Categories: Programmatic Advertising|By |3.1 min read|


When advertisers and publishers first begin their research into programmatic marketing, the terms Data Management Platform (DMP) and Demand Side Platform (DSP) tend to pop up in the context of media buying via online markets. But what exactly are these two concepts, and why do they matter? Should one be utilized over the other?

DMPs and DSPs are cloud-based programmatic advertising software applications used for a couple of different purposes, namely data collection and automation of ad space purchases. Even though they are often used together, these two concepts are actually quite different. Read on to learn more about what DMPs and DSPs are and how they differ.

What is a DSP?

A Demand Side Platform is simply the bidding system that bridges the gap between ad buyers and supply-side platforms with one simple individual interface. Based on Programmatic or Real Time Bidding, DSPs essentially represent the “middleman” between advertisers and publishers. DSPs provide the repository needed to buy and sell ad inventory online. Effectively, they are the core media trading platform for programmatic media buyers. Back in the day, advertisers and publishers had to do the buy-and-sell dance manually. DSPs make the process easier, especially in the context of programmatic ad buying.

What is a DMP?

A Data Management System is a central piece of cloud based technology used by an advertiser or their advertising agency to collect, organize, and manage substantial chunks of data from different sources. DMPs have become really popular as the buying and selling sides of programmatic advertising have grown in the last few years, namely because both advertisers want to track and target the right audiences throughout different advertising channels via analytics. The basic “pillars” of DMPs are collection, unification, organization, activation, and analytics. Basically, a DMP will unify all different kinds of data under one platform, namely data from publishers, partners, and advertising brands.

How Do the Two Differ?

A DMP is very different from a DSP. However, the two can be used together as part of a programmatic media buying ecosystem. They work side by side, for example, say you use Lotame as your DMP, then you can permission (or allow) all or certain audience segments to be exposed into your DSP. So, say you use Xandr as your DSP, then after you have allowed this, you can now activate against the segments as part of your media buying. There maybe associated costs to have these segments exposed, but that will be determined by your DMP and DSP agreements.

Now, where there might be some confusion, is that DSPs like Xandr have basic DMP functionality natively within their platform. This allows you do practically everything you’d expect to do in a DMP ie collect data and activate for campaigns. The biggest difference here is that the DSPs with basic DMP functionally lack advanced features like data cleansing or applying demographic data etc. So both DMP and DSP platforms can collect data from websites or landing pages etc, but DSPs can only really collect basic level of user data, which is limited. By using a DMP, data can then be pulled in from different sources like offline data sets, CRMs, email, media properties, etc. In addition to data collection, DMPs tend to have better export functionality than a DSP. Data collected within DSPs is generally not portable.

So, because some DSPs have Data collection features, does this mean that DMPs are not really needed? It depends. Advertisers who run media campaigns by leveraging their own 1st party data, should invest in data management platforms in order to aggregate data from virtually any source of their choosing, which is incredibly valuable when it comes to marketing.