Ecommerce Keywords That Produce Orders

A store owner once told me his best month for traffic was his worst month for revenue. That happens more often than you would think. Traffic is not the goal. Orders are the goal. Most ecommerce keyword research quietly optimises for the wrong one.

This guide gives you a repeatable process for picking keywords that produce orders. You can run it yourself. It takes a few hours the first time and about an hour a month after that.

Informational and buying intent are not the same customer

Every search sits somewhere on a line between "I am learning" and "I am buying". Same product, different person.

Informational queries:

• how to season a cast iron pan
• are merino wool socks worth it
• what size dog crate do I need

Buying queries:

• pre-seasoned cast iron skillet
• merino wool socks men's
• 42 inch dog crate

Look at what changes. Buying queries name a product, a spec, a size, a material, or a brand. Informational queries ask a question. The second type buys today. The first type might buy in three months, from anyone.

There is a middle group, often called commercial investigation:

• best cast iron skillet for gas hob
• merino vs alpaca socks
• large dog crate reviews

These people will buy soon but have not chosen. They are worth targeting, but they are worth targeting with a comparison page, not a product page.

Here is the honest part. Informational content is not worthless. It builds the topical depth that helps your commercial pages rank, and it is what AI assistants pull from when someone asks ChatGPT for a recommendation. But if your whole blog is informational and none of your category pages target buying terms, you will get traffic and no orders. That is the trap.

Category pages are your highest-value asset and your most neglected one

Ask most store owners which page they optimised last. They will say a blog post or a product page. Almost nobody says a collection page.

That is backwards. On Shopify a collection page, or a category page on any other platform, is the natural match for a buying query. Somebody searching "merino wool socks men's" wants a page with twenty pairs of merino socks on it, filterable by size, with prices visible. That is a collection page. It is not a single product, and it is not a blog post.

Category pages also carry the most commercial weight per visitor. A visitor landing on a collection sees your range, your prices and your delivery promise in one screen. They convert at a higher rate than a blog reader, by a wide margin, in almost every store I have looked at.

The neglect is usually structural. Most platforms generate collection pages automatically, with a title, a grid of products, and nothing else. No unique text, no buying guidance, no answers to the questions people ask before ordering. Google has very little to work with. Neither does an AI assistant summarising the category.

What a category page needs:

• A title tag using the actual buying phrase, not your internal category name
• 150 to 300 words of genuinely useful text explaining what is in the range and how to choose
• Answers to the two or three questions buyers always ask (sizing, care, delivery, returns)
• Internal links to the sub-categories below it and the parent above it
• Working filters that do not create thousands of crawlable duplicate URLs

That is a day of work across a typical store. It is the highest-return day in ecommerce SEO and most shops never do it.

Keyword cannibalisation, explained without the jargon

Cannibalisation means two or more of your pages compete for the same search. Google picks one, usually not the one you wanted, and both pages rank worse than one strong page would.

The classic ecommerce version looks like this:

• Collection page: /collections/merino-wool-socks
• Blog post: "The Complete Guide to Merino Wool Socks"
• Product page: "Merino Wool Socks - Charcoal"

All three target "merino wool socks". The blog post often wins, because it has more text. The customer lands on an article instead of a shop. Your conversion rate drops and you never notice why.

The fix is not to delete the blog post. The fix is to give each page a different job:

• Collection page targets the buying term: merino wool socks
• Blog post targets a question: how to wash merino wool socks
• Product page targets the specific item: charcoal merino wool socks

One target term per page. If two pages need the same term, merge them or redirect one. Then link the blog post to the collection page with the buying phrase as the anchor text. The article feeds the page that takes orders.

Mapping keywords to page types

Use this map every time you add a keyword. It removes most of the guesswork.

Query type	Example	Page type
Broad buying term	running shoes	Top-level category
Specific buying term	trail running shoes women's	Sub-category or collection
Product-level term	Salomon Speedcross 6	Product page
Comparison	best trail running shoes for wet ground	Buying guide or blog
Question	how long do trail shoes last	Blog post
Brand plus category	Salomon trail shoes	Brand collection page

If a keyword does not fit a page you already have, that is your content plan. Build the page.

Search volume alone is a trap

A term with 5,000 monthly searches looks better than one with 200. It usually is not.

Three reasons.

First, volume ignores intent. "Running shoes" has huge volume and mixed intent. Many of those searchers want Nike's website, not yours. "Wide fit trail running shoes women's" has small volume and one intent: somebody about to buy a specific thing.

Second, volume ignores your ability to rank. Difficulty is only meaningful relative to your own authority. A site with 13 referring domains and no ranked keywords cannot compete for a term where the top ten results all have thousands of links. Our own site is in exactly that position, which is why we target difficulty 0 to 9 terms and say so publicly rather than pretending otherwise.

Third, volume ignores value per visitor. Two hundred people searching for a $180 product are worth more than 5,000 people searching for a definition.

A practical rule: sort your keyword list by difficulty ascending, not volume descending. Then read the top twenty and delete anything with no buying intent. What is left is your target list for the next quarter.

To judge difficulty relative to yourself, do this manually. Search the term. Open the top five results. If they are all sites far larger than yours with deep, specific pages, that term is not for this year. If two or three are thin pages, forum threads, or general articles that do not really answer the query, you have an opening.

Finding the terms your competitors rank for and you do not

This is the fastest source of good keywords, because it removes the guessing. Somebody has already proven those terms convert in your market.

The process:

1. List five to eight direct competitors. Not Amazon. Not Etsy. Stores your size or one step above.
2. Put each domain into a keyword tool that shows ranked keywords. Ahrefs, Semrush, and DataForSEO all do this. Export the list.
3. Filter to positions 1 to 20 only. Below that the data is noise.
4. Remove branded terms. Nobody is searching your competitor's brand name and buying from you.
5. Cross-reference against your own ranked keywords. What they have and you do not is your gap list.
6. Apply the intent filter. Keep buying and comparison terms. Park the informational ones for later.
7. Apply the difficulty filter. Keep what your site can realistically reach this year.
8. Map each survivor to a page type using the table above.

You will finish with something between thirty and eighty terms. That is a year of work for most stores. Order them by how close they sit to an order, not by volume.

Run steps 2 through 8 again every quarter. Competitors publish. Rankings move. The gap list changes.

What this looks like as a monthly habit

• Week one: pull the competitor gap list, filter, map to pages
• Week two: fix the category pages for the top five buying terms
• Week three: publish the supporting articles and link them to those categories
• Week four: check for cannibalisation on anything new, and track movement

If you would rather not run it yourself, that is roughly what our ecommerce SEO service does each month: ten articles written and published, technical fixes, keyword tracking, and AI-search visibility tracking, at a flat $299 a month with the price on the pricing page rather than behind a sales call.

Should I target keywords with high search volume or low competition?

Low competition, if your site is small. A term you can actually reach in position three beats a high-volume term where you sit at position forty and get nothing. Check difficulty against your own authority, not against an abstract score. Move up to harder terms once you have rankings and links to support it.

How many keywords should one page target?

One primary term per page, plus close variations that mean the same thing. "Merino wool socks" and "wool merino socks" are the same page. "Merino wool socks" and "how to wash merino wool" are two pages. Splitting them prevents cannibalisation and makes each page clearer to both Google and AI assistants.

Do product pages or category pages matter more for SEO?

Category pages, in most stores. They target broader buying terms, they hold more internal link value, and they keep working when individual products sell out or get discontinued. Product pages still matter for brand and model searches, and for long-tail specs, but the category layer is where the neglected volume sits.