Ecommerce Meta Descriptions That Actually Get Clicked

You run a store with sixty categories and one template handling every meta description: "Buy {category} in our online store. Wide selection, fast delivery." Google will discard that description most of the time and write its own from the page content. Or it shows it and nobody clicks, because it says nothing that ten other stores don't say too.

This article is about writing ecommerce meta descriptions that have a chance of surviving and of lifting click-through rate. No theory you can read anywhere else.

What a meta description does and doesn't do

A meta description is not a ranking factor. It won't move you from page three to page one. It affects exactly one thing: whether the person who already sees your result clicks on it.

That is still worth something. If a category sits at position 6 with a 1.2% CTR, getting it to 2% means almost twice the visits without the position moving a millimetre. Across dozens of categories, that adds up.

The second thing to say up front: Google rewrites descriptions often. If it decides the page text answers a specific query better than your description does, it uses the page text. So treat your own description not as a guarantee but as a proposal that should be better than whatever Google would pick on its own.

Where stores get this wrong

When I crawl an online store, I almost always run into the same four problems.

1. The description is missing entirely. On product pages generated from a feed this is normal. Google then pulls the first sentence of the product copy, which is often a technical spec or a chunk of navigation.

2. One template for everything. A hundred categories, a hundred nearly identical descriptions differing only by the category name. For a search engine that isn't an error. For a human it's empty text. The same problem as duplicate title tags, one line lower.

3. Manufacturer copy pasted in. The same paragraph sits on ten other retailers' pages. It gives nobody a reason to pick you.

4. A description as long as a paragraph. 300 characters, of which the first third shows and the rest ends behind an ellipsis. Often including the part that mattered.

Length: what survives truncation

Google truncates by pixel width, not character count. In practice that works out to roughly 150 to 160 characters on desktop, considerably less on mobile.

Practical rule: put everything a person must see into the first ~120 characters. Treat the rest as a bonus that may or may not render. If your one differentiator — free shipping, store pickup, warranty length — sits at the end, nobody on mobile will ever see it.

Also keep in mind that Google sometimes prepends a date or splices in text from the page, which eats even more of your space.

What belongs in a category description

A category description has to answer one question: "will I find what I'm looking for here, and on what terms?" In practice that means four building blocks.

• What's here — specifically, not "wide selection". Product types, brands, sizes, or the number of items.
• Range or starting price — helps the visitor judge whether your offer is in their bracket. Prices change though, so be careful with numbers (see below).
• Purchase terms — shipping, availability, pickup, returns.
• A clear next step — "Compare by power output", "Filter by size". Not "Click here".

Before and after, for a cordless drill category:

Before: Cordless drills in our online store. Wide selection of quality products at great prices. Fast delivery. After: Cordless drills 12V to 20V — Makita, Bosch, DeWalt. Kits with two batteries or bare tools. Delivery in 2 days, 30-day returns.

The second one says roughly the same thing at a similar length, but you can actually learn something from it. Someone searching "18v cordless drill" sees both the voltage and the brands.

Product pages: don't just copy the first paragraph

Writing descriptions by hand makes no sense if you have thousands of products. A workable approach:

1. Build a template that pulls differentiating specs, not marketing sentences. So not "Buy {name} at a great price" but something like "{name} — {key spec}, {second spec}. {availability}, {shipping}."
2. Hand-write only the products that actually earn impressions. That is usually dozens of items, not thousands.
3. Check that the template doesn't produce text that gets cut off halfway through a spec.

If the specs are missing from the template, every product in the category ends up with the same description and the whole exercise loses its point.

Prices, stock and data that goes stale

Don't put a specific price in a meta description if it changes every week. Google reindexes descriptions with a delay, and a mismatch between the search result and the page is worse than no price at all.

Handle price, availability and ratings through structured data (Product, Offer, AggregateRating). That's where they stay current and where they can show up as a rich result. If your schema is broken, you lose them completely — and no meta description will make up for it.

Filters and pagination

Stores often generate descriptions for filter combinations and for page two, three, twenty of a listing. The result is hundreds of pages with near-identical text.

First decide which of those URLs you even want indexed:

• Filtered pages with real demand (for example "white kitchen sinks") deserve their own title and description.
• Combinations with no demand should be canonicalised to the main category or left out of the index. Writing descriptions for them is wasted work.
• Pagination — on page 2 and beyond, let the description inherit from page one, or don't bother at all. These pages aren't supposed to win clicks from search.

How to decide which descriptions to rewrite first

Don't work through them alphabetically. Open Search Console, go to the Performance report, switch to Pages and filter so you see URLs with high impressions and low CTR. Those are the pages people see in the results and don't click. That's where a new description has the biggest effect.

The second source is queries. If a category collects impressions on queries the description never mentions, you already have your brief — work them into the text naturally.

If you don't yet know how many pages have a missing, too-short or duplicated description, start with a crawl. The URL list is the foundation; without it you're tidying up blind.

Measuring: without a comparison it means nothing

Meta descriptions are one of the few changes whose effect can be measured fairly cleanly. The procedure:

1. Record the page's CTR and average position for the 28 days before the change.
2. Deploy the new description.
3. After 28 days, compare the same two numbers.

You're mainly looking at CTR at roughly the same position. If the position jumped significantly, the comparison is skewed and you need to wait longer. And expect that in some cases Google won't show your new description at all — then nothing changes, and that's a result too.

How I handle it

When I crawl a site I find the pages where the description is missing, shorter than it should be, longer than what gets displayed, or repeated across several URLs. For the findings that matter I write the finished description text — specific, at the right length, using the specs from that page. Not a recommendation saying "add a meta description", but text you can paste in.

Fixes get deployed through the WordPress plugin, the API, or a CSV export. You just approve them. For every deployed fix I then compare 28 days before and after, so you know whether the description did anything.

If you plan to handle SEO in-house, meta descriptions are one of the jobs that genuinely can be done that way. You just need to know which pages to take first.

Quick checklist

• The important information inside the first ~120 characters
• No two identical descriptions on two URLs
• Concrete specs, brands or types instead of "wide selection"
• No prices that change weekly — those belong in structured data
• Filter combinations without demand: don't index them, don't write descriptions for them
• Work order set by impressions and CTR from Search Console
• CTR measured 28 days before and after deployment