What is Taxonomy in a DMP?

There's a popular joke among software developers that goes: there are two hard things in computer science — cache invalidation and naming things.

This guide covers the latter — naming things — specifically in the context of taxonomy inside a data management platform (DMP).

What is Taxonomy?

Taxonomy is a broad concept found across many fields, from biology and education to business and data management. Even the Wikipedia page for taxonomy lists it under six distinct categories.

Despite spanning disciplines, the core definition stays consistent: taxonomy is a system for naming things and organizing them into groups based on their similarities. Originally rooted in the sciences, it's now a foundational concept in data management.

Taxonomy in Data Management

Consider how many tools and platforms a typical business uses to collect data. Each one tends to use different terminology than the next.

A straightforward example: one platform might refer to a "user" while another calls the same entity a "visitor." Another common source of confusion is the term conversion — some platforms define it as a purchase, others as a signup, and still others as something like a white paper download. As a DMP collects data from a range of online and offline sources, the need for defined, unified terminology becomes critical.

The goals of taxonomy in a data management platform are to:

• Organize data into groups based on their similarities and the relationships between them.
• Create a hierarchy across all available data.
• Make it straightforward to search for and use individual entries and groups — for example, during audience segmentation.

Creating Taxonomies in a DMP

Three principles should guide taxonomy creation:

• Relevance and uniqueness to the business — taxonomies should reflect the actual structure of the business and its data, not a generic template.
• Alignment with advertising and marketing strategies — or whatever the primary use case happens to be.
• Unambiguity — taxonomies must not contain duplicate terms or overlapping categories. Ambiguity leads to misclassification and unreliable segments.

Creating Taxonomy Rules

Taxonomy rules allow a DMP to automatically assign new data entries to the appropriate categories as they arrive, rather than requiring manual classification.

For example, an insurance company could configure a rule so that any new data entry carrying the attribute product=car-insurance is automatically placed in the Car Insurance category. This kind of automation is what makes taxonomy practical at scale.

Examples of Taxonomy in a DMP

The following examples illustrate how different types of businesses might structure their taxonomy inside a DMP.

Taxonomy for an E-commerce Site

Taxonomy matters for all types of companies, but it's especially important for e-commerce operations given the sheer volume of products involved.

Here's a basic example of how an e-commerce store might structure its taxonomy:

Taxonomy for a Car-Sales Website

Car-sales websites face a similar organizational challenge to e-commerce stores, with inventory spread across many makes, models, and conditions. Here's a basic example of how a car-sales website could structure its taxonomy:

Taxonomy for an Airline

Airlines operate across multiple business lines — not just ticket sales, but also frequent-flyer programs and membership programs. Each of these areas represents a distinct branch of a well-structured taxonomy. Here's a basic example:

Building Audience Segments from Taxonomy Categories

Once a taxonomy is in place, it becomes the building block for audience segmentation. Categories can be combined to define precise audience segments that reflect meaningful groupings of users.

Using the car-sales website as an example, here's one possible audience segment constructed from taxonomy categories:

A car-sales website could create a series of segments like this one for several purposes: selling them directly to other companies (insurance companies being the obvious fit), or pushing them to a supply-side platform to increase the value of its advertising inventory.

Taxonomy is, ultimately, what makes a DMP's data actionable. Without a coherent naming and classification system, even rich datasets become difficult to query, segment, and operationalize. Getting the structure right from the outset — relevant, unambiguous, and aligned with actual business goals — is what separates a DMP that delivers value from one that simply stores data.