A DMP (Data Management Platform) can provide audience targeting for advertising campaigns through demographic labeling and establish user profiles based on campaign data. It manages these labels and facilitates retargeting, thereby helping advertisers or agencies review and optimize their advertising strategies more efficiently.

Classification of DMPs

Based on the ownership of the DMP platform, DMPs are categorized into first-party, second-party, and third-party DMPs.

● First-Party DMP: Refers to an internal DMP built by large advertisers themselves or with the help of external technology providers. It is used for analyzing and managing user data, providing decision support and user data support for marketing processes and is widely used in industries such as e-commerce, gaming, and travel.

● Second-Party DMP: Refers to a DMP built by demand-side service providers (usually DSPs) to assist advertisers in better campaign deployment, enhancing effectiveness while increasing the volume of placements, indirectly boosting the advertiser’s spending on the demand-side platform.

● Third-Party DMP: Refers to a DMP primarily engaged in data transactions, offering services such as data exchange and sales to demand-side entities. It typically requires integration with DSPs before being applied to advertising campaigns. If it involves PC data, a cookie mapping process is also necessary between the DSP and DMP.

Classification and Enumeration of Data Suppliers

There are numerous third-party data providers in the market. Here are some of the larger ones:

● BAT: Alibaba’s e-commerce data, Tencent’s social data, and Baidu’s search data. Generally, these data sources have relatively high barriers to access.

● Companies with a wealth of valuable offline data: It’s important to highlight offline data, as we live in the real world where online behavior may not truly reflect our intentions. For instance, a user browsing cars online may not necessarily intend to purchase, but a visit to a car dealership suggests a high likelihood of intent. Typical representatives include companies like Zhanhui Zongying, which holds real user activity data from airports, high-speed railways, and the automotive industry chain, and UnionPay Smart, which has offline transaction data from POS machines. Recently, these data-rich companies have been activating data monetization models, offering advertisers superior programmatic advertising solutions through self-built DSPs and unique DMPs.

● Third-party monitoring companies also hold a large amount of advertising campaign data due to the nature of their business. Typical representatives include companies like MiaoZhen and AdMaster; mobile representatives include TalkingData and Umeng.

● Media companies also provide demographic data services (such as gender, age,andinterests), but the coverage of a single media source is inherently limited.

● Traditional CRM technology service companies are also present, but integrating CRM data with online data has always been a challenging issue.

● DSP companies also possess some data, primarily sourced from advertising traffic. Advertising exchanges provide user and media information data such as the current media, position, and IP of the user’s advertisement to help DSPs make better bidding decisions based on user behavior. As a result, DSPs have accumulated a wealth of data based on this advertising traffic and their past advertising performance data. However, since this data is carried within the advertising traffic and much of the RTB traffic is “remnant traffic,” it has a certain degree of fragmentation and may not reflect the entirety of a user’s online behavior. Especially on mobile ADX, unlike PCs, it cannot provide the URL of each ad content page, only the App in which the user’s ad is displayed, and the latitude and longitude obtained is merely the user’s offline location when the App is opened to display the ad, which may not reflect the user’s entire movement trajectory. Therefore, such fragmented datamakes itdifficult to analyze and label user behavior as continuously and precisely as on the PC side.

Basic Functionality and Core Process of Data Sample Learning

The most fundamental function of a DMP is to collect various online and offline data through different means and channels. The types of data can be diverse, not limited to advertising campaign data but also including CRM, surveys, third-party, and other data sources. Advertisers focus on different aspects of data from different sources:

● ForFirst-Party DMPdata, advertisers are more concerned with the analytical capabilities of first-party data, such as consumer behavior analysis on official websites and offline, and media attribution analysis.

● ForSecond-Party DMPdata, advertisers are more concerned with the application of second-party data in advertising campaigns and the impact of media content and categorization on advertising efficiency.

● ForThird-Party DMPdata, advertisers focus more on the efficient output methods and connectivity and effectiveness of the data.

The basic functionality of a DMP mainly revolves around various stages such as data collection, cleansing, integration, management, analysis, and application.

Data collection, cleansing, integration, and management focus on aspects such as timeliness, accuracy, reliability, stability, scalability, and automation of data processing.

Data analysis and application focus on the mining of “people” and “patterns” within the data. Creating audience profiles, classifying and tagging data, and providing guidance for marketing and decision-making are key.