In a thriving global business, with a multitude of languages, it’s a relief to be able to turn to a thesaurus or dictionary to locate the correct word or phrase and break down any language barriers that face us.
When it comes to technology, however, it’s a different story. There is no one single solution…no universally agreed set of principles to determine the best technology to solve a client challenge.
Deep levels of technological expertise, knowledge and understanding on both the side of the client and the consultants will determine what course of action to take and what solution is appropriate.
“Should we be using a CDP or a DMP…or both” is a question that arises often. Whilst they share similar attributes (both are based on managing customer data and designed to provide consumer insights) they each have their merits.
The purpose of this article is to help standardise and differentiate between CDPs and DMPs from our experience as an expert consultancy in both technologies.
Platform
In this aspect the DMPs are more rigid, since the client does not have the ability to choose the cloud, hardware or system. The platform is built, and its ownership is relegated to a license. On the contrary, the CDPs tend to have a more flexible approach, providing various options regarding the storage location, cloud choice or the need for any ad hoc development effort from the organisation.
Data Ingestion
Although both technologies offer real-time or batch data ingestion capabilities, the DMP usually has a more user-friendly approach, providing simpler methods such as JavaScript libraries or SDKs. The CDPs can feature them too, but it is not uncommon to see that they only offer the API connectivity, and the library implementation and maintenance relies on the technical skills of the user. It should also be mentioned that DMPs usually have a data purchase marketplace that gives direct access to second and third party data.
Data Modelling
This is one of the DMP's weakest spots, generally this technology allows creation of Boolean user attributes, that is, either you have it or you don't. There are exceptions in which you can create different types, numeric attributes, dates, arrays, etc. but there are not many. The CDPs are more focused on technical users and therefore give visibility to how and where the data is stored. This allows for several attributes and a very important functionality - being able to create entities of different types. This means that, unlike a DMP, which only allows for user information upload, a CDP is capable of incorporating other types of data to then be mixed and matched in many different ways. For example, you could upload Open Data information to have data on traffic, pollution or the stock market.
Data Storage
Since the DMP is a purely activation-oriented tool, it does not require a long data storage period, which is why the data stored in the platform is given an expiration date, typically within 120 days. On the other hand, given that the CDP is a technology with a greater capacity for analysis, it usually behaves like a Data Lake and the data can be stored for a vast period of time, subject to data retention policies.
Single Customer View
Building a unified view of customer data is key when talking about technologies that incorporate multiple data sources. The DMPs have functionalities that allow you to join a cookie with a client ID (a hashed email, CRM ID, etc.), or even give flexibility regarding how this union is made to manage devices used by more than one user. The CDPs take this capacity to the next level and have specific modules to generate the Single Customer View that offer the possibility of incorporating external Identity Graphs, matching based on personal data (names, physical addresses, etc.) or applying proprietary models.
Data Capabilities
This is where the biggest difference between the two technologies lie. While the DMP allows generation of segments based on Boolean rules, applies pre-configured look-alike type models or basic reporting, the CDP offers capabilities focused on Data Scientist and Data Engineer profiles, offering a work environment where you can process and clean the data, as well as create and execute proprietary models and perform advanced analysis and reporting.
External Connections
This is where the use of DMPs stand out and where we see the complementarity of both technologies. Due to the DMPs modus operandi and their longer presence in the industry, the match rate with media buying/selling systems is higher than for a CDP. This becomes a key differentiator since media spend is one of the largest budget items for a marketing department, and improving the efficiency is easily translatable into business wins. Regarding this, it should be mentioned that the DMPs open a line of business that is not directly available in the CDPs, and it is the monetisation of the audiences (publisher approach).
Our Vision
If we focus exclusively on the aforementioned capacities of both technologies, it seems that the CDP is the most solid option in comparison to the DMP. Nevertheless, we have to introduce a new variable no less important than the previous ones - the ability and maturity that are crucial when leveraging such technologies. Following the "start small and grow" philosophy, we believe that there are at least two phases of maturity where each of the technologies would be more prominent and better utilised:
-
Phase 1: for organisations that do not yet have the possibility of working with a unified user profile or are starting to develop multichannel initiatives, a DMP tool is the perfect catalyst to start generating the Single Customer View and activating it in an orchestrated way through several channels;
-
Phase 2: if the company already began to generate work dynamics without silos or took advantage of the segmentation capabilities offered by a DMP, it stands in a position to start introducing a CDP on the scene.
This does not mean that one technology replaces the other; they are in fact complementary. The DMP data ingestion exercise and the connection with activation tools can be successfully reused, since the CDP can ingest data from the DMP and inherit the generated audiences. Moreover, we must bear in mind that DMP tools have a longer presence in the media field and that the cookie matching tables play a key role - this is why the conversation DMP+DSP currently has a higher match rate than what is offered by a CDP by default. Also, the inclusion of computing and modelling capabilities via the CDP is a plus of intelligence that the DMP will be able to activate through more refined and profitable use cases.
Most organisations are still in Phase 1, which is why the DMP technology is so relevant. With the correct use of the DMP platform, which helps companies to give value and catalyse the data when optimising its marketing initiatives, they will be preparing a solid basis for an exponential leap in its maturity level and future exploitation of the CDP platform.
Authored by:
