Data management

Data management

Data is a gold mine and needs to be refined

Omni-channel identification and automatic merger

Nowadays, the same customer is often in contact with multiple different channels, which reflects different identity IDs. For example, general identity information such as membership numbers, mobile phones and email addresses, but also identities on various platforms, such as Taobao nock, Jingdong pin, and WeChat openID. The first step of data integration is to automatically identify and manage the unique identities of these channels, and use cross-channel IDs (mobile phones, membership numbers, etc.) to merge omni-channel data to form omni-channel customer profiles.

Customer life cycle model

Each customer's relationship with the brand will go through a complete life cycle, from cognition, understanding, interest, purchasing and using products and services, to loyalty or loss. Brands need to establish a customer life cycle model that suits their own conditions, and based on insights into each customer, determine which stage of the cycle belongs to.

Progressive profile

As the contact between customers and the brand gradually increases, the brand has a more complete and in-depth understanding of each customer. This process of understanding is not accomplished overnight, but gradually achieved over time. This process is called "progressive profile". Brands need to have a data platform that can carry out automated portraits, accumulate bits and pieces of customer data from various channels at all times, and ultimately each customer will become clearer with the future.

Label system

Each customer behavior will generate a lot of raw data, but most of the raw data may not be suitable for direct insights to guide further marketing actions. A better way is to extract "features" from the raw data, which is the label system. The construction of the label system is also an operational process. A basic label library can be established at the initial stage, and labels can be added and adjusted according to actual business needs and data quality during the marketing process.

Data mining in line with industry characteristics

After the initial collection, sorting and merging of data, further mining is needed to complete the process from "data to information". Different industries often have different data insight requirements, so algorithms and models that meet the characteristics of the industry are required to complete this step. For example, the retail industry needs to calculate customers’ brand preferences, style preferences, and RFM characteristics; the B2B industry needs to calculate potential customers’ intention scores.

Combination of first-party data and second-party and third-party data

In addition to the first-party data collected through its own channels and touchpoints, brands should fully consider the integration of second-party data (such as data related to brand flagship stores on Tmall and JD platforms). In addition, certain platform development provides limited use of third-party data, often requiring companies to upload part of the first-party data for cross-matching to generate new data insights and data value.

Apply for a trial of DM Hub now,
Experience "super growth"

Apply for trial