Every SEO check findable runs, and how to fix each one

By findable team. Published 2026-10-14. Updated 2026-10-14.

You ran an audit and got a list of ids like title_length and img_dimensions. This guide explains every one of them. Each check is a simple test on your HTML or response headers. Fix the failing ones and your score goes up.

What does the SEO family cover in findable?

SEO is 50% of your findable score, the biggest share. It has 98 checks covering metadata, structured data, performance and accessibility - the basics: metadata, page setup, headings, content size, links, images, social tags, security headers and markup. Each check has a fixed id, so you can search for it in any report.

How do I fix the title and description checks?

Four checks cover this: title_exists, title_length, desc_exists and desc_length. The title must exist and be 30 to 60 characters. The description must exist and be 120 to 160 characters. Search engines use both to build your result snippet, so write them for people.

<title>Every SEO Check ID and How to Fix It</title>
<meta name="description" content="A plain guide to the SEO checks, with the fix for each one.">

The length ranges are common guidance, not a hard rule from Google. Google may rewrite your snippet anyway. See the SEO starter guide for what it says about titles and snippets.

How do I fix the page setup checks?

Six checks cover basic page setup: viewport, charset, lang_attr, html_doctype, meta_refresh and robots_meta. Most are one-line additions to the head. The last two are about things you must not have: a meta refresh and a noindex directive. html_doctype ignores leading whitespace but fails if a byte order mark comes before the doctype.

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">

If robots_meta fails, look for noindex in <meta name="robots">. It tells crawlers to skip the page. If meta_refresh fails, replace the tag with a 301 redirect on the server.

How do I fix the heading and content checks?

Six checks look at structure and size: h1_exists, h2_exists, heading_hierarchy, word_count, text_html_ratio and dom_size. Use one h1, then h2 sections, then h3 inside them. Never jump from h2 to h4. Write at least 300 words of real text. A page with two h1 tags fails h1_exists just like a page with none.

Be honest about three of these. word_count, text_html_ratio (15%) and dom_size (1500 elements) are heuristics. Google does not publish these cutoffs. A short page can still be great. Use them as a smell test for thin or bloated pages.

How do I fix the link checks?

Four checks cover links: internal_links, external_links, empty_hrefs and link_noopener. You need at least 2 internal links and 1 external link. No link may have an empty href. Every target="_blank" link needs rel="noopener". A link to another host counts as external, so a link to your own subdomain is external too.

<a href="https://example.org/source" target="_blank" rel="noopener">Source</a>

The noopener value stops the new page from reaching your page through window.opener. rel="noreferrer" also passes the check, since it implies noopener. Read the MDN page on rel=noopener for details.

How do I fix the image checks?

Two checks cover images: img_alt and img_dimensions. Every image needs an alt attribute and explicit width and height. The alt check passes when the page has no images. The size attributes let the browser reserve space, so the layout does not jump.

<img src="/chart.png" alt="Bar chart of audit scores by family" width="800" height="450">

One catch: findable treats alt="" as missing, because the check trims whitespace and tests for empty. Decorative images with an empty alt are correct HTML, but they will still fail img_alt. Decide which matters more to you. Do not stuff keywords into alt text. Describe the image.

How do I fix the canonical and social tag checks?

Six checks cover sharing and duplicates: canonical, og_title, og_desc, og_image, og_url and twitter_card. The canonical link names the preferred URL. The Open Graph and Twitter tags control how your link looks when shared. See the Open Graph protocol. Each check only tests that the tag has a non-empty value, so a broken og:image URL still passes.

<link rel="canonical" href="https://example.com/page">
<meta property="og:title" content="Page title">
<meta property="og:description" content="Short summary.">
<meta property="og:image" content="https://example.com/card.png">
<meta property="og:url" content="https://example.com/page">
<meta name="twitter:card" content="summary_large_image">

How do I fix the security header checks?

Five checks look at transport and headers: https, hsts, x_content_type, x_frame_options and csp. They test your server response, not your HTML. A check passes when the header is present. It does not judge how strong the value is. If you audit a local file with findable ./page.html, there are no response headers, so expect these five to fail.

Strict-Transport-Security: max-age=31536000
X-Content-Type-Options: nosniff
X-Frame-Options: SAMEORIGIN
Content-Security-Policy: default-src 'self'

Set these in your web server or CDN. Start the CSP in report-only mode so you do not break scripts. The MDN CSP guide explains the options.

How do I fix the structured data and markup checks?

Three checks cover markup: jsonld, aria_landmarks and duplicate_ids. Add at least one JSON-LD block, use both <main> and <nav>, and keep every id unique. Duplicate ids break anchors and scripts, and they confuse screen readers. The jsonld check drops any block that is not valid JSON, so one stray trailing comma makes it fail.

<script type="application/ld+json">
{"@context":"https://schema.org","@type":"Article","headline":"Every SEO check"}
</script>

Read Google's structured data intro to pick a type that fits your page.

Key takeaways

• SEO is 50% of the findable score, so start here.
• There are 98 SEO checks. Each has a fixed id and a simple pass test.
• Fix the head first: title, description, charset, viewport, lang, doctype.
• word_count, text_html_ratio and dom_size are heuristics. Use judgment.
• Security header checks pass when the header exists. They do not rate its strength.
• Run findable -fail-under 80 URL in CI to catch regressions.