Bulk Meta Description Checker

What checking descriptions in bulk finds that one URL cannot

Some description defects belong to one page and you can see them one page at a time. A description that is too long is too long on its own. A missing one is missing on its own. If that were the whole list, a bulk checker would just be a faster single checker.

It is not the whole list. The expensive problems are properties of a set, and they stay invisible until the descriptions sit beside each other.

The first three rows are the argument for pasting a list rather than checking a page, and they scale the same way: a duplicated description is rarely one duplicate, because whatever produced it is producing it everywhere that template runs.

The description is not a ranking factor, and that is the point

Google removed the meta description from its ranking signals a long time ago and has said so repeatedly since. That fact gets repeated as a reason not to bother, which gets the conclusion backwards.

• Ranking gets you into the list. The description is most of what decides whether anyone picks you out of it, which makes it a click-through rate lever rather than a ranking one.
• It is the only copy you write for the result page. The title tag is a label. The description is the sentence that argues for the click.
• Google rewrites most snippets anyway. It does that most often when the description is absent, boilerplate or off-topic, so writing a good one makes the rewrite less likely rather than impossible.
• A better CTR at the same position is free traffic. No content, no links, no crawl budget. Just the line that was already going to render.

What the tool reads, and where other checkers get it wrong

Reading a meta description looks like one regular expression, and the obvious one is wrong in ways that return a confident answer rather than an error. The first is not an edge case, and it is the reason two checkers disagree about the same page.

• Attribute order is not fixed. <meta content="…" name="description"> is legal and common, and a checker requiring name first reports a page with a perfectly good description as having none. This tool matches both directions.
• Entities are characters. Tom & Jerry renders as twelve characters and is measured as twelve, not the fourteen the raw markup holds.
• Templates and comments are not the page. A description inside a client-side template or a commented-out block is not what a crawler resolves, and counting it invents a duplicate that does not exist.
• Empty is not absent. An element present with nothing in it is reported as its own finding, because the fix is in whatever rendered the empty tag.
• A head can declare more than one. Google takes the first, so the count is reported rather than silently collapsed.

Duplicate descriptions across a set are also an early symptom of duplicate content, and when the duplicated pages target the same query they are worth checking for keyword cannibalization as well. A template writing one sentence for forty pages is often a template that should not have generated forty pages.

How to check meta descriptions in bulk

Seven steps, ordered so the findings that are never judgement calls get dealt with before the ones that are.

1. 01

   Bring a list that belongs together

   The findings that matter here are about the set, so paste URLs that share a template: every product in a category, every post in a series, every page one layout renders. A random sample cannot show you that a template is writing the same sentence everywhere.

2. 02

   Read the duplicates first

   One description repeated across a set is the most common finding on any templated site, and it is the one no single-URL checker can make. Google treats boilerplate as a reason to write its own snippet instead.

3. 03

   Separate the empty tags from the absent ones

   A page with no description element and a page with an empty one look identical in a search result and are fixed in different places. The empty one means something is rendering the tag and failing to fill it, so the fix is in the template.

4. 04

   Check the pixel column, not the character count

   Google gives a desktop snippet two lines of roughly 600 pixels each and cuts when the text fills them. The 155 character rule is a hedge that works until a description is full of wide characters, which is exactly when it matters.

5. 05

   Rewrite the ones that just repeat the title

   The result already shows the title on the line above. A description saying the same thing spends the only free copy you get saying it twice, and gives someone deciding whether to click nothing new to decide on.

6. 06

   Find the ones your CMS already cut

   A description ending in an ellipsis was truncated before Google saw it, which is what an excerpt filter does when it takes the first 155 characters of the body. That ellipsis ships to the search result and the snippet trails off mid-thought.

7. 07

   Order the rewrites by impressions, not by page count

   A weak description on a page nobody finds costs nothing. Take the CSV into the same sheet as a Search Console export and rewrite the rows that already have impressions, because those are the ones where a better snippet turns into clicks.

For the list itself, the sitemap URL extractor pulls every URL out of an XML sitemap and the website URL extractor crawls a site that does not publish one. Run the same list through the bulk title tag checker afterwards, and draft the replacements in the SERP snippet preview, which measures a single description the same way this page measures a hundred.

What this checker reports

Frequently asked questions