To validate an e-commerce product category, run a card sort with 15-20 target shoppers to see how they naturally group your products, then follow up with a tree test to confirm they can actually find items within your proposed taxonomy. Most merchandising teams build categories around supplier logic or internal inventory systems, and the result is a navigation structure that makes perfect sense to the team but baffles actual buyers. Ecommerce category validation flips this by letting real shoppers tell you where products belong before you commit to a structure that tanks conversion rates.

Key Takeaways

Time required: 1-2 weeks from study setup to validated taxonomy
Difficulty: Beginner to Intermediate
What you need: Product list, 15-20 shoppers matching your target audience, a card sorting and tree testing tool
Key tip: Always validate product taxonomy with both a card sort (how shoppers group) and a tree test (whether they can find)---one method alone leaves blind spots

What You'll Need

A list of 30-60 representative products from your catalog (or the full list for smaller stores)
15-20 participants who match your target buyer profile
ValidateThat account (free at validatethat.io)
Your current or proposed category structure (if you have one)
A spreadsheet to track findings and compare structures

Step 1: Build Your Product Card List

Pull together the products you want to test. If your catalog has fewer than 60 SKUs, use them all. For larger catalogs, select a representative sample that covers every proposed category proportionally---include bestsellers, long-tail items, and anything that's hard to categorize.

Write each card using the language shoppers see on your site, not internal SKU names or supplier terminology. "Women's Oversized Linen Blazer" works. "WLB-OVR-2026-NTL" does not. If you sell products with overlapping attributes (a moisturizer that's also SPF protection), include them---these cross-category items are exactly where taxonomy breaks down and where validation matters most.

Pro tip: Pull your top 20 site-search queries from analytics. If shoppers are searching for terms that don't map cleanly to your current categories, those queries should inform how you write your product cards.

Step 2: Run an Open Card Sort

Set up an open card sort in ValidateThat where participants group your product cards into whatever categories make sense to them and name those categories themselves. Open sorting is critical here because it reveals the mental models shoppers actually use, not the ones you assume they use.

Give participants clear instructions: "Imagine you're organizing these products on a shopping website. Group them in a way that would help you find what you need." Keep the study unmoderated so participants sort without outside influence, and set a reasonable time limit of 15-20 minutes to keep completion rates high.

Aim for 15-20 completed responses. This sample size consistently surfaces the dominant grouping patterns without requiring a massive recruitment budget.

Pro tip: If you already have a category structure you're questioning, resist the urge to run a closed card sort first. Open sorting prevents you from anchoring participants to your existing (potentially broken) taxonomy.

Step 3: Analyze Grouping Patterns

Once responses are in, look for product pairs that 60% or more of participants placed in the same group. These high-agreement clusters are the backbone of your validated taxonomy. Pay equal attention to participant-created category names---if 12 out of 15 shoppers call a group "Skincare" but your site calls it "Face & Body Care," the shoppers win.

Build a similarity matrix to visualize which products cluster tightly and which float between categories. Products that land in different groups across participants are your cross-category items. You have three options for these: place them in the strongest cluster, list them in multiple categories with cross-links, or reconsider whether the product name is causing confusion.

Watch for unexpected groupings. Shoppers might group products by use case ("Date Night Outfits") rather than by product type ("Dresses," "Shoes," "Accessories"). These insights can reshape your entire category approach.

Pro tip: Export the raw data and compare participant-created category names side by side. Synonyms are valuable---they become your category SEO keywords and filter labels.

Step 4: Draft Your Proposed Category Structure

Translate your card sort clusters into a concrete taxonomy. Aim for 5-8 top-level categories, each with 3-7 subcategories. This range keeps cognitive load manageable while giving shoppers enough specificity to narrow their browsing.

Use the most common participant-generated category names as your starting point. Where participants disagreed on naming but agreed on grouping, pick the term that aligns with how shoppers search (check your site search data and Google keyword volumes).

Map every product in your catalog to the new structure, not just the ones you tested. The card sort gives you the organizing principles; now apply them consistently. Flag any products that don't fit cleanly---these will become your tree test's toughest tasks.

Pro tip: Create two or three structural variations if your card sort data supports multiple valid groupings. You'll test them head-to-head in the next step.

Step 5: Validate Findability with a Tree Test

A card sort tells you how shoppers group products. A tree test tells you whether they can find a specific product within your proposed structure. Set up a tree test in ValidateThat using your drafted taxonomy (without any visual design---just the text hierarchy).

Write 8-12 task scenarios based on real shopping intents: "You're looking for a birthday gift for a 10-year-old. Where would you find it?" or "You need a waterproof jacket for hiking. Navigate to it." Include tasks that target your cross-category products and any areas where the card sort data was ambiguous.

Recruit a fresh set of 15-20 participants (not the same ones from your card sort) to avoid familiarity bias. Measure three things: success rate (did they find the right category?), directness (did they navigate straight there or backtrack?), and time to completion.

Pro tip: A success rate below 70% on any task signals a category problem. Check whether the issue is naming (shoppers don't recognize the category label), placement (the product is buried too deep), or structure (the category hierarchy doesn't match shopper logic).

Step 6: Run a Purchase-Intent Survey on Weak Categories

For categories where tree test performance was shaky or card sort agreement was low, run a short survey to understand shopper expectations. This is especially valuable for new categories you're considering adding.

Ask targeted questions: "If you saw a category called [proposed name], what products would you expect to find there?" and "When shopping for [product type], which of these category names would you click first?" Use multiple-choice and open-text formats to capture both quantitative preferences and qualitative reasoning.

A quick survey in ValidateThat with 30-50 responses can settle naming debates, validate whether a new category has enough perceived value to justify its own section, and surface terminology gaps between your team and your shoppers.

Pro tip: Include one question about purchase intent: "How likely are you to browse a [category name] section?" Low intent on a proposed category is a signal to merge it into an existing one rather than launching it standalone.

Step 7: Finalize, Implement, and Monitor

Merge your card sort clusters, tree test results, and survey data into a final taxonomy decision. Prioritize changes where all three data sources align---these are your highest-confidence moves. Where data conflicts, lean toward the tree test results, since findability directly impacts conversion.

Document the rationale behind each category decision so your team can reference it during future catalog expansions. When you add new product lines later, you'll know whether to create a new category or fold products into an existing one based on the validated mental models.

After launch, track category-level metrics for 30 days: bounce rate per category page, exit rate, add-to-cart rate, and internal search frequency. If shoppers are still searching for items that live in your new categories, the labels may need adjustment even if the structure is sound.

Pro tip: Schedule a follow-up tree test 6 months after launch or whenever you add a product line that represents more than 15% of your catalog. Taxonomies drift as catalogs grow.

Pro Tips

Test with shoppers who match your actual buyer demographics, not colleagues or friends. Internal teams are blind to their own organizational assumptions, and the whole point of ecommerce category validation is to escape those assumptions.
Run your card sort and tree test as separate studies with different participant pools. Using the same participants for both creates a learning effect that inflates your tree test success rates and gives you false confidence.
Include "distractor" products in your card sort that sit at category boundaries. How participants handle a product like "Tinted Moisturizer with SPF 30" (skincare? makeup? sun protection?) reveals exactly where your taxonomy needs the most attention.
Name categories using shopper language, not brand language. If your shoppers say "workout clothes" but your brand says "performance apparel," go with what shoppers say. You can reinforce brand terminology elsewhere.

Common Mistakes to Avoid

Using supplier or warehouse categories as your starting point. These exist for inventory management, not for helping shoppers find what they want. Always start with shopper mental models from a card sort, not operational convenience.
Testing too many products at once. Card sorts with more than 60-70 cards see completion rates drop sharply. Sample strategically from your catalog rather than dumping everything in.
Skipping the tree test and going straight from card sort to implementation. A card sort tells you how shoppers group, but grouping and finding are different cognitive tasks. Categories that feel intuitive to create can still be impossible to navigate.
Letting one stakeholder override the data. If your card sort shows shoppers group "Candles" under "Home Decor" but your merchandising lead insists they belong under "Gifts," the data should win. Validate the product taxonomy with evidence, not opinions.

Frequently Asked Questions

How many products should I include in my card sort?

For most e-commerce stores, 30-60 product cards hit the sweet spot. This is enough to represent your full catalog without overwhelming participants. If you have thousands of SKUs, sample proportionally---pick 5-8 products from each major category, making sure to include bestsellers, niche items, and anything that's hard to classify. Card sorts with more than 70 cards see significantly higher abandonment rates.

Can I validate a single new category without retesting my whole taxonomy?

Yes. Set up a focused card sort with products from the new category plus products from two or three adjacent existing categories. This tests whether shoppers see the new category as distinct or whether they'd fold those products into what already exists. Follow up with a tree test that inserts the new category into your current structure and measures whether findability improves or degrades for related products.

What's a good tree test success rate for e-commerce categories?

Aim for 80% or higher direct success (participants navigated to the correct category on their first path) for your top-selling product categories. For less common products, 65-70% is acceptable. Anything below 60% means the category name, placement, or depth needs reworking. Compare rates across tasks---consistently low scores on a specific category point to a structural issue, while scattered low scores suggest a naming problem.

How often should I revalidate my product taxonomy?

Revalidate whenever you add a product line that represents more than 15% of your catalog, enter a new market segment, or see category-page bounce rates climb above your baseline. For fast-growing stores adding products monthly, a lightweight card sort every 6 months keeps your taxonomy aligned with evolving shopper expectations. Stores with stable catalogs can stretch to annual validation cycles.

How to Validate an E-commerce Product Category