Missing meta description: how to find and fix it
The page has no meta description, so search engines assemble the result snippet from text on the page. It is not a ranking factor, but it is the one line of sales copy you control on the results page. Add a 70 to 160 character summary, starting with the pages that already get impressions.
Updated 28 September 2026
- Severity
- Warning
- Issue type
missing-meta-description- How the audit flags it
- An HTML page returning 2xx is flagged when it has no
<meta name="description">tag, or the tag's content is empty. Descriptions under 70 or over 160 characters are reported separately, as too short or too long. - How to check
- One URL at a time in the free SEO checker, or every page in a site audit.
What happens without one
Google writes the snippet itself, from whatever text on the page best matches the query. Sometimes that works well. Often it picks a cookie notice, a navigation label or a sentence cut off mid-thought, and the result looks worse than its neighbours.
Google says it uses the meta description when it describes the page better than the page text does. It may still rewrite it for some queries, so a description is not a guarantee. Without one, you have no say at all.
Meta descriptions do not affect ranking. What they affect is whether someone clicks, which is why the pages worth fixing first are the ones people already see in results.
Which pages to fix first
A site audit usually finds this on dozens of pages at once, because it is a template problem: the blog, product or docs template has no description field, or has one that nobody fills in.
- Start with pages that have impressions in Search Console. Sort by impressions and look at click-through rate; a page ranking on page one with a weak CTR is the clearest win.
- Then templates with many pages, where one field and a sensible fallback fix every page of that type.
- Leave pages that are noindex, or canonicalised elsewhere. They do not appear in results, so their descriptions are never shown.
How to fix it
Add a description field to the template, and write descriptions by hand for the pages that matter. A good one says what the page gives the reader, in plain words, in one or two sentences. It is not a list of keywords.
For the long tail, a fallback beats nothing: the page's excerpt or first paragraph, trimmed at a word boundary. Do not use one site-wide description as the fallback. Every page would then share it, and the audit would flag them as duplicates instead.
Aim for 70 to 160 characters. Shorter wastes the space results give you; longer gets truncated. The free meta tag checker previews how the title and description may look in a result for one URL.
Fix it with a coding agent
Connect Claude Code, Cursor or Codex to IndexZero's MCP server, open your site's repo, and paste:
Run a site audit of example.com and wait until it completes. Read the missing-meta-description issues and group the URLs by template. Find each template in this repo, add a description field with a fallback to the page's excerpt trimmed to 160 characters at a word boundary, and write descriptions by hand for the ten pages with the most words. Show me the diff before committing.
Behind that prompt the agent calls run_site_audit, polls get_audit_status until the crawl completes, and reads get_audit_issues with issueType: "missing-meta-description". Your repo is where the fix happens: IndexZero reads your site and never changes it.
Frequently asked questions
- Is a missing meta description bad for SEO?
- It does not lower rankings. It can lower clicks, because the snippet search engines write in its place is often worse than one you would write.
- Why does Google show different text from my meta description?
- Google picks the snippet per query. If it decides a passage from the page answers the query better than your description, it shows the passage. A specific, accurate description is used more often than a generic one.
- Should I generate meta descriptions automatically?
- For the long tail, a fallback from the page's own excerpt is fine. Write the important pages by hand, and review generated copy before it ships, because it is shown to searchers as if you wrote it.
Audit the whole site
Find every page with this issue
An agent connected to IndexZero's MCP server can crawl the whole site, read the issues of one type, find the templates in your repo and fix them, then re-crawl to confirm. The crawl itself spends no credits; only the optional Lighthouse sample does. The free plan includes the MCP server, audits of up to 50 pages, and 2,000 credits a month.
Related issues
- Duplicate title tags: how to find and fix themWhy pages sharing one title compete with each other, the templates that cause it, and when a canonical beats a rewrite.
- Thin content: how to find and fix itWhat a low word count does and does not tell you, the client-side rendering trap, and whether to expand, merge or noindex a thin page.
- SEO issues and how to fix themEvery check the site audit runs, grouped by severity.