Keyword research for ecommerce: how to choose what to rank for in your store

By Julián, founder of Rankedora · 2 November 2026 · 9 min read

Almost all online stores end up ranking for what they already sold: their products’ internal names and the categories the developer set up. The problem is that shoppers don’t search with your catalogue in mind. They search by product type, by attribute, by problem or by comparison, and each of those searches calls for a different page. Good keyword research is, above all, a job of deciding which page answers which search.

The five types of search in an online store

Before opening any tool, it helps to know what types of search exist and which page should answer each one. The examples are from a sports footwear store.

Category and attribute searches are usually the ones that move the most sales per visit, because the shopper already knows what they want. Informational ones attract more traffic and convert less, but they open the door to people who haven’t decided yet and are the ones AI assistants cite most.

Step 1: gather ideas from your own data before tools

The best ideas are usually already in your business. Start there and leave the tools for expanding.

• Search Console: in the performance report you will see the queries you already appear for, even in low positions. A query with impressions and no clicks is a page that exists but doesn’t convince, or a page you are missing.
• Your store’s internal search: what your customers type there is the real vocabulary of your catalogue. If they search for it and don’t find it, you have a category pending.
• Customer service and reviews: repeated questions about sizes, materials, compatibility or delivery times are topics for guides and product pages.
• Autocomplete, “People also ask” and related searches: Google shows you how users complete the sentence. It is free and very useful for vocabulary and questions.
• Google Ads Keyword Planner: gives volume ranges and suggestions. Use it as guidance, not as an exact figure.
• Your competitors’ navigation: their categories and subcategories are a free list of groupings that already work for them.

Step 2: validate with the SERP which page can rank

It is the step most often skipped and the one that avoids the most mistakes. Search for the keyword on Google and look at which type of page occupies the top results:

• If you mostly see store categories, an informational guide won’t compete for that term. You need a category.
• If you see guides, listings and comparisons, a category doesn’t fit that intent. You need editorial content that links to your categories.
• If there is a mix, there may be room for both formats, but each with its own keyword.
• Also look at the extras on the results page: product carousel, video, AI answers. They change how much traffic actually reaches the website even if you rank well.

Also look at who is competing. If the top ten are brands with a lot of authority and you have just started, look for the more specific version of that search, for example by adding the attribute that sets you apart. That is usually where your first win is.

Step 3: one main keyword per URL

With the ideas validated, build a map. Each row is a URL and each URL has a main keyword with its group of variants of the same intent. This map prevents two pages from competing with each other, a problem we cover in SEO cannibalisation in ecommerce.

The “Action” column is the most important: create, optimise, merge or leave as is. A map that only adds pages makes the problem worse. It is common in stores with years of history for some of the actions to be merges.

Step 4: prioritise by business, not just volume

Not every keyword deserves the same effort. Score each row from 1 to 3 on these criteria and start with the ones that add up to the most:

• Margin and order value: rank first for whatever leaves the most profit per order.
• Stock and continuity: don’t build a category around something you won’t be able to supply. If the catalogue changes a lot, the structure of the out-of-stock page matters as much as the keyword.
• Closeness to purchase: a category or subcategory weighs more than a general guide.
• Realistic difficulty: how many strong brands occupy the SERP and which of your pages could compete.
• Effort: optimising a page that already has impressions usually delivers results sooner than creating one from scratch.
• Seasonality: if demand rises at a specific time, the page must be ready in advance, not when the campaign starts.

Step 5: add the questions people ask AI

Conversational searches in ChatGPT, Gemini or Google’s summaries are looking more and more like a question to a salesperson: “which running shoes do you recommend to start running with sensitive knees”. There is no official tool that gives the volume of those questions, so they are worked on differently: collect your customers’ real questions, write them up as they are and answer them at the start of a guide or product page, with concrete data. The step-by-step method is in how to create a buying guide that ranks and that AI can cite, and the difference between the two worlds in SEO vs GEO for ecommerce.

Common mistakes

• Choosing only by volume and ending up ranking for generic terms that don’t buy.
• Putting the same keyword on several pages and letting them compete.
• Trying to rank a category with a type of content that Google doesn’t reward for that search.
• Creating a page for every colour or size variant with no demand of its own.
• Not reviewing the map after catalogue changes and being left with orphan pages.
• Taking tool volumes as exact figures.

Quick checklist

• List of queries exported from Search Console and the internal search.
• Search types classified: category, attribute, product, guide, comparison.
• SERP reviewed for every important keyword.
• Map with one main keyword per URL and an assigned action.
• Priority by margin, stock, closeness to purchase and effort.
• Real customer questions incorporated into guides and product pages.
• Review every quarter and before each campaign.

Frequently asked questions

Sources consulted

• Google Search Central: SEO starter guide
• Search Console Help: search performance report