When manufacturing meets big data

"Minority Report," starring Tom Cruise and directed by Spielberg, describes the use of technology to read images of "prophet" brainwaves in 2054 to detect criminal attempts and accurately Predict criminal behavior and arrest criminals before they commit a crime. The "prophet" in the film is a "human" with superpowers, but in the real world, the "prophet" is what we have often mentioned in recent years - big data analysis. Hermann Wimmer, president of Teradata International Group, a global software company specializing in big data analysis, believes that big data mainly includes three parts: first, traditional data, such as the original transaction system of the enterprise, Data warehouses such as network systems and ERP systems; the second is data generated by sensors; the third is data on social media.

"Now more and more industries have to adapt to the trend of big data, not only limited to the original high-tech and Internet companies, but now including communications, finance, manufacturing, energy and other industries. Competitiveness in this area is fostered by following the trends.” Hermann Wimmer said, “Using data to drive business growth” is the way of the future. "For example, how can the marketing department use real data to help formulate market growth strategies; how to improve customer experience or customer satisfaction; how to make enterprises operate smarter and more efficiently by reducing the operating costs of warehousing and logistics; how to combine production The data of the department and other departments to optimize production and operation capabilities, these are the 'useful places' of big data." Aaron Hsin, CEO of Teradata (Teradata) Greater China, said as an example.

For traditional manufacturing, what aspects can big data "subvert" and "improve"? McKinsey & Company's recently released report "How to Use Big Data to Improve Manufacturing" lists 10 paths for big data to subvert the manufacturing process, including optimizing production schedules; improving manufacturing performance; precise supplier management; tracking product quality and improving Work flow; determine production based on sales, formulate production plans; quantify production capacity, track equipment operating efficiency; and provide preventive maintenance recommendations for production equipment.

It can be said that big data affects all aspects of manufacturing, operation and management. From the current application scope of big data in the manufacturing industry, we want to focus on customer relationship management (CRM), production optimization and Three aspects of supply chain management spy on the infinite possibilities of big data.

The "needle in the haystack" is possible

In today's economic climate, good customer service and customer experience are critical. More and more companies are improving customer relationships and understanding customer needs by mining customer | data. Today's CRM data analysis capabilities are not limited to customer emails, phone calls and other data, but can identify customer purchasing behavior and understand customer sentiment. Xin Erlun has personally felt the changing trend of data analysis in customer management: "In the past, more data warehouses were used for customer relationship management and experience, especially customer data and CRM data to analyze and explore to promote marketing growth. Ways and means. With the evolution of technology and data architecture, data has now extended to many areas, such as location data, base station data, as well as call records and consumer behavior on the mobile Internet, etc. Use these data from multiple channels Build an analytical model to observe customers’ interests and hobbies from 360 degrees, and predict future behaviors, so as to formulate personalized marketing strategies.”

A marketing story that happened in Haier can reveal from this aspect The "magic" of big data. In 2012, Haier launched Dizun air conditioners. How to accurately predict which users may purchase Dizun air conditioners? How to send a personalized service plan? Haier extracted tens of thousands of user data from the SCRM member database, matched it with China Post's name and address database, and established a "look-alike" model. In addition, Haier SCRM membership platform also cooperates with travel and health magazines. Haier found that there were people who subscribed to travel magazines in a community in Beijing, and one of them was Mr. Chen. Haier came to the conclusion that "he should be interested in the environment and nature", so he speculated that he was very likely to be interested in the PM2.5 removal function of Dizun air conditioners. Then, Chen received a single page of direct mail delivered by Haier. In addition to the knowledge of public welfare and environmental protection, he focused on the PM2.5 function of Dizun air conditioner. The next story came naturally. Chen took the direct mail single page to a nearby supermarket to buy an air conditioner, and also logged on to the official website of Haier to register as a member of Haier.

It can be seen from this case that in terms of customer management, the target of corporate marketing is not only a group of people, but a specific person. Secondly, the integration of cross-domain data is also very important. Of course, enterprises should firstly identify which areas of data are needed and how to obtain them. Hermann Wimmer cites the business value of data sharing between two industries - the automotive industry and the insurance industry. "Everyone who buys a car needs to be insured. Because of their different driving habits, insurance companies will evaluate them differently. How can we more accurately assess whether a driver belongs to high-risk or low-risk driving habits? It depends on the car he drives. Through the data sent back by more than 100 sensors mounted on the car, we can understand his driving habits, and then determine what level of risk category he belongs to. For example, if he does not speed and drives smoothly, just Low-risk, on the other hand, driving quickly falls into the high-risk category.” Hermann Wimmer says that the two industries are closely linked by data from sensors.

Digital and intelligent production process

In traditional manufacturing enterprises, a large amount of data is distributed in various departments of the enterprise. There are certain difficulties with the data. For example, enterprise resource planning system (ERP) data, manufacturing execution system (EMS) data, etc. are located in their respective systems. In addition, in some intelligent factories, equipment, raw materials, etc. are embedded in microprocessors, Sensors, these devices generate huge amounts of data. Digitizing the manufacturing process also poses challenges for data processing and analysis. How to place this data on a technical processing platform has important implications for optimizing production processes, etc. Kong Yuhua, director of the Big Data Division of Teradata (Teradata), pointed out that new technologies can find out the correlation between people, things and things and events, but the premise is that this kind of big data analysis is based on On a unified platform that can realize data flow. This accessible platform enables the integration of data from disparate systems.

The easiest and most straightforward way is to create a Product Lifecycle Management (PLM) platform, which is also an enterprise management software, but the advantage is that it can fully integrate data from R&D, engineering, and production Virtual model of the production, optimize the production process, ensure that all departments in the enterprise work together with the same data, thereby improving the operational efficiency of the organization and shortening the product development and time-to-market. Zhou Kehu, senior business consultant at Siemens Industry Software (Shanghai) Co., Ltd., said: "Taking the automotive industry as an example, automotive R&D is an extremely complex process. On the one hand, it requires the cooperation of multiple functional teams. There is also a lot of data to deal with. In order to avoid poor communication and ensure the smooth running of the production process, the engineering team must not only manage the data within the team, but also keep abreast of the progress of the quality control team in the production department.”

PLM brings together all relevant information from the first draft, through the detailed design process, to actual production. As a result, companies can use this type of data collected by PLM to optimize their design and production processes. For example, Chery Automobile uses the PLM platform to link production planning, simulation and actual production, and link manufacturing and product development. For example, dimensional analysis plays an important role in car body design. Chery's R&D personnel use PLM tools for dimensional analysis, which can determine whether the design structure and production method meet the technical specifications at an early stage of design, so that solutions can be formulated early to optimize these factor. At the same time, these simulation programs can also be used to test various vehicle safety performances.

For example, variance analysis is available on Siemens' PLM software platform, which simulates the production process with the aid of a computer-generated 3D model, providing insight into weak points in the production process before actual production is carried out. Chery once used it to find out a problem in the production of a certain model's headlights, saving the company more than $100,000 in losses. Because the product design and production process can be simulated in a virtual environment, the efficiency of factory planning can be improved, and the production efficiency of the production line will also be improved.

Big data is the basis for intelligent manufacturing, which enables mass customization. Due to the large number of consumers and different needs of each person, the specific information of the demand is also different, and the demand is constantly changing, which constitutes the big data of product demand. Manufacturing enterprises process these data and then transmit it to smart devices for data mining, equipment adjustment, raw material preparation and other steps to produce customized products that meet individual needs. "The manufacturing of the future will be data-driven," says Hermann Wimmer.

Efficient and scientific supply chain management

The predictive function of big data greatly improves the role of big data in supply chain management. The manufacturing industry collects vast amounts of data from supply chain channels, as well as from instrument or sensor networks at the production site. Tighter integration and analysis of these databases using big data can help improve inventory management, the efficiency of sales and distribution processes, and the continuous monitoring of equipment. Big data can make logistics in the supply chain more efficient: electronic onboard video recorders in trucks can provide the location of the truck; how to quickly actuate sensors and radio frequency tags, etc., to help fully loaded trucks more effectively combine road conditions, traffic information and Weather conditions as well as the location of the customer, resulting in significant time and cost savings.

Kong Yuhua said that big data analysis in the supply chain can allow companies to formulate sales strategies scientifically, instead of relying on experience and taking risks as in the past. For example, a brand that produces down jackets has thousands of stores across the country, how to distribute 100,000 items to various stores across the country. 1,000 sets per store on average? Obviously not scientific enough. Because the supply and demand markets in the north and the south are different, the north has large demand but there are also many competing brands; in addition, the demand for clothes size in different regions is also different, the number of clothes worn by southerners is smaller, and the clothes of northerners may be larger. Through big data analysis, analysis of historical data, weather information, etc. can give enterprises reasonable suggestions: which freight is the most suitable to where, so as to avoid the problem of backlog or out-of-stock inventory.

Retailers are leading the way in the use of big data. The retail giant Walmart has developed a big data tool, through which suppliers can know the sales and inventory situation of each store in advance, so that they can replenish the goods themselves before Walmart issues orders, greatly reducing the situation of out-of-stock and the overall supply chain. inventory level. In this process, suppliers can also control the display of goods in the store, and Walmart also reduces the manpower and capital investment in this, which can be described as a win-win situation. Therefore, it is very meaningful for manufacturing enterprises to learn from these experiences to optimize supply chain management, from raw material procurement to logistics and distribution. According to big data and corresponding analysis tools, suitable suppliers and materials are selected in time or even in advance to be put into production and processing, and at the logistics stage, a reasonable distribution plan and sales strategy can be selected. With the support of big data, everything is scientific and reasonable, which not only improves production efficiency and service quality, but also reduces costs.