Enterprise big data combat cases
One, the home appliance industry
Taking a home appliance company as an example, in addition to the familiar air conditioner, refrigerator, rice cooker, but also do smart home, there are hundreds of kinds of products. In its group structure, the IT department and HR, finance and other departments side by side to operate in the form of business units.
Currently, the home appliance and consumer electronics industry is worth "internal and external problems", overcapacity, price wars and homogenization phenomenon is serious; Internet companies involved in subverting the competition model, millet's "fan economy", LeTV's "platform + content + terminal". The "platform + content + terminal + application", the core is to operate the "user" rather than production. The company hopes to create the ultimate product and personalized service, the right product through the right channel recommended to the right customer, but in the CPC model currently only has CP matching (product channel), the lack of user panoramic view support, can not get through the "CP (customer product)" as well as "CC (customer channel)". CC (Customer Channel)" matching.
Based on the above internal and external environments and business drivers, the company hopes to make big data the hub of all business solutions. With Big Data DMP as the core of enterprise data, it makes full use of internal data sources, external data sources, and organizes enterprise data according to different domains to form a complete enterprise data asset. Then, utilize this system to serve various applications throughout the enterprise value chain.
So the problem comes, the company's data is scattered in different systems, more Internet e-commerce data scattered in the major e-commerce platforms, can not be effectively utilized, how to solve? The company's response strategy is: 1) start with external Internet data, the introduction of big data processing technology, on the one hand, to solve the external Internet e-commerce data utilization short board, on the other hand, you can test the waters of big data technology, due to the Internet data does not exist a large number of issues that require internal coordination, it is easier to quickly out of the effect; 2) construction of the DMP as a unified data management platform for the enterprise, the integration of internal and external data, the user portrait to build a panoramic view of users.
Phase I construction content: the technical implementation of the custom Spark crawler to capture daily Internet data (mainly Tmall, Jingdong, Gome, Suning, Taobao user reviews and other data), the use of the Hadoop platform for storage and semantic analysis and processing, and finally realize the "industry analysis", The Hadoop platform is used to store and process semantic analysis, and finally realize the three modules of "industry analysis", "competitor analysis" and "single product analysis".
The effect of the first phase of the construction of the home appliance company's big data system is quickly reflected in market insights, brand diagnosis, product analysis, user feedback, etc.
The second phase of the construction of the company's big data system has been completed.
The goal of the second phase of construction: to build a unified data management platform that integrates the company's internal system data, external Internet data (e.g., e-commerce data), and third-party data (e.g., third-party consumer data provided by external collaborations and Tabu, etc.).
The biggest value of the company's big data project to the enterprise is to transform the precipitated data assets into productivity. IT department, through the construction of a unified data management platform for the enterprise, integrates internal and external data for the enterprise, for the rapid support of the new application, and plays the role of agile IT; business department, through the insights of the product, brand, and the industry, assists the enterprise to make optimization and improvement on the product design, advertisement and marketing, and the optimization of the service. The business department, through product, brand and industry insights, assists enterprises in product design, advertising and marketing, service optimization, etc. to optimize and improve, and helps enterprises to carry out refined operations, accurate marketing and personalized recommendations based on user profiles, helping enterprises to create the ultimate service experience for users, and enhance customer stickiness and satisfaction; the strategy department, through market and industry analysis, helps enterprises to carry out product layout and strategic deployment.
Second, FMCG industry
Taking P&G as an example, in the cooperation with the marketing department of P&G China, we found that it is not necessary to integrate the internal and external data to do user profiling and customer insights. P&G captured all the data related to P&G evaluations on mainstream websites, used semantic analysis and modeling to grasp the shopping preferences and habits of different consumer groups, and quickly achieved customer insights using only external public data.
In addition, P&G is innovating in channel management. Using Internet user review data for community listening and monitoring user reviews related to the 50 retail stores P&G works with, the online data is used to conduct channel/shopper research and guide channel management optimization.
Implementation process:
1. Targeting microblogging, Dianping and other Internet data sources to collect millions of consumers talking about P&G shopping-related content;
2. Utilizing natural language processing technology to conduct multi-dimensional modeling of user comments, including more than 10 first-level dimensions and 50 second-level dimensions such as shopping environment, service, value, etc., to achieve a quantification;
3. Continuous monitoring of 50 retail channels such as Walmart, Watson's, and Jingdong, and the results are presented through DashBoard and periodic analysis reports.
As a result, P&G is able to correlate internal corporate data to more effectively grasp the overall situation of the KA channel, and even further grasp the key details, strengths and weaknesses of the KA channel, to guide the adjustment of the channel rating system, and to help formulate the product promotion planning.
Third, the financial industry
For consumer finance, home appliances, FMCG cases are also applicable, especially in terms of precision marketing, product recommendations. Here mainly share the application of credit risk control. Obviously, the Internet finance if the small loans are like banks to do field visits and invest a lot of manpower to analyze the judgment, the cost is very high, so there is a credit scoring model based on the bulk of the big data. The ultimate goal is also to achieve enterprise portraits and portraits of key people in the enterprise, and then use data mining and data modeling methods to build credit models. Yirendai, Sesame Credit and others are essentially this structure.
In contact with financial customers, we found that both banks and financial companies have more and more urgent needs for external data, especially data on strong external characteristics, such as records of lost credit, records after third-party authorization, and online behavior.
The above is what I have shared with you about the enterprise big data practical cases, more information can be concerned about the Global Green Ivy to share more dry goods