Ecommerce Keyword Research: How to Find Keywords That Actually Sell

The stakes are higher than most store owners think. Jungle Scout’s Q1 2024 consumer survey found that more than half of US shoppers start online product searches on Amazon. And Adobe reports that AI traffic to US retail sites grew 393% year over year in early 2026.

So your keyword list now has three jobs. It has to win on Google, hold up against marketplace listings, and give AI assistants clear language to quote. Below is the exact process I use when I plan keywords for online stores, including the parts most guides skip.

What Makes Ecommerce Keyword Research Different?

A blogger wants traffic. A store owner wants orders. That one difference changes the whole job. A keyword with 20,000 monthly searches and no buying intent can be worth less than a phrase with 90 searches from people who already have their card out.

Stores also have a structure problem that blogs don’t. You have collections, product pages, filters, variants, brand pages, and guides, all competing for similar words. So keyword research for a store is really page planning. You are deciding what should exist, and what should never be indexed.

When I first look at an online store, the issue is rarely a lack of keywords. It’s usually the wrong page ranking for the right term, or a money keyword with no page at all. That’s why my ecommerce SEO consulting work always starts with a keyword-to-page map.

• Wrong page ranking
• Money keyword, no page

The Five Types of Ecommerce Keywords to Collect

Before you open any tool, know what you’re hunting for. I sort every store keyword into five buckets, because each bucket belongs on a different kind of page. Mixing them up is the fastest way to build pages Google never ranks.

Branded searches run across all five buckets. Track them, but don’t count them as wins from keyword research. People typing your brand name were coming anyway. Those searches show demand you already created, not demand you took from competitors.

How to Do Ecommerce Keyword Research in 8 Steps

The research phase feels slow. On a CBD store I worked on, we spent over a month researching topics before publishing. The site went from almost no traffic to more than 10K monthly visits in about four months. I shared that approach in my guide.

Step 1: Start With Your Catalog, Not a Keyword Tool

Export your product catalog and list every category, product type, material, use case, and audience it covers. This becomes your seed list. Starting here keeps you honest, because you only research terms you can actually sell today, not terms a tool thinks look exciting.

For a jewelry store, seeds might be “moissanite ring”, “tennis chain”, “cuban link”, and “grillz”. For a pet store, “grain free dog food” and “orthopedic dog bed”. Write them the way customers talk, not the way your supplier labels them in a spreadsheet.

Step 2: Mine the Words Your Customers Already Use

Your best keyword data is sitting inside your own business. Product reviews, support tickets, return reasons, live chat logs, and sales call notes show how buyers describe their problems. Their words rarely match the tidy product names your team picked during setup.

Don’t skip your internal site search report either. Shoppers type what they expected to find, typos and slang included. In an older but still telling benchmark, Baymard Institute found that 70% of top US ecommerce search engines failed on product-type synonyms. Stores and shoppers simply use different words.

Then open Google Search Console, go to the Performance report, and look for pages with high impressions but few clicks. Those queries prove Google already sees you as relevant. They are usually the fastest wins a store has, because the page exists and only needs sharper targeting.

Step 3: Expand the List With Google, Amazon, and AI Tools

Now widen the net. Type each seed into Google and note the autocomplete suggestions, People Also Ask questions, and related searches. Then repeat the same seeds in Amazon’s search bar. Amazon suggestions come from shoppers close to buying, so they lean heavily transactional.

Marketplace and social search deserve real attention. The same Jungle Scout survey found that nearly 40% of Gen Z shoppers start product searches on YouTube or TikTok. If your buyers skew young, spend ten minutes in those search bars too and note the phrases that keep coming up.

AI assistants are good for phrasing. I give ChatGPT or Gemini a buyer profile and ask how that person would search. It’s quick for spotting long-tail wording. But treat the output as ideas, not data, and check every phrase against real search or sales numbers before it earns a page.

Step 4: Read the SERP Before You Trust Any Metric

Search every keyword that might deserve its own page and look at what Google ranks. If the top ten are collection pages, Google sees buying intent. If they’re listicles and guides, shoppers want to compare first. Match the format Google rewards, or you’ll push uphill for months.

Watch for Shopping ads, free product listings, image packs, and videos too. Google says people shop across its surfaces more than a billion times a day, drawing on over 60 billion product listings. A product-heavy results page tells you your product data matters as much as your copy.

Also check who ranks. A first page full of Amazon, Walmart, and category giants means a head term will take real authority to crack. A first page with small stores and forum threads is an opening, and it’s worth moving on quickly before someone else notices it.

Step 5: Treat Volume and Difficulty With Healthy Suspicion

Search volume and difficulty are estimates, and every tool calculates them its own way. According to Ahrefs’ research, Google Keyword Planner overestimates search volume 54.28% of the time. Use these numbers to compare keywords against each other, not to forecast exact traffic.

Low numbers don’t mean low value, either. Ahrefs’ long-tail data shows that almost 93% of keywords in its US database get fewer than 10 searches a month, and roughly 15% of daily Google searches are brand new. Plenty of profitable product phrases show zero volume and still bring orders.

For difficulty, I look past the score to the pages that rank. If they have strong link profiles, a new store needs links before competing. It helps to of the top results, so you can judge whether the gap takes months or years to close.

Step 6: Score Every Keyword by Revenue Potential

This is where store owners should think differently from bloggers. Instead of sorting by volume, estimate what a keyword could earn. I use a simple formula: estimated monthly clicks multiplied by your conversion rate, multiplied by your average order value. It’s rough, but it ranks opportunities by money.

Here’s a worked example with round, made-up numbers. Say a phrase gets 1,000 searches a month and a top ranking earns about a quarter of them, so 250 clicks. At a 2% conversion rate and a $120 average order, that page brings in roughly $600 a month.

Now take a 90-search phrase for a $900 product. With the same share of clicks, you get around 22 visitors. At a 6% conversion rate, which is realistic for a very specific query, that’s about $1,200 a month. The smaller keyword wins, and a volume-sorted list would have buried it.

Then run two more checks. Can you realistically rank within six to twelve months? And do you have the stock and margin to support the sales? A keyword that scores high on revenue but fails either check goes on a later list, not the current sprint.

• Rank in 6 to 12 months?
• Stock and margin?

Step 7: Map One Primary Keyword to One Page

Group keywords that share the same intent, then give each group one home. A collection page might target “moissanite tennis bracelet” plus close variants like “moissanite tennis bracelets for women”. Close variants don’t need separate pages. Separate intents do.

The quick test is the search results again. If Google shows mostly the same pages for two phrases, they belong on one page of yours. If the results are mostly different, split them. This one check prevents most keyword cannibalization before it has a chance to start.

Cannibalization is very common in stores that blog. A guide called “best moissanite rings” and a collection targeting “moissanite rings” can end up fighting each other. Keep the commercial head term on the collection, and point the guide to it with a clear internal link.

Here’s how I usually match page types to keyword patterns when I build the map for a store. Treat it as a starting point and let the search results for your own terms confirm each choice:

Step 8: Decide Which Filter Pages Deserve to Be Indexed

Faceted navigation is where ecommerce keyword research meets technical SEO. Every filter for color, size, brand, and price can create a new URL. Google’s own documentation warns that these combinations can create near-endless URL spaces, wasting crawling on pages nobody searches for.

My rule is simple. A filter combination earns an indexable, optimized page only when it has real search demand, enough products to fill it, and an intent of its own. “Black leather sofa” might qualify. “Black leather sofa under $900 sorted by newest” never will.

Everything else stays usable for shoppers but out of the index. Depending on your platform, that means canonical tags, noindex, or robots.txt rules. The keyword research tells you which few filter pages to promote, and the technical setup quietly handles the thousands you don’t want.

Platform Quirks That Change How You Map Keywords

The keyword process is the same everywhere, but each platform decides which pages you can create and how URLs behave. I’ve seen solid keyword research fail simply because the platform quietly generated duplicate pages. Here’s what to check on the major ones.

Shopify

Shopify stores target category terms on collection pages. Products can also load under collection paths, though Shopify points canonicals to the main /products/ URL. Tag-filtered collection URLs are the bigger risk, because they create thin duplicates. My Shopify SEO consultant work usually starts by deciding which tags deserve real collections.

WooCommerce

WooCommerce gives you product categories, product tags, and optional attribute archives. Categories should carry your main commercial keywords. Tags and attribute archives often repeat them, so I either give each one a unique keyword and copy or keep it out of the index. It’s a regular WooCommerce SEO fix.

Magento (Adobe Commerce)

Magento stores tend to have big catalogs, and layered navigation generates parameter URLs very fast. Here, keyword research has to feed straight into indexing rules, so the few filter combinations with real demand become clean, optimized landing pages. That’s where most of my Magento SEO time goes.

BigCommerce

BigCommerce creates brand pages alongside category pages, which is handy if you stock recognizable brands. Map “brand plus product type” keywords to those brand pages instead of forcing them into categories. Faceted search still needs the same indexing discipline covered above in any BigCommerce SEO setup.

Wix and Squarespace

Both platforms work within tighter templates. Before you assign a keyword cluster to a category page, check which text, heading, and meta fields that page actually supports. If it can’t hold helpful copy, target that cluster with a product page or a buying guide instead.

Smaller catalogs are common on both, and that’s good news. With fewer pages, each keyword gets more care. I usually advise Wix SEO clients to win five or six strong category terms before chasing long lists of product variations.

The same logic holds for Squarespace SEO. Build fewer, better pages, give each one a single primary keyword, and support it with FAQs written from real customer questions rather than guessed ones.

Keyword Research for AI Search and Shopping Assistants

AI search has changed what a keyword looks like. Shoppers now type full prompts like “best moissanite chain for daily wear under $500 that won’t tarnish”. In that same Adobe report, 39% of US consumers said they have already used AI for online shopping.

You can’t pull prompt volumes the way you pull keyword volumes. What you can do is collect the questions, limits, and comparisons buyers mention, then answer them plainly on collection and product pages. Specs, materials, sizing, and honest use cases give AI tools facts they can quote.

I’ve seen this pay off. On a high-ticket jewelry store I worked on, AI visitors grew from 117 in January to 1,399 a month by December, and that AI traffic produced $29,174 in revenue over the year. Clear, factual product content did most of the work.

For the wider picture, my breakdown of covers how AI answers are reshaping clicks and what that means for store owners. The short version: the brands with the clearest product facts are the ones AI tools recommend.

Plan Ahead for Seasonal Keywords

Many store keywords spike on a calendar. Gift terms, holiday decor, swimwear, and back-to-school supplies all follow predictable curves. Pull five years of Google Trends data for your core terms and note when interest starts climbing, not just when it peaks.

Publish or refresh seasonal collection pages two to three months before that rise, since pages need time to be crawled and ranked. I keep evergreen URLs like /collections/christmas-gifts live all year and just swap the products, so the page builds authority across seasons instead of restarting every December.

Free vs Paid Ecommerce Keyword Research Tools

You don’t need an expensive tool stack to start. Free tools cover most of the basics, and paid tools mainly save time and add competitor data. Here’s how I’d split them for a store owner doing this for the first time.

One tip: the top-of-page bid in Keyword Planner is a decent proxy for commercial value. Advertisers pay more for terms that convert. A keyword with a high suggested bid and modest volume is often a better SEO target than a cheap, high-volume one.

Ecommerce Keyword Mistakes I See in Store Audits

How to Track Whether Your Keywords Are Working

Give new or updated pages at least two to three months before judging them. Then track rankings by page type, not only by keyword. I want to see how collections perform as a group, because collections usually carry the commercial terms that drive store revenue.

Search Console has blind spots. Ahrefs found that 46.08% of clicks in Search Console go to hidden, anonymized queries. So pair Search Console with revenue per landing page in your analytics. A page earning organic sales is working, even when you can’t see every query behind it.

Revisit your keyword map every quarter. Add new products, retire dead terms, and check whether two pages have started ranking for the same query. Ecommerce keyword research is never really finished, because your catalog and your buyers keep changing.

Frequently Asked Questions

Written by

Abdullah Mahmud

When I first look at an online store, the issue is rarely a lack of keywords. This is the exact process I use when I plan keywords for online stores, including the parts most guides skip.

Keyword researchCustomer languageRevenue scoringKeyword mappingFilter pagesAI search

Ecommerce SEO consulting Shopify SEO consultant