By Rahul Saini, Founder and SEO Strategist at SearchCounselCo. Updated September 30, 2026.

1. Why ecommerce keyword research is different

Generic keyword research often stops at relevance, search volume, and difficulty. Ecommerce adds a second question that is just as important: what type of page should satisfy the search?

A broad product-class query may belong on a category page. A model-specific query may belong on a product page. A feature or use-case query may deserve a subcategory. A comparison query may need a buying guide. Some attribute queries may justify an indexable filtered landing page. Others should remain ordinary filters that are useful for shoppers but should not become search landing pages.

That is why the goal is not “find the biggest keywords.” The useful chain is search demand → intent → SERP pattern → page type → URL owner → action.

If you want the broader method outside ecommerce, use the general keyword research guide. This page focuses specifically on ecommerce architecture and URL decisions.

2. Build seed keywords from your catalog

Start inside the store, not inside a keyword tool. Your catalog gives you the commercial vocabulary that matters to the business.

• Category seeds: product families such as “standing desks,” “running shoes,” or “ceramic cookware.”
• Product seeds: product names, brands, model numbers, materials, and specifications.
• Attribute seeds: size, color, capacity, compatibility, ingredients, fit, material, or technical features.
• Use-case seeds: “for small kitchens,” “for wide feet,” “for travel,” or “for beginners.”
• Problem seeds: the problem the product solves, such as “desk for back pain” or “quiet blender for apartment.”

You do not need hundreds of seeds. You need enough to represent the real structure of the catalog. A focused set is more useful than a giant export filled with products or topics you do not sell.

3. Find the language shoppers actually use

Expand each seed with several sources because no single keyword database captures the entire market.

The strongest ecommerce research combines third-party estimates with first-party evidence. Tool volume can suggest where demand exists. Search Console and customer language show what your own site and customers are actually exposing to you.

4. Classify intent with the live SERP

Modifiers are useful clues, but the live search results are the better final check. Search the term and record the dominant result type. If the results are category pages, a long informational article may be the wrong asset. If the results are comparisons, a thin product listing may not satisfy the task.

SearchCounselCo’s search intent guide explains the four standard intent types. For ecommerce, use them like this:

Do not force a page type because a keyword “sounds transactional.” Verify what the results reward. A mixed SERP can indicate that users need both selection help and a path to products.

5. Map each keyword cluster to the right page type

This is where ecommerce keyword research becomes architecture.

For the implementation side, see category page SEO, subcategory and attribute pages, product page SEO, and buying guides and comparison pages.

6. Category, subcategory, facet, or no new URL?

One of the hardest ecommerce decisions is deciding whether a valuable attribute deserves a permanent search landing page or should remain only a filter.

7. Cluster keywords without creating cannibalization

Several phrases can describe the same underlying task. If the same searcher would be satisfied by the same page, treat those phrases as one cluster.

For example, “waterproof trail running shoes,” “trail shoes for rain,” and “waterproof running shoes for mud” can often live on one strong page. Creating three thin URLs may split internal links and make URL ownership less clear.

Use two checks:

1. Intent check: Would the same page satisfy the same user task?
2. SERP-overlap check: Do many of the same URLs rank for both searches? If yes, that is evidence one page may serve both.

When multiple pages already compete, use the keyword mapping and cannibalization guide before creating another URL.

8. Run a competitor keyword-gap workflow

Do not treat every keyword a competitor ranks for as an opportunity. The goal is to find relevant commercial gaps that fit your catalog and deserve a page.

1. Choose 3–5 search competitors. Use domains that repeatedly appear for the product categories and queries you care about. They may not be your closest business competitors.
2. Export their ranking keywords and URLs. Use a competitive keyword tool to see which terms and pages generate visibility.
3. Filter aggressively. Remove brand-only queries, products you do not sell, irrelevant countries, and informational terms that do not support your commercial architecture.
4. Group the remaining gaps by intent. Separate product-class demand, attribute/subcategory demand, specific product demand, comparisons, and informational questions.
5. Inspect the SERP for the best clusters. Record the page types that rank rather than assuming the competitor’s page type is correct for you.
6. Map each cluster to your own architecture. Decide whether an existing URL can own it or whether it represents a genuinely new page.
7. Apply the five-action framework. Protect, Improve, Consolidate, Create, or Do Not Target.

The key rule: a competitor gap is only useful when the search intent, products, inventory, and page type also fit your business.

9. Prioritize what to work on first

Do not sort the roadmap by search volume alone. Use four questions:

• Observed demand: Is there real search activity in Search Console or credible third-party data?
• Business relevance: Does the query map to products, categories, or decisions that matter commercially?
• Ranking feasibility: Does the SERP include page types and sites you can realistically compete with?
• Implementation value: Can one improvement strengthen an important existing URL, or does it require a new page, taxonomy change, or development work?

An existing page with relevant impressions at positions 11–30 may deserve attention before a brand-new head term with much larger estimated volume. Existing visibility is evidence that Google already associates the URL with the topic.

10. Mine Google Search Console for existing opportunities

Search Console is especially valuable because it shows searches where Google has already tested your URLs.

1. Filter by a commercial URL group. Start with categories, products, or one content cluster rather than the whole site.
2. Inspect queries with impressions but weak clicks. Separate promising near-page-one terms from queries buried far beyond page one.
3. Group by intent. Do not treat every spelling variation as a separate opportunity.
4. Check which page Google selected. If the wrong URL appears, investigate internal links, overlap, titles/H1s, and page purpose.
5. Compare equivalent periods. Account for seasonality, device mix, country mix, and partial dates before declaring a gain or loss.

11. First-party example: how SearchCounselCo applied the framework to this page

This guide itself is a useful example of why keyword research should lead to a URL decision instead of an automatic rewrite or new page.

In SearchCounselCo’s supplied Search Console export for https://searchcounselco.com/keyword-research/, the page recorded 8 impressions, 0 clicks, and an average position of 24.88 across the exported roughly three-month window. The query rows were too limited to provide a reliable query-level opportunity map.

That evidence did not justify creating another ecommerce keyword-research URL. The page already existed, matched the intended topic, and had at least some search visibility. The appropriate decision was:

This is a page-level editorial example, not a claim that the page is already ranking successfully. Future performance must be measured after publication using comparable Search Console periods.

12. Build the keyword-to-URL decision map

The deliverable should be a working decision sheet, not a raw keyword export. At minimum, track:

Starter CSV:

13. Validate the keyword before you build or rewrite

Before committing a URL to a keyword cluster, run a short live-SERP check:

• What page types dominate the first page?
• Are the results product-led, category-led, editorial, marketplace-heavy, local, or mixed?
• What information appears repeatedly because users clearly need it?
• What useful decision support is missing or weak?
• Can your page add something real: better comparison logic, first-party data, clearer taxonomy, stronger attributes, original visuals, or a practical tool?

SearchCounselCo’s SERP-intent guide goes deeper on using current results to validate page type before building.

14. Ecommerce keyword research for AI search

Do not create a separate “AI keyword list” and assume it needs different fundamentals. Google’s current guidance says AI Overviews and AI Mode use the same core Search systems and do not require special schema or a separate machine-readable file for eligibility.

What changes is the shape of the question. Shoppers may ask longer, more specific comparison or recommendation questions. That makes long-tail research, product attributes, buying criteria, comparison content, and structured catalog information even more useful.

For generative search surfaces, keep the basics strong: crawlable pages, clear internal links, important information in visible text, useful images where relevant, accurate structured data that matches the page, and content with enough original value to justify a visit. SearchCounselCo also covers store-specific implications in AI shopping recommendations and the broader topic in AI SEO.

15. Common ecommerce keyword research mistakes

• Chasing volume before relevance. Large demand does not help if the query cannot reasonably lead to your products.
• Ignoring page type. A good keyword on the wrong asset is still a bad targeting decision.
• Creating one page per keyword. Close variants usually belong in clusters.
• Making every filter indexable. Useful ecommerce filtering does not automatically deserve a search landing page.
• Using difficulty as a hard rule. Treat tool scores as comparative evidence, not absolute thresholds.
• Copying competitor gaps blindly. Their catalog, authority, market, and architecture may not match yours.
• Throwing away long-tail terms because a tool reports little volume. Specific queries can reveal real subcategory and product needs.
• Leaving GSC out of the process. Your own impression data can expose opportunities generic tools miss.
• Publishing without assigning URL ownership. A new page can compete with one you already have.
• Never revisiting the map. Products, inventory, search behavior, and SERPs change.

16. A 20-minute starter workflow

1. Choose one category. List the category name, key products, attributes, problems solved, and common use cases.
2. Expand the language. Pull suggestions from Google, one keyword tool, marketplace autocomplete, competitors, and first-party data.
3. Mark intent. Label each cluster informational, commercial, transactional, or navigational.
4. Open the SERP. Record the dominant page type for the most important terms.
5. Assign one preferred URL. Reuse an existing page where it fits; create a new page only for a genuinely distinct task.
6. Choose the action. Protect, Improve, Consolidate, Create, or Do Not Target.

Repeat category by category. You will end with a usable search architecture and prioritized work queue instead of a keyword spreadsheet nobody acts on.

Where to go next

Once the map is built, move from research into implementation: strengthen category pages, build justified subcategory and attribute pages, improve product pages, and use buying guides and comparison pages for commercial-investigation intent. For the full store-wide framework, return to the Ecommerce SEO guide.

Sources and further reading

• Google Search Central: AI features and your website
• Google Search Central: optimizing for generative AI features
• Google Search Console: Performance report overview
• Shopify: Keyword Research for Ecommerce
• Semrush: Ecommerce Keyword Research
• Ahrefs: Ecommerce SEO
• Backlinko: Ecommerce Keyword Research

Editorial note: Search volume, keyword difficulty, and traffic estimates vary by tool and market. Treat third-party figures as estimates and validate important decisions against first-party data, product reality, and the live SERP.