E-commerce Keyword Research: Matching Every Shopper's Search to One Page in Your Catalog

Two kinds of shopping search, and the page each one wants

Someone who types "espresso machines" is browsing. Someone who types "espresso machine with grinder" has narrowed the field. Someone who types a maker's model number has their card half out of the wallet. All three are shopping, but each expects a different page, and a store that sends all three to the same product page will rank for none of them.

The quickest way to learn which page a search wants is to run it. Search the broad term in a private window and look at what ranks: usually category pages, comparison articles and big retailers' collections. Search a model number and the results fill with product pages. The results page is a free statement of what Google thinks answers the search, and arguing with it is expensive. Our page on keyword intent goes deeper into reading it.

Category searches carry most of the reach; product searches carry most of the immediate orders. A store needs both, on different pages, which is the whole argument of the e-commerce SEO hub in one sentence.

Modifiers: the small words that tell you which shelf

Shoppers rarely type a bare product type for long. They add words that narrow it, and each family of modifiers points at a different kind of page. Build a modifier sheet for every major category before choosing a single target.

• Attributes such as color, size, material and finish (black leather, queen size, stainless). These usually belong in filters and product titles, and earn an indexable page only when the combination has demand and enough products to fill it.
• Audience such as women's, kids' or for large dogs. Often a category in its own right, with its own copy.
• Use or problem such as for flat feet or for small kitchens. Strong material for subcategories and buying guides.
• Price such as under $100 or budget. Honest only when the page genuinely sorts by that range.
• Comparison such as best, versus, review or alternative. These want guides and comparisons, not product pages.
• Brand and model, including part numbers and SKUs. Product pages, and brand pages when you stock enough of one maker.
• Purchase such as buy, sale, in stock or free shipping. Readiness signals; use them in titles only when they are true.
• Place such as near me or a city name. Relevant only when there is a counter, a showroom or local pickup.

The sheet does two jobs. It shows which modifiers repeat across categories, so you can design filters around them, and it shows which ones deserve pages. A modifier that appears in dozens of phrasings with clear demand is a subcategory waiting to be built. One that appears once is a line in a product description.

Mining Google autocomplete and Amazon's search box

Google says its autocomplete predictions reflect real searches, drawn from common queries that match what someone starts to type, and shaped by language, location, trending interest and the searcher's own past searches. It also says autocomplete does not simply show the most common queries and should not be compared with Google Trends. Treat predictions as a phrasebook of how real people word things, not a ranking by volume.

1. 1

   Search signed out, in a private window

   Your own history shapes the predictions you see. A clean window gets you closer to what a stranger sees.

2. 2

   Type the category, then walk the alphabet

   Try dog bed a, dog bed b and onward, then dog bed for, dog bed with and dog bed vs. Record every prediction word for word.

3. 3

   Repeat it in Amazon's search box

   Amazon's box shows how people word a product inside a store, where every search is a shopping search. It shows no volumes, so you collect phrasings, not numbers.

4. 4

   Read the results page, not just the box

   Note the related searches at the bottom, the questions Google lists, and which page types rank for each phrasing.

5. 5

   Tag each phrase with its source and intent

   A spreadsheet column for where it came from and one for category, product or question saves an hour of sorting later.

The best keyword data is already in your own accounts

Third-party tools estimate. A store owns four sources that do not, and most never open them for keyword work.

Google's title guidance for Merchant Center allows up to 150 characters, says to put the most important details first, and asks for terms customers actually search, such as the product name, brand and defining details, without prices or promotional text. Volumes, when you want them, come from Google's own Keyword Planner, read as rounded estimates rather than counts, and the tradeoffs among paid tools are in keyword research tools.

Which searches have earned a page of their own

A store can generate a page for every phrase on the list, and that is precisely how stores end up with thousands of thin, near-identical pages Google ignores. Before a search gets a page, put it through five questions. This is our working practice, not a Google rule.

1. 1.Demand: does anyone search it, by any evidence: autocomplete, Search Console impressions, site search logs or a tool's range?
2. 2.Inventory: do you stock enough products to fill the page? A category of two products is a product page wearing a costume.
3. 3.Distinct results: does Google show different pages for it than for the parent term? If the results are the same, one page should cover both.
4. 4.Margin: do the products on that page earn enough to justify the copy, photos and upkeep?
5. 5.Odds: who ranks now? A page of marketplaces and national chains is a long campaign; a page of small specialists is a fair fight.

Mapping keywords to the catalog, one page per search

The map is the deliverable. A list of keywords is trivia; a list of keywords with an owner page each is a plan someone can carry out.

1. 1

   Export the catalog and the category tree

   Every category, subcategory and product, with its URL, title and product count.

2. 2

   Group phrasings into clusters

   Phrasings that mean the same thing and return the same results belong together. Keyword clustering describes the method.

3. 3

   Give each cluster one owner page

   A category, a subcategory, a product or a guide, never two. Synonyms share a page.

4. 4

   Mark the gaps

   Clusters with demand and inventory but no page become new collections or guides. Clusters with demand and no inventory are a buying decision, not an SEO task.

5. 5

   Find the collisions

   Two pages aimed at one cluster split the signals. Merge, redirect or re-aim one; see keyword cannibalization.

6. 6

   Hand the map to whoever writes

   Titles, collection copy, product names and feed titles all come from the same sheet.

Where the mapped words go on each page

• Category pages: the primary phrase in the title tag and the H1, a short introduction written for a shopper, and filter labels in customers' words rather than warehouse codes.
• Product pages: a product name built from what people search, typically brand, product type, model and the defining attribute, not an internal code. The feed title follows the same order.
• URLs: Google recommends descriptive words in URL paths over bare numeric IDs. On an established store, changing handles means redirects, and Shopify redirects only from URLs that no longer load.
• Images: alt text that describes the product as a shopper would, which helps both accessibility and image search.
• Internal links: anchor text from guides and related categories that uses the cluster's own words, pointing at the owner page.

The page-level detail is covered in title tags, on-page SEO and product page SEO. Once the map exists, an e-commerce SEO audit checks whether the live store matches it.

Why most store keyword lists never leave the spreadsheet

A store with a few hundred products can do this in a week of evenings. A catalog of thousands, or one about to move platforms, usually needs the clustering and mapping done once properly. DataCram's keyword research and keyword mapping services produce exactly this sheet, tied to the products that pay, and our e-commerce development work builds the categories the map calls for.