Explore the Data management platform

Most marketers have begun to realize the role of data management technology as a strategic competitive weapon for advertising marketing activities. In order to meet the goal of deepening, understanding and transforming data into this power, the DMP-data-management platform will be an excellent choice for brand success.

Like many other terms in the industry, DMP-related data management technology has grown rapidly and is still somewhat confusing to many. So, it’s easier to access if we do not go too deep into the process details that start with basic operations such as cross-data between different platforms or cross-device user identities (Cross-device). At the most condensed level, DMP’s data management process becomes the story of ‘Data in’ and ‘Data out’, helping Marketer to discover the benefits. that the DMP platform can bring an easier way.

Input data

With most marketers, ‘charms’ that occur within the management and operation of the input of a DMP platform are not attractive because of the technical nature. However, it is the starting place for all future marketing activities.

DMP input allows Marketer to identify the user. “Knowing who anonymous people are surfing the web and using the applications is” is a feature that makes DMP useful in the eyes of Marketer. It facilitates the deployment of one-to-one marketing (direct-to-consumer marketing) or understanding of customer shopping flows.

DMP not only handles cookies

The early design of DMP focused on sorting the cookie IDs according to a defined classification rule and matching them to the operating platforms. Most DMPs, from simple levels like “media DMPs” built into demand-side platforms (DSPs) to fully independent platforms with full functionality, can easily handle Cross-data-based cookies.

More specifically, the first generation DMPs are designed as a platform for collecting and distributing cookies. This means assigning cookies to a segment of your customers and passing this information to the DSP platform to perform targeting.

In particular, DMP must be capable of capturing a variety of identity data – such as the customer data subdivision in the CRM system, data from in-store (“point-of-sale” – “system”, such as a cash register at the counter), or beacon data (a wireless sensor that tracks and identifies users in a real-world space of the receiving range).

DMP also needs the ability to capture data from the device

With the “less rude” process in the past, users are identified separately through each device they use. For example, a desktop user who uses two different browsers will have two cookie IDs; when he or she watches Internet TV using the Roku television signal decoder, there will be a corresponding Roku device ID; or online with the iPhone and iPad will have two IDFAs corresponding to two Apple devices. Because each of these IDs is identified as an anonymous, independent user, even though the actual user, Marketer can turn into 5 different anonymous users.

In fact, most DMPs are created only to store and manage cookies for display advertising, not as cross-device applications. Thus the ability to ‘absorb’ different data types such as structured data (such as name, gender, etc.) or unstructured data free and unstructured data such as text, video, image, sound, …) of the DMP is quite spread. But as users spend more and more time on non-desktop devices, in addition to managing cookies, DMP also needs to be able to capture data across devices to better identify users.

Device Diagram (Embedded device graph)

Identifying “cross” devices is notoriously difficult and is usually done through two methods: Deterministic Matching and Probabilistic Matching.

Over the next few years, the term “device graph” will become commonplace in the industry. In particular, as more and more companies seek to identify users cross-device without data from walled gardens such as Google or Facebook. It can be said that ‘Data in’ is a key component of any DMP platform because if it is not possible to collect data from identification devices and not identify the right user, the one operation -to-one marketing or multi-channel models is almost impossible.