Dry goods: lightweight construction of user labeling system in the automotive industry

Recently, the digital transformation of car companies has been in full swing, and major car companies have invested heavily in the construction or reconstruction of CDP (customer life cycle data platform). First of all, open up the delivery system, lead management platform, sales follow-up assistant, customer order system, after-sales service system, membership system, mall system and other related systems throughout the customer's entire life cycle. Then build a customer labeling system based on the global data system (the essence of the label is the commercialization of data, that is, to realize the commercial realization of data, and the data must be transformed into a label that can help the business improve to have value, otherwise it is a data burden). Theoretically, this approach is logical, but from the actual construction experience of car companies, it is difficult:

01

First of all, the contact system that runs through the customer life cycle belongs to multiple departments of car companies, involving sales department, after-sales department, marketing department, etc., cross-departmental communication costs are high, the implementation cycle is long; some businesses across multiple departments are even difficult to find actual business docking people, and the current organizational structure of car companies has become a mountain of digital transformation. Judging from the many car company projects we have actually experienced, this is almost a common phenomenon.

02

Secondly, the construction time span of the car enterprise system is large, involving a large amount of historical data, and the various systems lack unified data standards. For example, most enterprises do not have a customer lifecycle OneID, and the customer master data system is only applied in some business systems; in addition, the data standards are inconsistent, such as the model name is not uniform in the sales and after-sales systems. All of this brings a huge workload to the data opening work.

03

Again, the construction of different suppliers of each system has great differences in the technology used, and the data synchronization ability of each system is also very different. The system built in the early stage is difficult to achieve in real-time data synchronization, which has become a bottleneck for the interconnection of various systems. Differences in technology have greatly increased the difficulty of getting through the system.

Although many car companies have implemented early data lake projects, now CDP projects. At present, the author does not see a car company can solve the customer data in various systems to achieve real-time or quasi-real-time interconnection. Even so, in fact, this does not mean that the labeling system cannot be built. Back to the essence of labeling is to achieve the commercial realization of data, and the construction of the accurate label system has two core key points that are scenario-based and real-time, which can transform data into labels that help business improvement.

The core of label construction is to use it as a scenario. Cut a business scenario into a small closed-loop business scenario. Starting from the current situation of the use of labels by auto sales consultants, the most used label scenario is that the consultant uses the personal WeChat "set notes and labels" function, such as "Mr. Zhang XX model", this simple note label is an efficient way to help consultants identify customers and quickly circle customers, labels can be uncomplicated, but must be valuable. Secondly, labels, phone numbers, and descriptions are used to supplement customer information. This is the scene of following up the customer in the personal WeChat scene.

Personal WeChat solves the basic function of the label application. However, there are also obvious deficiencies, these labels are manually labeled by consultants, and the channels for obtaining information are derived from the communication of the consultants themselves, which has great limitations.

Tencent's official enterprise WeChat automotive industry version, aiming at the pain points of the automotive industry business. The automotive industry scene has been comprehensively upgraded on the basis of the native functions of personal WeChat and enterprise WeChat. In particular, the two functions of customer labeling and customer dynamic behavior labeling (dynamic radar), the general feedback of sales consultants is very practical.

The label section is divided into three parts, personal label, enterprise label and automatic label. Among them, the enterprise label, the background provides the overall configuration of the enterprise label, can directly access the main engine factory CDP has built the label system, to achieve CDP tag and enterprise WeChat tag interoperability.

If the main engine factory does not build a CDP label system, it will not affect the application of the consultant label system. The tag library function itself provides a complete set of tag templates, OEMs, 4S stores, consultants, etc. can customize and manage the tag library, and quickly build a set of label systems for practical applications.

Auto Tag provides Group Behavior Tags, Asset Interaction Tags, keyword tags. In view of these WeChat communication customer scenarios, customers are automatically labeled. This greatly improves the limitations of consultants who rely solely on manual labeling.

01

Group behavior label: 4S stores in outdoor auto shows, storefront activities, closed store sales and other marketing activities, for scanning the code into the group of customers automatically labeled.

02

Material interactive tags: According to the behavior of sending different materials, customer clicks, forwarding, etc., the corresponding tags are marked for customers. For example, the test drive invitation letter is marked with labels such as "intended customer" and "test drive". Send the corresponding material to the customer, and the customer clicks to automatically trigger the rule and automatically label. It can greatly reduce the workload of consultant labeling and improve the accuracy of labeling;

03

Keyword tags: Apply the enterprise WeChat session archiving function, the consultant and customer conversation process, automatically extract keywords, such as model entity name extraction, automatic label intention model; price negotiation session, automatic extraction of car purchase budget tag, etc.

The customer labels summarized by sales consultants from the practical experience of following up on customers have important value in themselves. For example, the reliability of information such as customer car purchase intentions, purchased models, car purchase time, and car purchase budget grasped by consultants through communication and interaction is high. These labels complement the labels that CDP mined from the data. For example, algorithm labels, even if speculating on customer gender, need a large amount of customer behavior data to support, and consultants can quickly label these labels through meeting or communication; at the same time, there is a large number of behavioral data missing in the early stage of clues, and algorithm labels face an embarrassing situation of technology but no data available.

The sales themselves are only fast and not broken, the sooner the portrait is clear, the more helpful it is for sales strategies, quotation schemes, and transaction customers. Therefore, the construction of our label system must take the "mass route", give full play to the advantages of the dealer system of many sales consultants, with the help of the consultant's manual labeling of the label, on the one hand, enrich the data resources lacking in early customer information; on the other hand, the consultant's labeling is the natural machine learning algorithm label training sample source, laying the foundation for the subsequent large-scale automatic algorithm labeling.

From the perspective of the real-time nature of the label, the customer's own needs and cognition are constantly changing, so the customer's label is also constantly changing. For example, customers who have a bad impression of their own brands, one day, actually test drive their own brand models, have a new experience of their own brand screens and technology configurations, and will suddenly turn to accept their own brands. I believe that many people will have this experience.

If you speculate through algorithms from customer data over a long span of historical cycles, it is difficult to capture this change in business opportunities. Using historical data to calculate more is an average person, the more historical data, but to erase the customer's current situation changes, missed business opportunities to discover.

Therefore, real-time dynamic behavior labels generated by real-time customer behavior data are important. For example, through the customer dynamic radar, for the consultant to send the customer trajectory materials in real time, test drive invitations, model introductions, circle of friends, etc., real-time capture of customer behavior dynamics, and through timely reminders to consultants, let the consultant seize the time node of communication is very valuable to enhance customer conversion.

For example, when the customer opens a certain model introduction sent by the consultant, or browses the promotional activities of the consultant friend for a long time, Micro sends a message to notify the consultant in time. Consultants communicate with customers in a timely manner to introduce the model, or invite them to participate in promotional activities. At this time, the customer is also paying attention to the same thing, the customer just needs it, the consultant just appears, and this scenario-based communication becomes very warm, rather than the dry and dry boring customer without a goal.

This real-time customer dynamic behavior label, combined with real-time message reminders, forms a closed loop from the label to the business application, which plays the business value of the label well, and returns to the essence of the label".

To summarize, we start from the scene where the sales consultant applies labels, and build around the use of the building. For enterprises with different IT construction maturity levels such as OEMs, dealer groups, 4S stores, car dealers, and used car dealers, corresponding solutions are provided in label scenarios and real-time applications:

(1) For enterprises with perfect CDP construction: make full use of CDP's already built label system capabilities and the ability of Micro Automotive Industry Label System to communicate, and form a closed loop for label business applications with the help of Micro Direct Connection to Consumers. Give full play to the business value of labels in customer cultivation, transformation, fission and other business links.

(2) For enterprises whose CDP construction is not yet perfect: you can directly take the automotive industry version of Micro As a starting point, with the help of Micro's platform integration capabilities, run through the design, construction, and application of the closed-loop ability to support business landing. In the later stage, looking at the needs of enterprise development, you can simultaneously build your own CDP platform; you can also directly apply the powerful front-end application and background configuration of the micro automotive industry label system, which is also a solution with a very high input-output ratio of ROI.

Again, the construction of the label system must give play to the practical experience of the sales consultant group in sales practice, so that the consultant can effectively precipitate the labeling assets of the customer as the core assets of the enterprise; the labeling system that can be applied by the business is valuable.

If you want to continue to learn more about the content and communication: please pay attention to the WeChat public account: Nesch Technology