1. Introduction to Cdp System

1. Basic concepts

Customer-data-platform, CDP for short; Through the collection of multiple customer data (subjects and clues), so as to carry out accurate customer analysis and crowd segmentation, and then achieve efficient customer maintenance and discovery and daily marketing operations.

From the business perspective, Cdp is a node in the customer management process, while from the technical perspective, Cdp is a platform heavily biased towards data analysis.

The data structure

  • Subject data: multi-party customer (one party, three parties, offline) data collection, customer subject construction and behavior data mapping based on unique ID identification, to achieve structured model data management;
  • Behavioral clues: Collect behavioral data of customers’ various event types, such as registration, login, click, browse, purchase, etc. through SDK burying point, as the core clues for customer follow-up;

Constantly improve the customer subject data, improve the relevant image analysis, and then through the relevant behavior acquisition, accurate real-time follow up, such as browsing behavior of the new guest, old customers click and so on, all have the possibility of potential demand, after the Cdp system collected such clues, communicate to the customer to follow up quickly, accurate and efficient service.

Based on the process

  • Data collection: the client main body is the collection of multi-end (Web, APP, small program, etc.) registered users or channel expansion input, clues are mostly from the collection of buried means;
  • Customer model: based on the unique customer ID identification, build the main structure of the customer, business model, etc., collect and integrate the demand data in multiple business scenarios;
  • Data analysis: Basic analysis ability of customer data, common hierarchical refinement, label management, portrait and report analysis, so as to accurately identify customers;
  • Marketing operation: the above series of operations are to carry out fine operations for customers, so as to improve customer value and reduce maintenance time and marketing cost;

The core value of

In the context of traffic flow, the cost of acquiring customers is very high, so it is particularly important to conduct fine operation after acquiring customers to avoid mass loss. Establishing a group of customers with high loyalty is the means of operation with the lowest cost but the highest value, and CDP system is to support the realization of this strategy.

2. Compare Crm systems

There are many systems associated with the concept of customer management, such as CRM, CDP, DMP, etc., which need not be overly entangled with the concept of these systems, but only need to have an overall understanding. In most scenarios, they may be highly aggregated in a system, and only divided and controlled by permissions.

  • CDP: Focuses on the acquisition, management, refined operation and marketing of customer data to promote customers’ transaction behaviors;
  • CRM: the core of the customer transaction, the data level is relatively static, mainly lies in the transaction process management, recording, service, etc.
  • DMP: the core is a data management platform based on labels, which is partly involved and connected with CDP and based on the perspective of customer groups.

The essence of system platform division is the disassembly of business process nodes. When the business complexity is high, it is conducive to the rapid iteration and expansion of a single business system. In the early stage, there may be a management system, divided into many modules, to reduce the cost of development and operation and maintenance; There are different processing strategies in different periods, and a clear understanding of the whole process is the core ability to deal with business changes.

Second, business cycle

1. Core modules

The customer subject

Basic file information of customers, the biggest characteristic of this kind of data is relatively low frequency of change, not easy to obtain but easy to maintain, in addition to relatively perfect customer main body also includes: customer contact, system follow-up person module; This constitutes a complete client subject file.

Clues to the event

Through multiple product ends and business lines, buried data collection can be used as the core clue to follow up customers and accurately reach users’ needs, such as new customer registration, browsing and clicking, which are all demand-driven. Users’ needs can be captured through clue events to follow up sales promotion.

The client to follow up

Obtain the potential needs of customers through clues, and then quickly follow up, clarify the needs of customers, maintain the follow-up records of customers, and continue to provide accurate services. There are various ways to follow up, such as phone calls, visits, text messages, etc.

Dimensional analysis

Is several aspects for the analysis of the customer, the commonly used means of population subdivision, tabbed management, business reports, comprehensive early warning, dimension score, erosion cycle model, detailed analysis is the core of the data to identify the guest group, so that from the technical level of customer value evaluation, once under different business scenarios to follow up the corresponding key customers.

Marketing promotion

Through the analysis of customer group data and the establishment of the tagging system, accurate promotion and marketing can be carried out on customer groups, such as intelligent tagging marketing and customer acquisition based on seed groups. The more accurate the data, the lower the cost of marketing and the higher the value of feedback.

2. Customer cycle

The business essence behind Cdp platform is the identification and management of customer life cycle. Corresponding strategies and means are available at different stages, such as common customer cycle division:

  • New customers: newly registered users whose needs are not clear;
  • Ordinary customers: with specific needs, but with some uncertainty;
  • Membership: clear needs and certainty;
  • Super customers: provide specially-assigned follow-up and differentiated services;
  • Loss of customers: no effect of continuous follow-up, failed to wake up many times;

No matter what type of customer, there is a certain risk of customer loss. When customer loss occurs, analysis of the cause of customer loss should be provided from the business side, and reflection should be made from the technical side, whether the process cycle is not meticulous enough, and loss risk identification is not timely.

Focus on customer data collection and business model building, so as to clarify the customer cycle, build the existing customer base of fine operation capacity.

3. Architecture design

For the construction of any business platform, the first step is to define the business logic that needs to be solved behind it, and then disassemble the business process, implement modular management and implementation. When the basic structure is perfect, it is continuous iteration and optimization:

Customer growth

  • Data collection: mainly targeted at two aspects of data, the main body of data continues to improve, new data and clues continue to accumulate;
  • Channel management: data collection source management, not only online, but also offline, business expansion and other channels;
  • Scenario analysis: analyze the characteristics of data in different scenarios, identify high-quality acquisition environment, and give priority to data processing;
  • Quality monitoring: through the dimension of data collection, or the identification of information quality, initial filtering or cleaning;

Customer files

  • Subject data: the perfection of customer subject information, which is then used in various business links, needs to be relatively independent in structure;
  • Cue events: the cue data collected at different business nodes. Behind different cue events is demand mining;
  • Follow up management: follow up personnel allocation, follow up results feedback, based on the results of customer demand is clear, high value;

Customer analysis

  • Detailed stratification: Data stratification refers to basic capabilities, such as customer unit price, transaction times, regional division, business value, etc.
  • Tag recognition: tag system based on customer group management, portrait perspective, in customer accurate search and marketing is very key;
  • Dimension scoring: comprehensive evaluation of customer value, such as: activity, demand, purchasing power, data perfection;
  • Cycle model: based on the analysis of historical data, periodically evaluate the cycle nodes where customers are, and carry out strategic management;
  • Loss warning: silence analysis is provided for customers of different cycles or levels, and early warning information is provided to avoid loss;

The customer marketing

  • Operation activities: In the operation system, according to the characteristics of customers, provide activities with different characteristics to promote differentiated products;
  • Marketing strategy: differentiated marketing or marketing based on common characteristics of high-quality customer groups for customers under different label categories;
  • Feedback of results: No matter what marketing means are used, feedback of results is crucial to verify the optimization of marketing strategy;

Basic ability

  • Data storage: different characteristics of the data using the corresponding storage components, in the database selection to broaden the vision;
  • Search engine: highly dependent on data platform, for search engine construction is extremely important, support multidimensional data query;
  • Business connection: for example, Crm, DMP and other system connection, usually at the data level and application interaction;

The above analysis is based on the business flow of Cdp platform. Some basic system functions, such as permission control and operation log, are not described in detail. In fact, when the data volume keeps expanding, technologies such as big data-related components and rule engine will be gradually introduced to deal with them.

A lot of abilities are accumulated by finding solutions, learning, trial and error, dealing with, reflecting and summarizing in the case of problems.

Source code address

GitEE address https://gitee.com/cicadasmile Wiki, https://gitee.com/cicadasmile/butte-java-note/wikisCopy the code