Is There Life in the Old Dog? What a Data Management Platform Is and How to Build One

Like matter, ether, or bacteria, data is literally everywhere. In the ad tech world, it comes from the user’s purchasing intentions, bios, and preferences and is simply the foundation upon which the advertising builds.

To use this data effectively and not be overwhelmed by its amount, marketers have their own set of tools – one of the oldest and most powerful being a data management platform (DMP).

Brace yourself; we are making another step towards becoming ad tech gurus. Today, we find out what a data management platform is, how it works, and why/if you should use it in your ad tech stack.

What Is a Data Management Platform?

In essence, a data management platform (DMP) is a technology used for gathering and managing data. DMPs store all kinds of first-, second-, and third-party data, which helps to build anonymized user profiles and share them with other parts of the ad tech stack.

First-party vs. Second-party vs. Third-Party Data

Just before we move on, it’s important to briefly remind the newcomers what the difference between data types is. This will matter a lot later on in the text.

First-party data is data collected by you as a company. Suppose someone responds to your letter, scrolls through your website, or orders your services. First-party data has stricter privacy protocols and is more exclusive; therefore, it costs more.

Second-party data is the data that you as an advertiser, share with a partner company. In a general sense, the value of this data is lower since you can’t check how it was obtained. But truly, this all depends on your partner.

Third-party data is a more accessible type of data you don’t collect directly, like cookies. It is generally worse than first-party data, as you can’t check its credibility. Plus, third-party data can’t give you a full user profile and has its privacy risks. That’s why it will probably disappear in the non-distant future.

Why Use DMP?

Let’s clear this out at the beginning: DMPs are not as vital to audience segmentation as, let’s say, DSP/SSPs are to programmatic. But this solution opens some interesting possibilities, especially if you pair it with big data and AI.

Data management platforms are used by publishers, advertisers, and sometimes even scientific labs and government offices. Today, we’ll omit the latter, but we’ll look through DMP’s main use cases in digital marketing.

Use Case #1. Buy-side platform
Primarily, data management platforms serve as a media buying tool for advertisers. By separating the audience into segments and providing the marketers with insights, the software helps to drive down user acquisition costs.

Essentially, DMP serves as the core upon which advertisers build their targeting and personalized ad offers. Most of the data, in this case, comes from second and third-party sources like cookies.

Yes, third-party cookies are slowly becoming a thing of the past, but DMP is a more secure software than you’d think. You see, data management platforms can’t collect personally identifiable information (PII). We’ll explain how that works for advertising in a bit.

Use Case # 2. Sell-side platform

Despite the common misconception, publishers can also make use of a data management platform; however, they rarely do (so we’ll mostly focus on the first use case in this piece).

You see, instead of having to rely on third-party sources, publishers can utilize their own (first-party) data to drive revenue. How exactly? First-party data is more exclusive and, therefore, more expensive.

In this case, publishers store higher-valued first-party data with their DMPs and attract more advertisers with opportunities for unique targeting. Moreover, such a solution improves user experience with relevant ads, meaning better CTRs.

What Are the Benefits of a DMP?

Now that you understand how the data management platform could be used, let’s briefly sum up its benefits.

As for the advertisers, a DMP allows them to:

• Store and process large data sets;
• Segment users by audiences;
• Better target these audiences with more relevant ads;

A data management platform allows publishers to:

• Store first-party data like demographics, behavior, etc.
• Tie this data to randomized cookie IDs or hashed log-in usernames (almost like user profiling);
• Build audience segments and reference them when ad request happens;

As you can see, the functionality of data management platforms reaches beyond just “store data.” DMPs have benefits for both core players of the ad game; now, let’s figure out how they work.

How Does a Data Management Platform Work?

DMP is an optional component of a programmatic tech stack. Its implementation adds a few extra steps before it comes down to an open auction that we described in the programmatic article.

From the advertiser’s point of view, the process goes the following way:

Step #1. Let’s say the advertiser wants to target males aged 20-25, located in Kyiv, who were recently looking for a toaster. They set targeting criteria in their DSP, and the process starts.

Step #2. Before coming to an ad exchange, the DSP first comes to the DMP for data.

Step #3. The DMP looks through its data sets and searches for users who match the criteria.

Step #4. Upon matching, the DMP forms a Device ID or a Unique Advertising ID (a special ID code that contains the advertiser’s request about a user).

Step #5. The requested ID goes to an ad exchange, and the programmatic auction happens. The user sees an ad placement.

Important: As you might’ve concluded from step #4, the data management platform contains data about the device, not the user. That’s why it’s considered to be (relatively) anonymous.

In this case, what types of third-party data does the DMP collect? Usually, the Device IDs passed to the ad exchanges contain:

• Third-party cookies (mobile & desktop);
• AdID (Android devices);
• IDFA (Apple devices).

“Hold up, a piece of software designed to store & analyze data for marketing purposes? Isn’t that a CDP as well?” – asked the voice of an ad tech enthusiast in my head.

To not get him confused and shed more light on the peculiarities of data management platforms, let’s dive into the differences.

What’s the Difference between a Data Management Platform and a Customer Data Platform?

As mentioned above, both platforms serve as data warehouses, but their purposes and functions differ.

Difference # 1. Data

In most cases DMPs store third-party data; likewise, CDPs usually store first-party data.

This difference is fundamental. With DMPs, advertisers get anonymized information on the user’s device. Third-party cookies can’t cover cross-device information like multiple email addresses or detailed behaviors.

On the other hand, CDPs unify scattered cross-platform data into a single user identifier. So the abstract male Kyiv citizen aged 20-25 looking for a toaster turns into a specific person with an email address, transactional, demographic, and engagement data.

With such a vast list of parameters, an advertiser can differentiate frequent/loyal/win-back customers and be more efficient with their ads. For example, instead of sending five prospecting emails to the already engaged audience, the advertiser sends one with the right offer for the right customer.

Difference # 2. Expiry Date

The other significant point in the “data management platform vs. customer data platform” debate is how long the gathered data is stored.

CDP vendors build their platforms to retain data for longer periods of time. And that makes total sense if you think about their purpose. The more data the advertiser has on the user – the more detailed the user portrait it gets.

Better user profile = more relevant marketing = higher CTRs = higher incomes. DMPs work differently. They store data for 90 days tops. This, however, doesn’t make them better or worse, just different.

You see, if you’re an advertiser who sells toasters, you don’t need people who intended to buy one a year ago. That’s where DMPs come in handy. Third-party data is cheaper and easier to obtain, so sometimes, it’s more cost-effective to use a DMP.

Difference # 3. Pricing

Obviously enough, DMP vs. CDP is a matter of pricing.

An average monthly subscription to a CDP starts from 20 to 60k USD, and we don’t count maintenance costs the setup. At the same time, an average DMP subscription will cost you from 1 to 6k USD.

Can CDPs Work with DMP?

So yes, those really ARE different platforms, each working better for a specific use case. But it doesn’t mean that these two can not co-exist in a single solution.

If you’re big enough, you can make the two work together with features from both worlds, like using a CDP for customer profiles already inside your brand and a DMP for the audiences outside of it.

How Do You Build a Data Management Platform?

We don’t know if DMPs will still be relevant for advertisers once Google deals a finishing blow to third-party cookies. However, we know that many enterprise marketers still want to leverage its benefits, whether with a cloud-based SAAS environment or a more traditional on-premise infrastructure.

We’ve already mentioned how much vendors charge for a monthly DMP subscription on average, but what if you wanted to build one yourself? There’s no way this article can cover every coding aspect required; still, we’ll glide through some basic features and development aspects.

*Warning: a lot of complex terms and processes ahead. If you don’t care about tech – skip to the conclusions.

So, what are we aiming for? Building a competitive platform here is a matter of how savvy your tech team is rather than the amount of invested money. To match the industry standards, your DMP needs:

Data Collection

Naturally, a data management platform HAS to collect data in real-time. This operation is called “batch data processing,” and it’s a great way to analyze large amounts of data at once.

You’ll also need to create proper mechanisms for effective data ingestion (obtaining and importing data for usage & storage). To accomplish this. The development team has to create SDKs for mobile apps, code tags for web properties, and enable batch imports at scale (for offline data).

Audience Building & Management

The second vital feature of a data management platform is the ability to organize and analyze the audience based on all the data ingested. Where do you take data? For this, you’ll need to partner up with third-party data providers.

To make use of the integrated data, backend devs create data models, and frontend devs create visualizations. The combined efforts result in a convenient analytics feature.

Ecosystem Integrations

To make your DMP live and prosper, you’ll require a lot of collaboration with ad tech providers of ad servers, DSPs, SSPs, and other ad tech solutions we frequently talk about in this blog. To make it work, the vendors of this software will provide you with their APIs.

Your backend developers will work on integrating these APIs, while your frontend specialists will create an appropriate UI that shows shared audiences.

Data Security

Hiring a few experts in data security is essential. You’ll require someone who understands how to make your product comply with regulations and not lose profit while doing so.

This will be especially important when Google finally removes third-party cookie support. It’s easy to predict dozens of changes in data security protocols once this happens.

What Are the Examples of Data Management Platforms?

Okay, you’ve looked through the vast list of features covered and considered that building a DMP is not your option. Well, the market is full of companies-vendors with their own advanced products.

Some examples of DMPs you may have heard of include Google Marketing Platform, SAS Data Management, Oracle BlueKai DMP, and Adobe DMP.

Google Marketing Platform

Google Marketing Platform is a major player in the market launched by Google in 2018. Do not confuse with Google Ads and Google Ad Manager.

SAS Data Management

As you’ve probably guessed by the name, SAS Data Management offers a cloud-based solution to data federation, monitoring, and ELT & ELT support.

Oracle BlueKai DMP

BlueKai has been a pioneer in the market, and since Oracle bought the company, its product has evolved. It uses Oracle ID Graph to cover the need for cross-channel marketing.

Adobe DMP (formerly)

This one’s a peculiar example since Adobe’s advertising solution used to be a DMP and turned into a CDP to provide its users with cross-channel activation and a bunch of new features.

Summary. Should You Use a Data Management Platform?

Answering the question from the title, we think yes, there’s still sense in using a DMP, and there are still players in the market who could benefit from using it in the long run. Let’s briefly sum up the key points of this article:

• A DMP can be used by both publishers and advertisers, but it is mostly a tool for gathering third-party data
• A DMP contains device data, not user profile data.
• A DMP is less complex and versatile than a CDP, but it’s cheaper and still has its use cases
• You can create your own DMP and benefit from an in-house platform or you can subscribe to one. The market has plenty of solutions

Once again, we can’t predict the future of the platform; however, no matter if you buy or subscribe to a DMP solution, you’ll still need an ad tech stack to make use of it.

For example, Epom White-Label DSP has top-notch analytics on the performance of your ad campaigns. The reporting and data analytics features are made to provide you with detailed insights into what works and what doesn’t.