Ecommerce Keyword Research: How to Map Terms to Pages

What does ecommerce keyword research produce?

The output of a keyword research project is a table. Each row lists a term, the page intended to rank for it, and what that page still needs.

Match each term to a page you already have, and note the terms that need a new one.

Your platform determines which page types you can create. The 2025 Web Almanac identified 19.9 percent of the desktop sites it analyzed as stores, with WooCommerce at 35.4 percent of them and Shopify at 21.5 percent.

On either platform, you can assign keywords to collection and product pages.

Common ecommerce platforms let you create a home page, collection pages, product pages and articles. Every term you keep gets assigned to one of them.

One of your pages can rank for many search terms. Ahrefs sampled 3 million searches in 2017 and found that pages ranking first also ranked in the top ten for roughly 1,000 other keywords, with a median of about 400.

For example, say you sell desk lamps. Assign desk lamps to a collection page and brass desk lamps to a smaller collection or a filter.

A model name goes on the product page. A question about how bright a desk lamp should be goes in an article.

If you run Google Shopping, use the same model name in your product feed. The feed determines which products are matched to a query. For the feed rules, see optimizing your Google Shopping feed.

Page types and the terms to target

Google recommends a separate URL for each variant, with the canonical on the parameter-free version. Check that rule before you assign a term to a variant page.

This guide covers the assignment only, and ecommerce SEO covers the order of the rest of the work.

Which page should own each keyword?

Before assigning a term to a page, search that exact wording and review the results. Count the collection pages, the product pages and the articles.

If collection pages, product pages and home pages dominate, assign the term to a page that sells. If guides, definitions and comparisons dominate, assign it to an article.

Kevin Indig tracked more than 4,000 keywords and almost 40,000 product grids over nine months to February 2026, writing for Search Engine Journal.

He counted 3,367 category-page URLs and 90 product-page URLs ranking inside those grids.

AI Overview appearances also vary by search intent. Semrush tracked more than 600,000 US keywords across 10 industries from November 2025 to April 2026.

Semrush found that AI Overviews rose 71 percent on commercial-intent results and fell 5 percent on transactional queries.

For each target query, check which page types rank and how far an AI Overview pushes them down the results.

Use volume to order the list and the result types to choose the page. Ahrefs compared Keyword Planner volume with Search Console impressions for 72,635 keywords in 2021.

In Ahrefs’ sample, 54.28 percent of Keyword Planner figures were overestimates. Another 45.22 percent were within half of the Search Console count.

A term with little reported volume can still be worth targeting. For a calculation of a low-volume term’s value, see B2B keyword research without volume.

Which keywords deserve a filter page?

Every filter combination a shopper can click creates a URL a crawler can reach, but most combinations do not correspond to a term anyone searches. Each combination you leave crawlable costs crawl time and earns nothing.

A category page with three filter groups, and the address they produceCount the filter URLs before assigning keywords to them

Three filter groups with three options each produce 27 combinations; adding more groups increases the total.

Decide which filters deserve keyword targeting, then control how Google crawls the rest.

Oftentimes there’s no good reason to allow crawling of filtered items.

Google Search Central, Crawling and managing faceted navigation updated 18 December 2025, read 20 October 2026

Google warns that filter parameters can create an unlimited number of URLs. Crawlers then spend their time on those URLs, and your new pages are found more slowly.

So keep one unfiltered listing page and your item pages crawlable. Disallow the filter parameters you do not want crawled, and return a 404 for a filter combination with no results.

Create a filter page only if it meets all three conditions below, starting with a term that recurs month after month in your own data.

The page must also add something the filtered grid does not: what the material is suitable for, or how the sizes run. And the products in it have to stay in stock.

On Shopify you can build that page without hand-picking the products. Shopify gives four controls: automated inclusion, manual inclusion, automated exclusion and manual exclusion.

Shopify is also replacing its legacy manual and smart collections with a new collections model, so check which controls your admin offers before you plan the page. For the platform limits around filter pages, see Shopify SEO and faceted navigation.

Demand for that filter term can also move with the season. That is the next thing to check.

How to read seasonal keyword demand

Demand for a recurring term can vary through the year, and an annual average hides that variation. Use monthly search figures to schedule the work on a seasonal catalog.

Google Ads averages the searches for a term over twelve months by default, rounds what it shows you, and reports it for exact matches only.

Google Trends measures relative search interest by dividing each data point by the total searches in the selected location and period, then scaling the result from 0 to 100.

Salesforce estimated global online sales of $336.6 billion during Cyber Week 2025, based on the activity of more than 1.5 billion shoppers in more than 89 countries.

A twelve-month average smooths out a peak like Cyber Week, so check the demand in the month you plan to target.

For example, say you sell wool blankets. Pull the monthly figures for your top terms, find the month the climb starts, and schedule the page work for the month before it.

Then compare the pattern with last year’s search demand in your own store. Your store’s peak month may differ from the one the keyword tool shows.

How to map keywords to store pages

With the monthly and annual numbers in hand, map the keywords in six passes. Only the last two passes involve writing; the others are decisions based on data you already have.

1. List the pages you already have

   Export your collection URLs and your product URLs. For the full page-inventory process, see the ecommerce SEO audit.

2. Find search volumes for your candidate terms

   Start with the words in your collection names and product titles, then add the related and suggested terms the tools return. Keep the monthly figures as well as the annual average.

3. Count the page types that rank for each term

   Search each candidate term and count the collection pages, product pages and articles in the results. Write the most common type next to the term.

4. Assign one page per term

   Choose an existing store page of the type most common in the results. If two pages could target the term, assign it to the one already receiving impressions and choose a different target for the other.

5. Choose which new pages and filter pages to create

   For each unassigned term, decide whether to create a filter page or another type of page. Apply the three conditions from the filter section first.

6. Write only what the assigned page lacks

   Most assigned pages need a title, one short block of copy and a product grid that matches the term.

Avoid spending more time on copy than the page needs. Digitaloft collected 300 UK category pages ranking first for commercial terms on 13 August 2025 and measured the unique copy on them.

Digitaloft found an average of 310 words per page. Some 44 percent had fewer than 200 words, and 10 percent had nothing beyond the H1.

So do not use word count alone to decide how much copy a page needs. Use the term in the title, answer the shopper’s question in a short block, and move to the next row in the table.

For example, a brass desk lamp collection needs two sentences explaining where brass works well and a grid of brass lamps.

What ecommerce keyword research costs

Budget for two costs: the tool subscription and the time spent mapping, checking and writing each page. Subscription prices are published, and the number of collection pages decides how many hours the work takes.

If your store has 40 collection pages, you have 40 keyword assignments to make and 40 pages to check afterwards.

So ask how many pages a quote covers, since the keyword count does not show how much mapping and checking is required.

If you want to outsource keyword mapping, our ecommerce SEO service starts with that work.

Common ecommerce keyword research mistakes

Watch for these four mistakes when mapping keywords to catalog pages. Each one has a fix you can apply to the table you just built.

First, do not build a page per keyword. Group the terms one collection page can target, because splitting them across pages creates thin pages that compete with one another.

Second, do not leave two of your pages chasing one query.

Keep the term on the page whose type appears most often in the results, and give the other page a different target.

Link from the page you retargeted to the page you chose for the original term, using anchor text that fits the sentence it sits in. For advice on placing links, see the internal linking guide.

Third, do not discard a term solely because a tool reports low search volume. Ahrefs reports that 2.3 billion keywords in its US database, almost 93 percent of the total, receive fewer than ten searches a month.

Check whether those low-volume searches match products you already stock.

Fourth, do not assume Google sees all your demand. Feedvisor surveyed 1,006 US consumers with Zogby Analytics in August 2026.

Feedvisor put the share who start a product search on Amazon at 62 percent, down from 80 percent a year before.

Compare that with the demand you see in your own marketplace search data. Record Google terms and marketplace terms in separate columns.

Shoppers also search with AI assistants, and those select products on information you may not control. For advice on updating that information, see the ChatGPT shopping guide.

And do not generate filter pages at scale just because the terms exist. Each page you create needs stocked products and useful copy.

How to check the keyword map worked

Use a Search Console report to check whether the assigned pages receive impressions for their target terms. Group the performance table by page, select an assigned page, then switch to the query dimension to see which searches produced them.

The page’s leading query should be the term you assigned to it.

Check whether impressions for the target term go to the assigned page or are split between pages on your site.

Search Console, Performance and Queries, filtered to one pageCheck which queries bring impressions to the assigned page

One page, and the searches that brought it impressions.

Search Console reports your store’s own impressions for the countries and devices you select. A keyword tool only estimates them.

For example, if your brass desk lamp collection also gets impressions for desk lamps, reassign the broader term to the broader page.

Some low-volume queries will be missing from the table. Google says queries issued by fewer than a few dozen users over two to three months are left out of the tables.

These queries still count toward the chart totals, so a low-volume term can receive impressions without appearing in the table.

Ranking position alone does not tell you how many clicks a term earned. Advanced Web Ranking’s click-through study, updated on 19 October 2026, uses anonymized Search Console data from millions of US results.

Advanced Web Ranking reports that, when an AI Overview appears, click-through rates for the top result have typically fallen by around half. Transactional queries fall the least.

So judge an assigned page by the clicks it receives, and consider its ranking alongside that total.

Open your five biggest collection pages in Search Console this week and note the leading query on each one. Start with any page whose leading query differs from the term you assigned.