The Data Management Platform (DMP) Q&A

6 minute read

Team BlueVenn

When it comes to marketing technology, prospective users are faced by an intimidating bowl of alphabet soup. Attempting to understand what all the confusing acronyms mean, whether you need them and how you distinguish one from the other has become yet another challenge of the already overburdened marketer.

To help you cross at least one off your list, we’re going explain the DMP – Data Management Platform – and why you may (or may not) need one yourself.

What is a DMP?

A Data Management Platform is a platform that ingests, collates and stores audience and campaign data from online, offline and mobile sources. More than simply managing the data, it makes refined audience segments usable by other platforms to place ads and customize content.

What does a DMP do?

By managing and analyzing consumer data, it enables users to build target audience segments, assess how those segments are performing and continually optimize them, to ensure campaigns keep reaching those best-performing segments. For example, which convert after the least amount of impressions or which are most likely to engage with particular ad content.

Essentially, it allows digital advertisers and publishers (the most common users of a DMP) to create highly targeted adverts, personalizing the online experience to specific audiences of people.

One of the main attractions of a DMP isn’t just being able to target the customers within your own database. It’s that it can use this information (in combination with cookies and third party data from data vendors) to find other customers who fit your best audience criteria, and suggest ad space on other websites they visit.

How does a DMP work?

The processes of a DMP can be broken down into three stages:

  1. Import and collects data – taking structured information from different systems then organizes it. A DMP aggregates your own first party data, second party data and third party behavioral data into a consolidated view.
  2. Find audiences and segments – with this data aggregated, it allows users to define their audience based on a number of characteristics and help you find them.
  3. Deploy automated marketing plans – A DMP will gather the instructions needed with who to target, with what message, and to what channel and/or device you should send it. A DMP can automate the process of personalizing websites, email and digital devices with ads, creative and personalized messaging.

Okay, but what about DSPs?

Often Demand-Side Platforms (DSPs) get muddled with DMPs. Occasionally, this is because of the existence of ‘hybrid DMPs’ that do both. For the sake of explaining things here, a DSP steps in after the DMP has created its instructions.

Put simply, a DSP is the software that advertisers use to actually buy available media (including display ads, video, mobile and search ads) impressions across a range of publishers sites. Typically, a DSP will step in and ‘talk’ to a DMP after it has done its thing.

What sort of data does a DMP contain?

A DMP can collect and unify customer data from many difference sources. This includes:

  • First party data – This is your own audience data. It is collected through website behavioral data, customer subscriptions, CRM data, and data collected through your social media channels or mobile apps. It’s the most valuable, high-quality data at your disposal.
  • Second party data – This is data that hasn’t been given to you directly, but obtained through a direct relationship with another entity – typically one where there are mutual benefits to share each other’s data sets.
  • Third party data – This is data collected by an entity that does not have a direct relationship with you and is collected from internet interactions and behaviors. This anonymous data will be aggregated by a DMP to create comprehensive audience profiles. It’s not as valuable as first party data, but can be collected at a far larger scale.
  • Offline data – Offline data, including transactional information collected in-store, or through customer service calls. This information may be collected through personal contact, but that does not mean it is non-digital. It offers hugely valuable insights that marketers can incorporate into their online experiences.

Does a DMP contain personal information?

First, we need to understand what personal data, or rather, ‘Personally Identifiable Information’ (PII) data is. Essentially, PII data contains information that can be used to distinguish or trace an individual’s identity. For example, their name, date of birth, email, phone number and home address.

Unlike a CRM, SCV or CDP (we can save those acronyms for another day), a DMP cannot legally contain personally identifiable information, as digital advertising requires specific, anonymized identifiers to maintain its privacy compliance obligations.

Although this is not an issue for behavioral tagging and purchased third-party data (which is mostly non-PII already), other data must be stripped of PII before upload.

Effectively, this means that while a Single Customer View (which are also often confused with DMPs) can use aggregated customer data to send marketing communications to a person, a DMP primarily delivers advertising to a device, with no direct relationship with a customer.

How is PII anonymized?

In the case of offline data, it needs to be stripped of its Personally Identifiable Information before a DMP can ingest it.

For use in a DMP, privacy is maintained through a process known as ‘onboarding’. This requires uploading the data to an onboarding platform, then matching it with online data with a person’s cookie and mobile ID, meaning the DMP is able to use this valuable (but now anonymous) demographic and behavioral data. Many organizations use third party data management services to enable the transfer of offline data to digital.

There are several methods to anonymize first party, second party and offline data. One is ‘hashing’ (an algorithm that will assign a new value to data that makes it impossible to link it to what was originally input). Alternatively ‘data perturbation’ and ‘generalization’ (modifying actual data values to ‘hide’ specific confidential record information) allows data to keeps its value to users, without identifying an individual.

Reliable products. Real results.

Every day, thousands of companies rely on Upland to get their jobs done simply and effectively. See how brands are putting Upland to work.

View Success Stories