it discards. If you want to influence the snippet, state the answer plainly in the page's opening paragraph in the words people actually search — then the passage Google lifts is one you wrote on purpose.