A meta description is the summary under your page title in search results. It does not affect rankings and Google has rewritten most of them in public studies, so merchants gain most by hand writing them for their highest impression pages with ad tested selling points.
Google replaces the meta description you wrote roughly six to seven times out of ten in the largest public studies. Filling 400 empty fields is homework. Writing the 10 to 20 descriptions that sit on your highest impression pages is ad copy, and only one of those two jobs moves clicks.
Most Shopify merchants I talk to treat meta descriptions as homework. There is a field in the Shopify admin, an SEO app flags it as missing, and someone spends a weekend filling in 400 product pages. Then nothing measurable happens, and the conclusion becomes that SEO does not work.
The weekend was not the problem. The order of operations was. A meta description is one of the few SEO elements you can write in minutes and test against real data, but only on pages that already appear in search often enough for a better snippet to matter. Whether you’re doing $10K months or $1M months, the job is the same: find the pages people already see, and make the line under the title worth clicking.
A meta description is an HTML tag that summarizes a page, and its only job is to persuade a searcher to click; it is not a ranking factor. It sits in the page code as a meta tag named description, and Google may show it as the two lines of grey text under your blue title link. Your title tag tells Google what the page is about. Your meta description tells a human why this result is the one to open.
Jeff Oxford, founder and CEO of the ecommerce SEO agency 180 Marketing, put it plainly when he joined me on episode 409 of eCommerce Fastlane: “Meta descriptions is not a ranking factor. It’s, it’s more of like ad copy for SEO.” That framing changes how you write one. You stop stuffing keywords into it to please an algorithm and start writing it the way you would write a search ad, for a person deciding between ten options in about two seconds.
The keyword still matters, but as a click benefit: words that match the search help the result stand out. If you want the deeper picture of how titles, headings and content work together, the Shopify ecommerce SEO beginner’s guide covers the basics, and our guide to writing H1 tags for readability and SEO covers the heading that sits on the page itself.
In Shopify, you edit a meta description in the Search engine listing section at the bottom of a product, collection, page or blog post, and on the homepage under Online Store, then Preferences. Shopify’s help article on adding SEO keywords recommends about 160 characters for the description, allows up to 70 characters for the page title with 60 or fewer advised, and notes that longer text gets shortened in search results. Leave a product’s field blank and Shopify falls back to the opening of your product description; on other pages the theme fills the gap with page copy. That fallback is fine for most pages, which is the point of the next section.
Google replaces the meta description you wrote with its own snippet most of the time, roughly six to seven times out of ten in the two largest public studies. An Ahrefs study of 20,000 keywords, published in 2020, found Google rewrote descriptions 62.78% of the time on desktop. A Portent study of about 30,000 keywords from the same year found first page results had their description ignored 68% of the time on desktop and 71% on mobile.
Both studies are six years old, and nobody outside Google knows the current rate. What has not changed is Google’s own explanation. Its documentation on controlling snippets says snippets are built mainly from the page’s own content, tailored to each search, and that the meta description is used when it describes the page more accurately than a passage would. There is no maximum length; Google simply cuts the snippet to fit the screen.
The detail in the Portent data that matters most for merchants is where rewrites happen least. Higher volume keywords were less likely to have their descriptions rewritten (Ahrefs saw the same direction, though a smaller gap: 59.65% for head terms against 65.62% for the long tail), and longer, more specific queries were more likely to be rewritten. The author’s explanation is sensible: people put real effort into descriptions for head terms, while a long tail query is too specific for any one description to match, so Google pulls a better passage from the page.
Read together, the studies give you a priority rule rather than a reason to give up. A description written for the queries a page is already shown for has a fair chance of appearing. A description written for 400 pages that each get five impressions a month will mostly be replaced, and even when it is not, five impressions cannot produce a meaningful number of clicks.
Write meta descriptions by hand only for pages that already earn real search impressions, which for most Shopify stores means the homepage, a handful of main collections and the bestselling products, usually 10 to 20 pages in all. Everything else can use the theme fallback or a reviewed bulk draft until it earns its way onto the list.
Finding that list takes about 20 minutes in Google Search Console. Open the Performance report, set the date range to the last three months, switch to the Pages tab, and turn on impressions, clicks, CTR and average position. Sort by impressions. The pages at the top are where the snippet is seen often enough to matter. If you run a catalog of thousands of SKUs, check what share of total impressions your top 20 pages actually hold; when the long tail holds the majority, the reviewed bulk drafts described below matter more than they do for a smaller store.
Then look for two patterns. The first is a page ranking on page one, roughly positions one to ten, with a click through rate well below its neighbours. That page is being seen and passed over, and the title and description are the cheapest thing to change. The second is a page with a lot of impressions ranking just off page one. A better snippet will not fix that page on its own; it needs better content or more links, and Jeff’s advice on collection pages applies, a few hundred words of genuine buying guidance on the collection itself. Our Shopify collection page SEO guide walks through that work.
Collections usually deserve the attention first. They match the broader, higher volume searches like “linen bedding” or “trail running shoes”, which are exactly the head terms where Portent saw the fewest rewrites. Product pages earn hand written descriptions when they are bestsellers or when a single product is the answer to a common search. For the rest of your product pages, a clean, accurate first paragraph of product copy does the job, because that is what the theme and Google will fall back to anyway.
The fastest way to write a meta description that earns clicks is to reuse the selling points and calls to action that already won in your paid search ads, then fit them to what the searcher wants from that specific page. You do not need to guess what persuades your customers if your ad account has been testing it for you.
That was Jeff Oxford’s most practical tip on episode 409: “look at your top performing Google Ads on, on Google Search, see what unique selling points get the highest click through rates, what calls to action, the highest click through rates and incorporate those elements into your meta descriptions for SEO.” If your ads have shown that “free returns for 60 days” beats “premium quality” by a wide margin, that line belongs in the description of the collection that ad was pointing to. If you only run Google Shopping ads, you have less copy data to borrow, and your customer reviews and support tickets are the next best source of the words buyers use.
From there, a few rules hold up. Match the intent of the page: a collection description should signal range and choice, a product description should signal the specific reason to buy this item, and a homepage description should say who you are and what you sell in plain words. Put concrete facts in it, such as price range, shipping threshold, materials or guarantee, because Google’s documentation explicitly lists product details like price as useful. Keep the important part inside the first 150 to 160 characters so it survives truncation. Make every description unique, since identical descriptions across pages give Google a reason to replace them all.
Two things to stop doing. Stop writing keyword lists; Google calls them out as unhelpful, and shoppers skip them. And stop treating the description as a place for vague brand language. My own reaction on the episode still holds: getting this right is “not generically keyword stuffed but like that sounds like a whole science and form in itself.” For deeper guidance on the mechanics, Moz’s meta description guide is a solid reference.
AI is a reasonable way to draft meta descriptions for the long tail of a large catalog, as long as a person reviews the pages that matter and the drafts are specific to each page. Google’s documentation allows programmatically generated descriptions for large, database driven sites, on the condition that they read naturally, vary from page to page, and describe the page in front of the searcher.
The tooling has caught up. When Mike Simmons, CEO of Plug In Useful, joined me on episode 415, he described merchants copying keyword recommendations out of Plug in SEO into ChatGPT and pasting the results back, and explained that Plugin AI was built to “improve every single content field that’s there from the page title to the meta descriptions to the image alt text.” He was just as clear about the risk: “AI just left to its own devices, can turn out some pretty shoddy work.”
My own hesitation on that episode was about bulk changes: “I don’t know if a lot of people would do it in bulk. I think they’d be a little bit scared.” I still think that instinct is right for the pages on your Search Console list. Those 10 to 20 pages carry your organic revenue, and they deserve a human who knows the product, the margin and the customer. For the other few hundred pages, an AI draft that is accurate and unique is an improvement on a blank field or a duplicated sentence, and it costs very little. The same principle runs through our piece on AI product descriptions at scale: raw output is a first draft, and the review step is where the value is.
Whichever tool you use, including Shopify’s own AI features, check the drafts for invented claims. A model will happily promise free shipping you do not offer or a material the product does not contain, and a false promise in a search snippet costs more trust than a plain one.
Meta descriptions matter less on informational searches where Google shows an AI Overview, and they still matter on the product and category searches where most store revenue comes from. The decline in clicks on informational queries is real and large. In its February 2026 update, Ahrefs found that AI Overviews correlated with a 58% lower click through rate for the top ranking page, comparing December 2023 with December 2025 across 300,000 keywords. Pew Research tracked 900 US adults in March 2025 and found they clicked a traditional result on 8% of visits with an AI summary, against 15% without one.
No meta description fixes that. This is one reason the blog posts many merchants wrote years ago now earn impressions with almost no clicks.
Commercial searches behave differently. Jeff described a study of 300,000 keywords in which AI Overviews appeared almost entirely on informational searches and did not appear on searches like “best running shoes” or “blue running shoes”. That matches what most merchants see in their own data: collections and products still compete in a classic list of results, where the snippet is part of the pitch. Jeff also argued that click through rate feeds back into rankings: “if you get clicked more often than, than your expected click through rate, Google’s going to rank you up very fast.” That is his read of the evidence rather than something Google confirms, but it is one more reason to spend your time on the commercial pages.
Where AI assistants do send shoppers, they pull from your product data and from what other sites say about you, not from your meta tags. Our guide to ecommerce SEO in the AI era covers that side, and our AI visibility tools comparison lists ways to track how your brand appears in ChatGPT, Gemini and Perplexity.
On the labelled composite store below, a 140 product brand, the backlog shrinks from 210 descriptions to 16 hand written ones, with reviewed AI drafts covering the rest. The store is a composite, built from patterns I see repeatedly in merchant conversations rather than from any single store, so treat the numbers as illustrative: a home fragrance brand doing about $900K a year, with 140 products, 18 collections, a blog of 60 posts and Google Ads spend on branded and category search.
Their starting point is typical. An SEO app reports 210 missing meta descriptions, and the founder has asked a freelancer to write all of them. Search Console tells a different story: over three months, the homepage, six collections and nine products account for most of the store’s search impressions, and almost every blog post is shown on informational searches where an AI Overview now answers the question.
Here is the call I would make. First, write 16 descriptions by hand: the homepage, the six collections and the nine products, each built from the two ad lines that have the best click through rate in that category, with price range and the free shipping threshold included. Second, skip the blog posts entirely for now; a better snippet cannot beat an AI Overview, and the time is better spent improving the two or three posts that still convert. Third, let a reviewed AI draft fill the remaining product pages over the following month, checked only for accuracy. Fourth, cancel the freelancer brief for 210 hand written descriptions, because most of that work would be rewritten by Google and seen by almost nobody.
Then measure. Annotate the date of the change, wait four to six weeks, and compare clicks and CTR on those 16 pages against the previous period. If CTR rises on pages whose average position stayed flat, the snippet did the work. If nothing moves, the problem is the page or its ranking, not its description, and that is useful to know too.
The call changes by stage. A store under $50K a year should write the homepage and its top three collections and stop there. A store above $5M with a large catalog should treat descriptions as an ongoing test, rotating new ad winning lines into its top 50 pages each quarter. If you would rather hand the whole program to specialists, an ecommerce SEO agency such as ClickSlice can run it, and Shopify’s guide to increasing website traffic covers the channels beyond search.
A meta description on a Shopify store is the short summary of a product, collection, page or blog post that can appear under its title in Google search results. You set it in the Search engine listing section at the bottom of each item in the Shopify admin, and for the homepage under Online Store, then Preferences. Shopify recommends about 160 characters. If you leave the field empty, your theme usually uses the opening text of the page instead. The description does not affect rankings directly. Its job is to persuade a searcher that your result is the one worth opening, so it should state what the page offers and why it beats the other results.
Google uses the meta description you write only some of the time and replaces it with its own snippet more often than not. Two 2020 studies found rewrite rates of about 63% (Ahrefs, across 20,000 keywords) and 68 to 71% (Portent, for first page results). Google says it prefers your description when it describes the page more accurately than a passage from the page itself. Descriptions written for the main queries a page already ranks for are the most likely to be shown, while descriptions on pages found mostly through long, specific searches are usually replaced. That is why it pays to focus on your highest impression pages instead of every page.
A meta description should usually run about 150 to 160 characters, with the most important information in the first part so it survives truncation. Google sets no maximum length and simply cuts the snippet to fit the device, which on mobile can be shorter than on desktop. Shopify’s admin recommends 160 characters for the description field. Length matters less than content: the Ahrefs study found Google rewrote overly long descriptions at about the same rate as the rest. Lead with the reason to click, such as a price range, shipping offer or specific benefit, and avoid trailing keyword lists that add length without adding persuasion.
You do not need to hand write a meta description for every product page, and for most stores it is not a good use of time. Write them by hand for the homepage, your main collections and the products that earn the most impressions in Google Search Console. For the rest, a clear first paragraph of product copy or a reviewed AI generated draft is enough, because those pages are rarely seen in search and Google often rewrites their snippets anyway. Google’s own guidance allows generated descriptions on large catalogs as long as each one is accurate, unique and readable. Add pages to the hand written list as they start earning impressions.
You can use ChatGPT, Shopify’s AI features or an SEO app to draft meta descriptions, provided a person checks the important pages and every draft is accurate and specific to its page. Google allows programmatically generated descriptions on large sites if they read naturally and vary from page to page. The risk is invented detail: a model may promise free shipping, a guarantee or a material you do not actually offer. Use AI to fill the long tail of your catalog, and hand write the 10 to 20 pages that carry your organic revenue, using selling points that have already performed in your Google Ads or customer reviews.
