From 2012, Midea has started to invest in centralized system, comprehensive reconstruction system under IT governance, and the system delivery has been transformed from outsourcing to self-control, investing more than a billion and realizing the enterprise standard and language at the group level.
At present, the annual production and sales volume of Midea has reached scale and level of 150 billion, therefore, refined management and refined data drive become extremely important. The present data have already been the core asset of an enterprise, so how to make good use of such core asset which in turn promotes enterprise to develop more orderly and healthily, and at the same time how to determine direction and control risk for big data project have become the most concern of Midea to build up data system.
The Client's Return
The cooperation between Midea and Yonghong began in 2015. The construction of underlying data of the whole big data system of Midea is very strict, requiring a very flexible deployment, perfectly meeting user's business scenarios, ability to help business users to quickly set up scenario of overall business analysis and visualization, strong tension of products themselves, a data platform to continuously improve and optimize products with us, in quick response to our demand, and well satisfy our service request, which are just advantages of Yonghong.
-- Huang Kan, deputy general manager of Meicloud, a member company of Midea Group
The data source of Midea big data is constituted on the basis of internal data and external data (including the user data) and other dimensions, and as to application program from bottom to top layer as well as construction of drive mode, the overall business operation of Midea is driven by data; the joint task force of inter-business unit of Midea communicates issues relating to big data through quarterly or monthly meetings and makes targeted improvements.
The following is the specific scenario that the application data analysis of Midea improves the business:
1. Understand user needs and create hot style products.
For response of the market to an electric oven after launched by Midea failed to compete with its competitors, after analysis of big data by tracking data, the team has found some problems with the products: compared to similar products in the domestic market, the product size was too big beyond general demand of domestic users; the bench-marking analysis showed the conclusion that domestic users have paid more attention to function and aesthetics, and some special functions were important decision factors for customers to make commodity choices; Midea has carried out demand evaluation to users in market to confirm major function of the product from the user's point of view, in addition to targeted operation and promotion. After three months of data analysis and targeted improvement, this product returned to market first.
Figure 1: big data analysis benefit -- optimize products, close to users, and improve profit.
2. Improve product details and decrease complaint rate by 40%
After Midea put a household product into the market, its e-commerce platform captured user evaluation data for analysis. There were a large number of users complaining about the product unavailable with accessories although its quality was OK. After repeatedly analyzing scenarios of product design and usage, Midea has found out the reason of potential problems was product accessories packaging and using habit of users and improved its package design to relaunch to the market, since then, by tracing and analyzing daily data, Midea has found that complaint rate for this product decreased by about 40%.
3. Users’ portraits and precision marketing
At present, Midea has nearly 150 million cellphone-identified users progressively increasing at least 120,000 data every day, and the big data platform helps users to realize labeling all information such as purchase records, purchase channels, territory, preference of users and so on, so that one user data record containing nearly 600 labels and multistage tag attributes contributes to complete users’ portraits in 360 degrees, and on the basis of those name tags, the success rate for precise marketing or product recommendation will be higher. At the same time, the recommendation model of Midea users’ portraits will help shopping guide after-sales service, etc., to understand and locate users in three-dimensional space from online to offline, therefore what users needed before, what they are concerning about now, what they may need in the future and what kind of service they need Midea to provide have very strongly instructional guidance for products introduction and services.
Figure 2: agile development, Yonghong assists Midea to realize unified control of collectivized data application