Get to know Dunnu
Dunnu Group, founded in 1987, is mainly engaged in clothing, hotels, and real estate. Its headquarter is located in Haining, the leather capital of China. Zhejiang Dunnu United Industrial Co., Ltd. (hereinafter referred to as "Dunnu") is a large-scale professional clothing enterprise integrating development, design, production and sales. Dunnu has two R&D and operation centers in Shanghai and Haining. It has three major brands, DUNNU, DDU and DIDIER PARAKIAN. Its sales network covers all parts of the country, with 500 chain stores nationwide. Dunnu has a garment production base of more than 80,000 square meters and has more than 2,000 employees. In 2016, the turnover of Dunnu's clothing business was more than one billion yuan.
Big data platform twists and turns
Dunnu started to build a big data platform in 2013. At the beginning of 2014, a foreign brand system was launched. Because of the general use effect, the employees were not used to using it, and the leaders did not want to read the reports issued by the system. After one year, it was basically abandoned. Data construction is the only way for enterprise IT, and the demand is expanding again. In 2016, the construction of the big data platform was restarted. Considering the four factors of mobility, interactivity, cost, and manufacturer experience, it was decided to use FanRuan reports to build this data platform.
Summary of the construction results of big data platform
1. In 2016, the production and sales rate increased by 5%, the profit increased by 15%, and the contribution of the big data platform accounted for half
2. The monthly report is accelerated from the 12th to the 5th of each month, and can be further improved
3. Reduce the workload of 20 professional IT personnel, and big data contributes half of the role
4. The management of the enterprise is from disorder to order (goods week operation management, salary management, etc.)
5. Changes in enterprise management, flattening the organizational structure, breaking down departmental barriers, improving collaboration efficiency, and greatly improving the efficiency of business data circulation
Dunnu big data platform construction
There are two special expectations for the data platform Dunnu.
First, for small and medium-sized enterprises, although the amount of data we have to deal with is not big data, the business logic is very complex. We hope that the big data platform can analyze the data generated by these complex logic businesses and guide business adjustments.
Second, the requirements for platform intelligence are higher. The selection of the platform is not only used to produce fixed-format reports, but also to implement all the business processes and indicators of the whole company with the system, which requires higher intelligent requirements for the platform.
Based on the above two expectations, in 2016, our big data platform construction mainly planned and completed 4 modules: master data management, address book management, goods management, and BOSS interactive screen.
master data management
The core purpose of master data management is to open up various business systems and form a unified data interface specification. Before the platform was launched, there were no rules for Dunnu's basic data. As a result, related facilities such as property and points mall cannot be effectively linked with member data, member point management is a mere formality, and member data in the business system cannot play its due value. In the process of master data management, in addition to batch building with ETL tools, the maintenance platform is also built with FanRuan.
For example, we have used FanRuan to develop pages such as "Master Data Basic Addition Process" and "Dealer Information Modification Application Process", which are integrated into the OA system. Personnel from all departments can apply through the unified OA portal, and the maintenance of the master data of each system is uniformly received to the IT department. After the approval of the OA process, it will be automatically generated if the system can be connected. relevant departments and personnel. (Flowchart as above)
Master data management is the foundation of our data analysis. The two pages in the figure above realize the master data management of our membership system, order system and OA system. In addition to helping us open up different systems and realize data sharing, FanRuan master data management also effectively saves IT human resources. Rough statistics, only the master data collation module saves 6 to 7 people.
-
How to perform master data management operations?
For example, an interesting challenge was encountered in the early stage of construction: what is the "broken code" of clothing. From the perspective of the leadership, if all the sizes of the clothing stored in the company are incomplete, it will be broken; from the perspective of the warehouse clerk, if the clothing size in the warehouse is incomplete, it will be broken; from the perspective of the store salesman, customers need The size is currently out of stock in the store. Before the membership system, inventory system, order system has not completed the master data management. Then there will be goods in the warehouse, but the store can't match the goods, and it is impossible to get the corresponding goods from the warehouse in time to provide our customers. Now, after perfecting and getting through the product information and member information, you can transfer the correct clothing from the warehouse to the customer on the spot, and you can also choose to express the clothing to the customer's home according to the customer's preference. So thinking about it in reverse, the meaning of "broken code" is that Sister Ling of the business does not agree, resulting in inconsistent indicators and difficulty in going through the process. Later, we defined the broken code as that the customer cannot get it on the spot in the store, and it cannot be mailed to the customer within 3 days, or the customer refuses to mail it.
Address book platform
The second result of our data analysis project is the address book of the big data platform. Before, our OA platform already had an address book (see the picture below), and it looked pretty good, so why did it need to be rebuilt?
Because the previous address book is not used. Our investigation found that the original OA platform address book has three major problems: too many people, little information, and weak correlation. There are nearly 1,000 contacts in the company, all of which are displayed in the list, and no one wants to see it; and the personnel information in the address book only has departments, telephone numbers, and mailboxes. Due to the need to communicate with others for a certain job, it is impossible to confirm whether or not to contact this person according to the address book. ; For each person, most of the more than 1,000 contacts are useless, but every time they search, they are mixed up, and the person they are looking for cannot be found.
In order to solve these three problems, we have redeveloped the "big data platform address book" according to the needs of business departments to communicate the results. According to the organizational structure, the new address book is divided into hierarchical groups and folded. Detailed information can be drilled down at each level, and it covers a relatively complete range of personnel information, including supervisor-level personnel information and cultural information in the region, including store information, etc. . This solves the three major problems of the original address book in one fell swoop.
-
Why address the address book issue first?
In fact, this is not only a challenge for us to build a big data analysis platform, but also an unexpected gain for us. During the project demand research, we found that it was quite difficult to find specific demanders. Every time we needed to contact each other, the efficiency was relatively low. At the same time, we also received complaints from the business department that it was inconvenient for them to find someone by themselves. This problem is only discovered now, because the company has grown and the number of personnel in the system has doubled, and these problems will not be reflected in performance and indicators. That is to say, the information system of the whole company has not been able to reflect the OA system address book to find the corresponding The problem of human inefficiency. With the assistance of the FanRuan reporting tool, we evaluated the workload and found that the data was readily available. This address book platform used on mobile phones was launched in less than a day. Of course, there are some optimizations in the later stage. In general, this is the first time that business personnel have felt the efficiency of this system and have been recognized for the first time. The subsequent cooperation work has also been very active.
Goods management
The core of Dunnu goods management system is data access and real-time reporting. At present, we have realized automatic daily closing at 12:00 and 24:00 every Monday, automatic adjustment at 13:00 in the region, automatic adjustment at 9:00 on Tuesday in the whole country, acceptance of goods on Friday, and a complete set of automatic sales on weekends. Before using FanRuan, it was impossible.
Our monthly report is mainly about goods management. The new monthly report system is mainly used by middle and low-level managers. Considering everyone's needs, we use the form of a table. It is not only convenient for managers to view complete data, but also easy to connect with our WPS system. Previously, our monthly report was expected to be released on the 12th of each month, but it was actually difficult to release it on the 12th. Now we can speed up to No. 5 out, we can further improve later. This monthly report alone can reduce the workload of 20 people. But we are not really laying off 20 people, we are assigning these people to other project modules of informatization construction.
BOSS interactive screen
Dunnu's BOSS interactive screen system is mainly responsible for efficient analysis of existing business system data information, and displays the analysis results on the leader's office display screen, allowing leaders to view the financial data indicators of various management departments intuitively and conveniently , and allocate resources reasonably. The BOSS interactive screen in our office mainly analyzes and displays the three important financial indicators of directly-operated stores, franchised stores, and the actual receipts, retail sales, and payment collections of the fifth quarter on the previous business day and the current month. The boss can click on the interactive screen in the office to understand the sales of all stores, and at the same time highlight the abnormal indicators of abnormal stores. Considering that leaders are away on business and also need to keep abreast of the store's business status, we have deployed a mobile interactive kanban using the FanRuan report platform. Now the inventory and financial data of more than 500 stores in Dunnu can be viewed directly by leaders on interactive screens and mobile terminals. According to the feedback from the business department, the leaders basically no longer call for business financial data, and the leaders themselves feel convenient. At the same time, every time the business department reports, directly facing the interactive screen in the office, reporting and operating, focusing on business analysis instead of wasting on reading reports and checking data.
Rethinking the challenges in big data platform practice
1. The strategy is ambiguous and the implementation of the policy is difficult. Due to the weak foundation of data construction, the data was chaotic at the beginning, and the data of different systems did not match. The leadership needs performance, the business department needs data, and the IT department is caught in the middle and it is difficult to promote. Since the formulation of a clear strategy, the decomposition of the policy implementation has been carried out smoothly. Therefore, setting a strategy is the first priority. In 2017, Dunnu formulated a clear big data platform development strategy.
2. The basic data is incomplete or missing, resulting in low model intelligence and low value. Nowadays, enterprises generally have the problem of poor basic data quality. In the construction of big data platforms, we must be patient and do a good job in data cleaning and data quality control, and spend our time on processes and specifications. Data quality management depends on the standardization of enterprise process management.
3. Change the corporate culture, change the organizational structure, and affect the interests, rights and control of some personnel. One of the goals of Dunnu's big data platform is to improve the work efficiency of business personnel. Efficiency is improved, less manpower is required for the same work, and some procedures even disappear. It affects the interests, rights and control of some people, and the implementation of natural projects will encounter obstacles. Therefore, the impact of this part was fully considered before the project was implemented, and the opinions of these people were also paid attention to during the implementation process. The success of the first phase of the construction of the Dunnu big data platform is also the success of the opposition.