How to Fix Duplicate Meta Descriptions

What a duplicate meta description is, and why it matters

A duplicate meta description is the same description text used on more than one URL. Audit tools flag it because the description is meant to summarise one page. When fifty pages all say Shop the best garden tools online, the text tells a searcher, and Google, nothing about any one of them.

Duplicates do not cause a penalty, and the meta description is not a ranking factor. The cost is indirect. Google is more likely to ignore a description that appears across many pages and build its own snippet from page text, so you lose control of what people read before they click. Large groups of duplicates are also a useful warning sign, because they often point at a deeper problem: a template printing a default, or many URLs carrying the same content.

How to find and group the duplicates

Search Console no longer reports duplicate descriptions, so you need a crawl. Our site crawler or any desktop crawler will list every URL with its description. Export the list, sort by description, and count how many URLs share each one. Then group them by template, because the fix is usually per template, not per page.

If you prefer the command line, this short Python script reads your sitemap, fetches each page and prints the descriptions used more than once, with a count:

import re, urllib.request
from collections import defaultdict

SITEMAP = 'https://www.example.com/sitemap.xml'
get = lambda u: urllib.request.urlopen(urllib.request.Request(u, headers={'User-Agent': 'desc-check'}), timeout=20).read().decode('utf-8', 'ignore')
urls = re.findall(r'<loc>(.*?)</loc>', get(SITEMAP))
seen = defaultdict(list)
for u in urls[:2000]:
    m = re.search(r'<meta[^>]+name=["\']description["\'][^>]*content=["\']([^"\']*)', get(u), re.I)
    seen[m.group(1).strip() if m else '(none)'].append(u)
for d, us in sorted(seen.items(), key=lambda x: -len(x[1])):
    if len(us) > 1:
        print(len(us), '|', d[:90], '|', us[0])

If your sitemap is an index of other sitemaps, run it against each child sitemap. Then look at the URL patterns in each group. Paths such as /tag/, /page/2/, ?sort= or /collections/ tell you which template is responsible.

Before fixing anything, remove from your list the URLs that do not matter: pages with a noindex tag, and pages whose canonical points elsewhere. Google does not show their snippets, so their descriptions are not worth your time.

Fixing the template default

The biggest groups almost always come from a default. Many themes print the site tagline, or one fixed sentence, whenever a page has no description of its own. Fixing that one line usually removes most of the duplicates at once. You have two good options: build a description from each page's own data, or print no tag at all when there is nothing specific to say.

WordPress with an SEO plugin

Set a template for each content type in the plugin settings, built from variables that differ per page. In Yoast, a product template might be %%title%% from %%sitename%%. %%excerpt%%; in Rank Math the variables use single percent signs, such as %title% and %excerpt%. Check that the theme is not adding its own description as well. Our guide to Yoast and Rank Math descriptions covers finding a hard-coded theme tag.

Shopify

Shopify themes print the description from the page_description object in theme.liquid. For products without an SEO description, Shopify uses the start of the product description, so duplicates usually mean the product descriptions themselves begin the same way. Collections and pages work the same way. A theme that hard-codes a fallback is the other common cause:

{%- if page_description -%}
  <meta name="description" content="{{ page_description | escape }}">
{%- endif -%}

If your theme has an else branch printing a fixed sentence, remove it. See fixing meta tags on Shopify for the rest of the head.

Next.js and other custom sites

Build the description from the page data in generateMetadata, and leave it out rather than falling back to a site-wide sentence:

export async function generateMetadata({ params }) {
  const { slug } = await params;
  const p = await getProduct(slug);
  const desc = p.summary
    || `${p.name} in ${p.material}, ${p.sizeRange}. From £${p.priceFrom}, delivered in ${p.leadTime}.`;
  return { title: `${p.name} | Oak & Iron`, description: desc };
}

Pagination, filters and true duplicates

Paginated pages

Page 2 and beyond of a category or blog listing can share page 1's description, or add the page number. People rarely search for page 3 of anything, so this is the lowest-priority group. The quick fix is a variable: Yoast offers %%page%% and Rank Math offers %page%, which add text such as Page 2 of 5. Each page in the series should keep a canonical pointing to itself, not to page 1.

Filter, sort and tracking parameters

Addresses such as /shoes/?colour=red&size=8 or /shoes/?sort=price usually show the same listing in a different order. Give them a canonical tag pointing to the main category, and do not link to them in ways that invite crawling. Once they are canonicalised, their descriptions stop mattering. Large filter systems can also waste crawling, which is covered in fixing crawler traps.

<link rel="canonical" href="https://www.example.com/shoes/">

Pages that are truly the same

If two URLs share a description because they share almost all their content, the description is a symptom. Merge them into one stronger page and add a 301 redirect from the old one. Common examples are two near-identical service pages for neighbouring towns, or an old and new version of the same guide. See fixing duplicate content and fixing canonical tags.

The same page at several addresses

If the crawl shows the same page under http and https, or with and without www, every description is duplicated twice over. The fix is a single 301 redirect to one version, not new descriptions. Check with curl -sI http://example.com/ that each variant redirects in one step to your main address.

Writing descriptions for the pages that matter

Templates handle the long tail. Your home page, main service or category pages, best-selling products and most visited posts deserve descriptions written by hand. Use your analytics or the Search Console Performance report to list the 20 to 50 pages that get the most impressions, and start there.

Each hand-written version names what the page offers, adds a specific fact such as a price, place or timescale, and is around 120 to 160 characters long.

A worked example

A garden tools shop on WooCommerce had a crawl report showing 904 URLs with the description Shop the best garden tools online. The groups were 640 products, 180 tag pages, 60 paginated category pages and 24 filtered URLs.

Viewing the source showed the theme printing the tagline whenever the SEO plugin had no description, because the plugin's product template had been cleared during a migration. The theme line was removed in a child theme, and the product template was set to build each description from the product name, its main feature from a custom field, and its price range. Tag pages were set to noindex, paginated pages gained the page number, and the filtered URLs already had canonicals to their categories.

The owner then wrote descriptions by hand for the 30 products and categories with the most impressions. A second crawl two days later found 14 duplicates left, all pairs of near-identical products, which were merged.

The fix, step by step

Edge cases worth knowing

• Language versions with the same description in English are duplicates too. Translate descriptions along with the page.
• Product variants on separate URLs, such as one per colour, share a description unless the template adds the variant name.
• Very short descriptions, such as a product code alone, can be unique and still useless.
• Staging sites that are crawlable duplicate everything. Block them with a login.
• Crawler limits may stop a crawl before it reaches deep pages, so the count can look smaller than it is.

Keeping descriptions unique

• Crawl the site monthly, or after any theme, plugin or template change, and compare the duplicate count with the last crawl.
• Do not fix duplicates by adding random words or numbers; it makes them unique but no more useful.
• Do not spend time on noindexed or canonicalised pages.
• Do not stuff descriptions with keywords to make them different from their neighbours.
• Add a description field to your content process, so new products and posts are published with one.

Google picks up changes as it recrawls, which takes weeks for a large site. Key pages can be sped up with URL Inspection and Request indexing. For how Google then chooses snippets, see why your meta description is not showing in Google.