What Is a Data Management Platform (DMP) and How Does It Work?

Around 1,683,862,863 people use the Internet every day. In the eyes of brands and online advertisers, that's 1.6 billion potential opportunities to get ads in front of a target audience.

The fact that the average person spends 6.5 hours a day browsing websites, watching videos, and chatting with friends on social media has created an almost infinite amount of user data. For online advertisers, this data is the equivalent of gold (or oil?) when it comes to building advertising campaigns — it enables targeting based on behaviour, interests, location, and much more.

To collect, analyze, and act on this enormous volume of user data within the interconnected online advertising landscape, advertisers, agencies, and publishers rely on a piece of software known as a data management platform (DMP).

What Is a Data Management Platform (DMP)?

A data management platform (DMP) is software that collects, stores, and organizes data from a range of sources, such as websites, mobile apps, and advertising campaigns. Advertisers, agencies, and publishers use DMPs to improve ad targeting, conduct advanced analytics, look-alike modelling, and audience extension.

The types of data DMPs collect include:

First-Party Data

First-party data is information gathered directly from a user or customer. It is considered the most valuable form of data because the advertiser or publisher has a direct relationship with that user — the user has already engaged and interacted with the brand.

First-party data is typically collected from:

• Web and mobile analytics tools
• Customer relationship management (CRM) systems
• Transactional systems

Second-Party Data

Second-party data is essentially first-party data from a different company. It is far less common than first- or third-party data. The information is initially collected as first-party data and then passed on to another advertiser through a partnership agreement, at which point it becomes second-party data for the receiving party.

For example, a website that sells sporting equipment (call it All Sports) may partner with a website that promotes sporting events (Half Time). When a user visits All Sports, a cookie is created. That cookie is then shared with Half Time and used to target ads to the same user across their platform.

Third-Party Data

Third-party data has earned a complicated reputation over the years, primarily because of the privacy concerns it raises. Nevertheless, it is still regularly used by marketers to reach and target desired audiences — even though it is not considered as valuable as first- or second-party data.

Third-party data is collected from a range of different sources and sold to advertisers for use in audience targeting. For example, a publisher may add a DMP's pixel to their website, allowing the DMP to collect data about visitors. Because this data is collected by a third party (i.e., the DMP vendor), it is classified as third-party data.

How Do DMPs Work?

A DMP works by first collecting data. It does this by integrating — server-to-server or via API — with other AdTech and MarTech platforms such as demand-side platforms (DSPs), ad exchanges, supply-side platforms (SSPs), and CRM systems. It also collects data by adding a tag (a JavaScript snippet or HTML pixel) to an advertiser's or publisher's website.

From there, the platform pushes that data through a series of internal processes.

The Core Processes Inside a DMP

While each DMP is different, most carry out the processes outlined below.

Data Normalization and Enrichment

Before collected data can be used, it needs to be organized into a common format — this is data normalization. The process includes collecting IDs from cookies, removing redundant or unnecessary data, and converting the source's data schema into the DMP's own data schema.

The next step is data enrichment: improving data quality by appending additional data points such as device type, location, browser type and version, and operating system.

Data Segmentation (Classification and Taxonomies)

When first-, second-, and third-party data arrives in a DMP, it goes through a process called data segmentation. Each piece of user data is analyzed and placed into different categories — also called data taxonomies — in order to build distinct user profiles.

A simple example of a segment would be users with the attribute "country = USA."

The goals of data classification and taxonomy creation in a DMP are to:

• Organize data into groups based on similarities and relationships between entries
• Create a hierarchy of the data
• Make it easy to search for and use individual entries and groups (e.g., for audience creation)

Profile Merging

Profile merging converts all profiles that share a common identifier — such as an email address or cookie ID — into a single consolidated profile. The goal is to eliminate duplicate profiles (profiles containing the same IDs) and duplicate identifiers within a given profile.

Audience Creation

An audience is a group of user profiles that share common user identifiers. For example, an advertiser might create an audience in their DMP called "Android users in the USA" — that audience would contain profiles with attributes such as "device = android" and "country = USA."

Creating audiences is by far the most important function of a DMP and is the foundation for the next step: data activation.

Data Activation

Data activation means putting audience segments to work. These segments can be deployed across a wide range of use cases.

The Main Use Cases of a DMP

Improving Targeting for Online Media Campaigns

One of the primary use cases for a DMP is helping advertisers improve the performance of their online media campaigns. This works through integration with a DSP: the DSP and DMP sync cookies together, allowing both platforms to identify users across different websites and mobile apps.

Advertisers using DSPs to purchase impressions can apply their DMP-built audiences for targeting. The data held in a DMP improves campaign targeting because it contains data that is not passed in the bid request — such as behavioural data.

It is worth noting a common point of confusion: DSPs and DMPs are often conflated. In most cases, an advertiser can create simple audiences for targeting and retargeting within the DSP itself, but only a DMP lets an advertiser build complete user profiles by connecting data from multiple sources.

Audience Extension

Audience extension is a process that allows advertisers to reach a publisher's audience across many different websites. For publishers, it is a way to generate more revenue from their audience without having to increase inventory (available ad space). For advertisers, it means reaching the same audience across a broader range of sites.

DMPs allow publishers to create audiences that can be synced with SSPs and DSPs, enabling advertisers to target those audiences via their DSP.

Onsite Personalization and Content and Product Recommendations

Showing the right product to a visitor on an e-commerce store can mean the difference between a conversion and a missed opportunity. By using a DMP, website owners can create profiles containing information about what products visitors have purchased in the past and what content they've consumed — enabling more targeted recommendations and increasing the likelihood of purchases.

These are just a few of the many use cases of a DMP. Others include advanced analytics, look-alike modelling, deterministic and probabilistic matching, and cross-device attribution.

Why DMPs Matter

Everyone within the digital advertising industry understands that data is the lifeblood of any marketing campaign. Knowing whom to target, when to target them, and what to show them is the foundation of effective advertising.

The data collected and organized by a DMP enables advertisers to get their message in front of the right audience, increasing the ad's effectiveness and reducing wasted spend. Without a DMP, advertisers frequently target the wrong audience, which erodes campaign performance.

DMPs Are Not Only for Advertisers

DMPs are typically associated with advertisers (or agencies representing a brand) and their goals of audience targeting and campaign optimization. However, publishers benefit from them as well.

By connecting a DMP to their supply-side platform, a publisher gains a richer understanding of their users. This insight is directly useful when selling inventory — whether through premium direct deals or via ad networks and ad exchanges. The result is the ability to increase inventory pricing and offer advertisers a better-matched audience.

For example, a publisher running a travel website who knows that most of their users are female, between the ages of 30–35, and based in the New York area can confidently sell inventory to a brand advertising beauty products. The ad reaches the right audience, and there is a reasonable probability that the users will engage with it.

The Future of DMPs

Over the past decade, DMPs have become a key component of online advertising and marketing infrastructure. Like many AdTech platforms, however, they face significant headwinds.

Privacy and data protection laws such as the GDPR require advertisers and publishers to obtain user consent before collecting data, which has materially reduced the availability of third-party data. Browser-level privacy features — including Apple's Intelligent Tracking Prevention and Firefox's Enhanced Tracking Prevention — block third-party cookies from being created, cutting off a major channel through which DMPs previously collected third-party data from websites.

The direction of travel is clear: the future of DMPs, and of AdTech more broadly, will be defined by the collection and utilization of first-party data. Platforms that can help organizations activate their own direct-relationship data will remain relevant; those dependent on third-party data pipelines face a far harder road.