Product attributes vs categories: How to get your taxonomy right

It’s simple – if customers can’t find the products they want or need, they can’t buy them. The bridge which successful businesses build between customer intent and purchase is the taxonomy: In short, that’s how you group, label and describe your product range. However, if you don’t get the balance between your product categories and their attributes right, everything is hard to shop – customer touchpoints become pain points. Bad news for your brand.

Conversely, if you get it right, product discovery becomes practically effortless and satisfies any potential customer from the get-go. In this article, we’ll outline exactly what ‘getting it right’ involves, and how to get there.

Taxonomy matters way more than you think

Taxonomy sounds like a pretty technical term, but it has an impact in very human ways:

• Can someone land on your site and reach the right “slice” of the catalogue in no more than three or four clicks?

• Are your search filters helpful, or do they regularly return empty (‘ZERO’) results?

• Can you syndicate product information to different marketplaces without having to rebuild your structure every time?

These three questions sit on top of two cornerstones:

• Categories – The “where”
• Attributes – The “what”

Categories: Your catalogue’s answer to “Where?”

Categories are the high-level map. They define where a product sits in the overall structure.

• They are hierarchical: Home > Tools > Power tools > Drills.
• They drive navigation, breadcrumbs, and reporting.
• Inside a PIM system, they also control who can see and edit which parts of the product range.

In a well-structured catalogue, each product has one primary internal category path, even if it may appear in multiple locations on a website. Categories should be relatively stable over time, requiring, say, annual reviews rather than weekly.

As a rule, high-performing categories are:

• Customer-centric (based on how people shop, not how internally organised)
• Designed to move from broad-to-narrow, with three to four levels at most
• Free of “feature masquerading as category” (so, that means no “Red dresses” or “Waterproof jackets” as top-level buckets)

Attributes: the “what” of your products

Attributes are the facts – what describe and differentiate products within a category.

Commonplace examples:

• Drills: Voltage, Chuck size, Power source, Brand, Weight
• Dresses: Size, Colour, Length, Fit, Material, Pattern
• High-volume B2B components: Diameter, Material grade, Pressure rating, Compliance code

Remember, it’s the attributes which underpin the power for:

• Filters and faceted search
• Comparison tables and spec sheets
• Marketplace templates (Amazon, Google Shopping, Alibaba, and all)
• Future capabilities like AI recommendations and better on-site search

To function truly efficiently in a PIM, attributes should be:

• Granular and standardised – “Navy” rather than a mix of “navy blue”, “midnight navy”, or “NVY”
• Validated – usually via picklists or controlled vocabularies
• Contextual – only shown where they make sense (screen size for TVs, not toasters)

The golden rule: ‘If it can be a filter, it’s probably an attribute’

Most taxonomy problems come from turning attributes into categories.

The broken pattern

Jackets > Waterproof jackets > High-vis waterproof jackets

You now have to build separate pages for every combination of features. Add two new safety ratings and three new colours and your category tree explodes.

The better pattern

Category: Jackets
Attributes: Waterproof = Yes; High visibility = Yes; Colour = Yellow

Now a customer can see all jackets, then filter down quickly. Your merchandising team can also change or add features without having to re-engineer the tree.

As a rule of thumb:

• Use categories for “what kind of thing is this?”

• Use attributes for “what properties does it have?”

Common pitfalls that break taxonomy

It’s the few recurring mistakes which do most of the commercial harm:

1. Over-categorising attributes
   Building unmanageable trees and painful navigation. For example, a category pats like Shirts > Slim fit > White instead of “Shirts” with Fit and Colour as attributes – this will result in an unnecessary challenge for customers to negotiate.

2. Messy attribute values
   Permitting “Black”, “Blk” and “Midnight Black” as separate values. In this particular example, product filters will miss out half your range for no good reason. Using controlled vocabularies in your PIM can solve this.

3. Confused data ownership
   Letting your PIM override operational data which really belongs in the ERP (such as price and stock) while ERP simultaneously overwrites content-friendly product attributes. You’ll simply end up with constant firefighting instead of one authoritative record per data type.

4. Ignoring inheritance for variants
   Repeating the same care instructions, brand, or warranty for every size/colour variant instead of inheriting from a parent. That creates maintenance overhead and guarantees inconsistencies later.

Using PIM to design a taxonomy that works

A good PIM doesn’t magically repair an entire taxonomy, but it will give you the tools to implement, police, and monitor a better and more user-friendly design.

Practical steps you can take:

• Start with customers, not org charts
  Base your category tree on how buyers think: applications, use cases, industries or lifestyles (obviously depending on whether B2B or B2C).

• Define attribute sets per category
  For instance, drills need Voltage and Chuck size; sanding discs need Grit and Diameter. The PIM allows you to attach relevant attribute groups to each category so that time-poor users don’t face irrelevant fields.

• Enforce controlled vocabularies
  Use picklists for colours, materials, compliance codes, etc. PIM can prevent new non-standard values from creeping in.

• Robustly govern changes
  Be sure to assign owners for both category structure and attribute definitions. Changes should be deliberate and not an accidental by product of someone importing a new spreadsheet.

If you handle your taxonomy this way, it will remain stable enough for reporting and integrations, as well as being flexible enough to support new products, channels and campaigns.

Final words

If your site navigation feels unnecessarily complicated, your filters don’t behave predictably, or you’re having to provide every marketplace with a different version of your hierarchy, it’s not just a UX issue for you – It’s most likely a problem with your taxonomy.

At Start with Data we’ve worked extensively with retailers, distributors and manufacturers to redesign their product taxonomies, model attributes in PIM, and untangle category/attribute divides. Get in touch with us today and we can talk more about your taxonomy and catalogue pain points, as well as how we can help you to enhance their performance so that your teams can thrive and your customers actually enjoy the seamlessness of their shopping experience!