Do meta descriptions matter? The latest research

Meta descriptions do matter, although studies show that meta descriptions influence neither ranking position nor AI.

Writing meta descriptions can be a tedious task, especially on large websites with many missing meta description tags. Determining whether adding them is worth the time depends on your website, the individual page, and your priorities.

This article explores how and when meta descriptions actually matter, shares supporting data for each argument, outlines how to write them efficiently (i.e., should you use AI?), and how to discern if Google is rewriting your meta descriptions and what to do about it. 

Be the brand customers find first.

Track, grow, and measure your visibility across Google, AI search, social, local, and every channel that influences buying decisions.

Start your free trial

Why meta descriptions are less impactful today

Aside from their influence on organic rankings, meta descriptions are also becoming less impactful from a user and click-through rate (CTR) standpoint.

Why?

In today’s search results, organic real estate is increasingly taken up by paid ads and AI search results, such as AI Overviews. SERP features like star ratings, pros-and-cons snippets, and other rich elements are competing for the same space, and these rich snippets are often more visually appealing than the text snippet Google generates from meta descriptions.

Here’s an organic search listing with both a pros-and-cons snippet and review stars:

Ultimately, traditional organic listings are getting pushed further down the search engine results pages (SERPs). Opportunities for humans to actually see a meta description are shrinking, and that trend is likely to continue.

Do meta descriptions matter for AI search optimization?

Studies suggest that meta descriptions currently carry no weight with AI tools, but anecdotal evidence suggests otherwise.

An AI crawler study from Writesonic tested 62 hidden markers across six AI assistants, including ChatGPT, Claude, Gemini, DeepSeek, Grok, and Copilot. 

Meta descriptions scored zero across every crawler tested, as did JSON-LD (used for schema) and OG tags (used for social media previews). The only tag reliably read by these AI assistants was the title tag.

Writesonic is assertive in their finding that meta data doesn’t reach AI crawlers: 

“I need to say this clearly because a lot of current GEO advice gets it wrong: the metadata on your website does not reach AI assistants. Almost none of it.”

But importantly, they also state:

“…metadata is still valuable for traditional SEO.” 

We’ll cover the impact of meta descriptions for traditional SEO in the next section.

A test of how meta descriptions affect AI Overviews

In support of Writesonic’s findings, Seer Interactive ran a live experiment pointed specifically at AI Overviews and AI visibility. 

Over a six-week period, they replaced the meta descriptions on 21 high-traffic pages with either a single period or a neutral placeholder, then they compared crawl and citation activity against a control group of untouched pages. 

The result: A statistically flat -1.3% effect on AI crawl activity, with no citation penalty on the de-optimized pages.

Opinions and observations offer conflicting information

The studies and their findings are hard to disagree with, but some consultants have shared opposing views that — while anecdotal and over eight months old — should not be discounted. 

Why?

AI SEO is still very new and our understanding of its best practices will change. We don’t always know how every unique AI platform will process the contents of a page. Unlike traditional SEO, where there’s realistically just Google’s algorithm to please, AI SEO involves several different popular platforms, with possibly more to come. They’re all unique in how they analyze content on the web, and it’s not yet feasible to track the nuances of every single one.

Turning our attention to the most popular AI search platform, SEO consultant Chris Long shared a post highlighting ChatGPT’s source code and pointed to a snippet he suggests shows ChatGPT drawing from a page’s meta description, as well as other elements, such as the introduction, bullet points, subheadings, and more.

Within Long’s comment section, one reply from Maciej Chmurkowski stands out. Chmurkowski agrees with Long and supports his findings that meta descriptions matter:

“I can confirm the findings. I have analyzed 1,000 AIO answers and compared cited versus non-cited sources for the same keyword. Among the top 10 most correlated factors, meta description and title clearly lead the pack.”

While studies show meta descriptions don’t matter for AI search, some SEO consultants strongly disagree. Next, let’s look at meta descriptions and SEO.

Do meta descriptions matter for SEO? 

Google tells us that meta descriptions are used to create snippets under the page titles of listings within the SERPs, and Search Advocate John Mueller confirmed this in 2022. He said, “…the meta description is primarily used as a snippet in the search results page. And that’s not something that we would use for ranking.”

However, just because something isn’t a ranking factor, that doesn’t mean it isn’t worth doing. Mueller even said, “Having a good snippet on a search results page…can make it more interesting for people to actually visit your page.” 

So if meta descriptions are used to generate snippets, a more engaging snippet may lead to more clicks, which could then impact rankings. 

Let’s explore that.

Meta descriptions can increase click-through rate (CTR)

Your website listing’s snippet on Google’s SERPs is one of the first impressions your listing makes. It’s well known that Google uses meta descriptions to write snippets for pages on the SERPs. 

Writing a meta description gives you some control over what Google might pull. If content is available in the meta description, it might be used.

Bruce Clay, SEO expert and owner of BruceClay wrote: 

“[Meta descriptions serve] as the first impression of your website from the search results. You know what you want your search listing to look like. So it’s a good practice to have your page descriptions how you like them. Like, maybe, including a call to action?”

Adding a call to action is a good idea. It may encourage the user to do something. 

You could also add authority signals, such as: 

• Awards and recognition: “Award-winning,” “top-rated”
• Numbers and data: “Backed by $5M in research,” “10+ years of experience”
• Brand association: Mentions of press features, like “as seen in Forbes”
• Expertise: Signals that a professional was involved, like “doctor-reviewed” or “CPA-vetted”

Ultimately, a well-crafted meta description can influence someone to click your organic listing. Plus, higher CTR and dwell time can, over time, signal to search engines that a page is a trusted, high-value result.

---

Tip: Google has a developer resource on the Google Search Blog with best practices on how to write a meta description. 

---

Google rewrites meta descriptions more often than not

Here’s the catch: Google frequently ignores what you write within the meta description. Instead, it often generates its own snippet.

Seer Interactive’s analysis of thousands of tracked keywords found that 70% of tracked queries surfaced a Google-written description instead of the original, with individual pages showing two to 11 different snippet variations in a single month.

That doesn’t mean the meta description is pointless, however.

Google tries to provide the best possible snippet, whether generated by the meta description or otherwise, for the query. It even bolds terms so the snippet stands out, entices clicks, and thereby increases CTR.

Terms bolded by Google in meta descriptions tend to draw the eye and can lift CTR on their own. These bolded words closely align with and are relevant to the search query.

Here’s an example of a meta description with bolded terms that appears for the search “SEO versus PPC.” It’s clear how the bolded terms relate to the search query. Both “SEO” and “PPC” are bolded, and the meta description almost answers the search intent by clearly stating that the two are complementary strategies.

While Google has technically rewritten the meta description, the contents of this snippet are pulled verbatim from the sub-heading, an H2 with structured data establishing the content as an “alternative heading.”

Bruce Clay wrote: 

“[Meta descriptions help] search engines determine what the page is about. Even if Google doesn’t ‘use the description meta tag in our ranking’ and rewrites the description for the search results page, it still uses the original information to understand what the page is about. Put another way: The index still has the meta description tag — the search result text is just selected on the fly when the search result is rendered. To put this into perspective, Google could create many different descriptions for the same webpage depending on the query.”

Every time Google swaps in its own snippet, it’s telling you something about what it thinks searchers want to see for that query. 

If you pay attention to meta description variations over time, you’ll get a rough idea of the terms, phrases, and angles Google considers most relevant to your page. Feed those back into the page itself and you can effectively re-optimize it, making the page more likely to rank — and more importantly, more likely to engage audiences by speaking their language.

---

Pro tip: Skip to the section “How to see if Google is rewriting your meta descriptions,” to learn how to track meta description variations, then read the section “What to do when Google rewrites your meta descriptions” for a direct example of how updating pages based on meta description rewrites improves performance.

---

Our research on the impact of meta descriptions on SEO

To test the impact of meta descriptions on SEO, we removed all the meta descriptions from ten of the best performing articles on a blog. The meta descriptions were removed for a total of 15 days.

Here’s what happened:

Overall, the site received +18 clicks (+1.8%) and -6,000 impressions (-5.88%). Average CTR remained essentially the same, increasing by only 0.1%, and average position decreased by 0.1%.

Here’s the Google Search Console graph:

Realistically, all of these changes could be chalked up to natural fluctuations.

We also looked at the top pages in isolation — those whose meta descriptions were removed. Consistent with the sitewide results, changes were unexciting.

Clicks increased by +5 (+0.77%), impressions decreased by -6,000 (-4.6875%), average CTR was exactly the same, and average position decreased marginally, by just 0.1.

Here’s the Google Search Console graph:

For traditional SEO, it seems meta descriptions have no noticeable effect on rankings, but to restate their importance: Meta descriptions are used to write snippets on the SERPs, which can impact CTR, so you should still include them.

Is it better for humans or AI to write meta descriptions? 

Seer Interactive tested whether it’s best to write meta descriptions using humans or AI, comparing descriptions across hundreds of pages that were manually written, ChatGPT-4-generated, or left blank for Google to generate.

Here’s what happened:

• Rankings barely moved for any group, although the ChatGPT-4 pages saw the largest average ranking improvement, which, unfortunately, was a shift that didn’t translate into more clicks.
• CTR had the biggest gap. Manually written descriptions delivered a 176% CTR lift over ChatGPT-4. They even beat Google’s own rewrites by 54.1%. Conversely, the ChatGPT-4 group saw CTR drop 21.5%.
• Time and budget were clearly an advantage for AI. ChatGPT-4 completed descriptions more than six times faster than the manual process, spending about 2.5 minutes per page versus 19.1 minutes when written by a human.

Seer’s own takeaway wasn’t “never use AI to write meta descriptions,” however. It was that the performance gap, as measured, wasn’t yet close enough to justify skipping the manual step. 

Do you even need to write meta descriptions anymore?

Seer’s research found that junior specialists can spend double-digit hours each month on bulk metadata updates, which adds up to roughly the same budget as a full technical audit or a new piece of content. 

One of their ecommerce clients redirected 22 hours of meta description work in a single quarter toward product page content instead, and they saw a 13% lift in organic traffic to their category pages. 

As Seer put it, “time reallocated is opportunity captured.”

While updating meta descriptions is time consuming and sometimes wasteful, this isn’t always the case. 

Many websites have hundreds or thousands of pages that aren’t impactful in organic search, such as PPC landing pages or pages generated for sales teams. Pages with the least amount of importance in search can be left without a meta description or may receive an AI-written meta description. As Seers’ study found, writing with AI is faster. If adding a meta description is done quickly and without wasting human time, it could be completed when time and priorities allow. Low stakes and time saved is a perfect use case for AI.

The most important pages on a website, such as money-generating product or service pages, should be written by a human (or at least reviewed and edited by one). These meta descriptions may influence what Google displays on the SERPs, and as Seers’ research found, human-written meta descriptions do increase CTR compared to AI.

Brand-new pages that Google hasn’t yet crawled, fresh launches without established rankings, and legal or medical content with compliance requirements can all benefit from a manual meta description.

How to see if Google is rewriting your meta descriptions

Here’s how to quickly check whether your meta descriptions are showing up in search or getting rewritten by Google:

First, you need to track keywords in your Semrush account.

Go to the “Position Tracking” tool and create a new project. Enter the target keywords you want to check > Click the “Add keywords to campaign” button.

It may take anywhere from a few minutes to about an hour to add all the keywords if you’re doing this for the first time.

Next, you can view the SERPs for the keywords you’re tracking.

Navigate to “Overview” > Scroll down to the list of keywords > Look for the small SERP icon next to your ranking position > Click to pull up a live screenshot with a text breakdown of exactly how the result appears in Google.

From there, compare the description shown to the one you actually wrote for the page. If they don’t match, Google has rewritten the snippet to better fit the searcher’s intent.

What to do when Google rewrites your meta descriptions

Once you know where Google is rewriting your meta descriptions, you can use this information to reverse-engineer the algorithm. 

If Google keeps adding specific phrases you didn’t include, work those exact terms into your page’s headings and paragraphs. That’s Google telling you — in its own words — what it thinks the page should be about.

SEO consultant Steve Toth used a similar trick on a blog post ranking in position two for a target keyword. He opened Google Docs, typed a snippet-style sentence, and let Google’s predictive text autocomplete it. He then copied that text under the page’s H1. Within days, the page won the featured snippet. 

Toth’s take is that “Google’s predictive text mirrors how it understands content.” If Google autocompletes a phrase, that phrasing likely matches how Google wants the information displayed. The whole process took him under five minutes.

It’s not guaranteed to work on every page, but the principle is sound, and it costs nothing to test.

Know that meta descriptions matter

Don’t guess on whether Google is using the meta descriptions you wrote — check and confirm instead. 

Semrush’s Position Tracking tool shows you exactly which snippet is live for each keyword search, so you can stop rewriting descriptions Google’s already ignoring.

For new content, Semrush’s AI Article Generator can draft a solid first-pass meta description in seconds, freeing up your time for the pages that need a human touch. For everything already published, run the AI Search Optimizer across your existing pages to flag which descriptions are worth a rewrite and which ones can be left alone.

Own the conversation before your competitors.

See where your brand appears, where it doesn’t, and exactly how to win more visibility across search, AI, local, social, and every channel that matters.

Start your free trial