A product taxonomy functions at the centre of your eCommerce operations, customer search, merchandising, data analytics, and channel syndication. When it works well, customers find what they need quickly, while your teams can scale product data with confidence. When it doesn’t work, everything downstream goes wrong: From poor navigation and broken filters to inconsistent feeds. You end up with a major headache and some very frustrated internal team members.
The fact is, the majority of taxonomy problems aren’t caused by ‘exotic’ edge cases. They tend to come from a limited set of repeat mistakes which businesses make again and again – whether it’s due to time pressure or lack of clear data ownership. Having outlined those doom-laden scenarios, the positives are that these mistakes are predictable, and avoidable.
Below, we outline the five most common product taxonomy mistakes and provide you with practical fixes.
1. Designing for internal structures, not customers
This is a classic. A business builds its taxonomy around internal departments, supplier groupings, ERP logic, and legacy catalogues. These structures may feel familiar (and ‘common sense’) internally, but they do not reflect the intricacies of how customers browse, search, or compare products.
Your customers are thinking in terms of needs, desires, and outcomes, not .org charts. To them, if category names or paths feel opaque and counter-intuitive, users will get frustrated and bad mouth your brand under their breath, at best. At worst, they’ll abandon their efforts and go elsewhere.
How to avoid it
- Base your taxonomy decisions on customer behaviour: search queries, navigation paths, and usability testing
- Use clear, market-facing language rather than internal codes or supplier terminology
- Use cases will test real customer missions (“cordless drill for masonry,” “office chair under £300”) to see whether users can complete the search intuitively
Naturally, internal references still matter, but they belong in attributes and metadata, not in the customer-facing hierarchy.
2. Over-categorising or under-categorising products
Avoid extremes. Some taxonomies become deeply nested jungles of subcategories. Others collapse everything into a handful of generic buckets. Both examples create friction in the customer journey.
Over-categorisation forces your customers to follow long click paths, as well as creating fragile structures with tiny categories. If you under-categorise, it pushes too much work onto filters and search, making the browsing experience feel vague, overwhelming, and onerous.
How many is too many?
There’s no one-size-fits-all ‘right’ number, but the depth of taxonomy tends to be a bigger risk than its breadth. As a rule of thumb, the majority of eCommerce taxonomies generally try to stay within two to four levels. If you’re regularly finding you need to add a fifth or sixth level, it’s pretty certainly a sign that you’re encoding attributes (colour, material, style, finish) into categories instead of using filters.
A good test is cognitive load: if you’re using a category solely to narrow down a single characteristic, it probably belongs as an attribute. A good taxonomy should define what the product is, not every possible variation of it.
3. Mixing categories, attributes, and tags
Another frequent mistake is using categories to represent product types, attributes, use cases, campaigns, and brands all at once. This overload ends up with duplication, inconsistent reporting, and a brittle navigation experience for the customer. For instance, if “Brand,” “Material,” or “Use case” all appear as categories in one place but filter to elsewhere, neither browsing nor analytics will behave predictably. Customers expect satisfaction and certainty.
How to avoid it
- Use taxonomy strictly for product classification
- Store intrinsic characteristics (size, material, compatibility) as structured attributes
- Use tags for temporary or marketing-led groupings, not core navigation
- Enforce mandatory attributes per category, using governance enabled by a PIM solution
Clear separation will make it a much more reliable system for faceted navigation, analytics, and syndication to channels.
4. Inconsistent naming and weak governance
We all need rules. Taxonomy management is no different. Without rules even the best-designed taxonomies will start falling into disrepair. Over time, subtle inconsistencies creep in: pluralisation differences, duplicated concepts, overlapping categories, or slightly different attribute names which end up breaking search filters. Individually,
These issues might not appear dramatic, but they stealthily accumulate in the background and customers end up with far too many zero results pages, resulting in an escalating bounce rate. Why? Basically, they don’t trust in how you use your product data.
How to avoid it
- Define naming conventions for categories and attributes (plurals, punctuation, abbreviations)
- Use category templates with standard sets of attribute
- Assign clear ownership for taxonomy changes
- Run periodic and planned audits to identify potential drift[1] and duplication
Consistency doesn’t mean being pedantic about it! Far from it. It’s simply what enables you to scale the taxonomy across teams, regions, and channels.
5. Treating taxonomy as a one-off exercise
Many merchants design their taxonomies during a replatforming or PIM implementation project. Then, they ‘freeze’ it. In the meantime, product ranges evolve, customer language and intentions shift, and new channels impose differing requirements.
The result? Bolt-ons, lots of exceptions, and the sticking plaster approach of channel-specific workarounds which, piecemeal, undo the original design.
How to avoid it
- Treat taxonomy as a living, organic structure, not a launch deliverable that’s done and dusted once
- Monitor navigation data, search logs, and zero-result queries
- Review and refine your taxonomy regularly
- Manage taxonomy centrally with a PIM platform and map it to channel-specific structures where needed.
A taxonomy which is intentionally designed to evolve is a whole lot easier to manage than one that evolves by accident!
Final thoughts
Product taxonomy isn’t only a navigation challenge for customers. It’s a foundational data structure which shapes a range of performances:
- Product discoverability
- Data quality
- Reporting capacity and range
- Scalability
And beware – taxonomy failures generally stem from a (partly) understandable desire to take shortcuts: it could be internal thinking, over-hasty decision-making, or a lack of ownership and accountability.
These aren’t the only challenges. However, by avoiding these five key mistakes you move towards transforming your taxonomy from a constraint into a smooth-operating competitive advantage. That means grounding it in
- Customer behaviour
- Clear governance
- PIM-led structure
If your product taxonomy has grown organically (or painfully) we at Start with Data can help you. We have extensive experience in supporting businesses with taxonomy design, PIM-led governance, and scalable product data models which will improve findability without overengineering. Get in touch with us today and we can talk in more detail about your situation and how we can help you with a practical review of your current structure or build a high-performing, future-proof taxonomy to power your growth.