Summary of foreign big data applications retail experience

The amount of data generated in the past three years is still more than the amount of data in the past 40,000 years. There is no doubt about the advent of the era of big data. Big data can come from all aspects, from shopping transactions in life to industrial manufacturing; from social media information to online video images; from enterprise information management systems to government-owned e-government, Data generation.

Therefore, we are about to face a revolution. Emerging big data will become a top priority for enterprise development. However, conventional technologies have been hard-pressed to handle large amounts of PB (1024TB) data. The challenge posed by this change is that successful businesses must face the future development process. Only those companies that can take advantage of these new data types can create a sustainable competitive advantage.

Data should be flexible, data should be rapid

Guess has a number of product lines, but how to choose among many apparel products, the main push style or accessory style is the key factor in the sales trend. To make the right decisions, the data analysis mechanism needs to query all the data in the spreadsheet. The process took hundreds of hours and Michael Relich, senior vice president and CIO at the company, had to deal with an endless stream of calls due to a large number of queries. After learning how Guess handled the master data, Relich and his technical team developed a two-dimensional code solution that customers can use to scan the iPad and view detailed information about the apparel on the screen.

This project is not part of any strategic planning or long-term move. On the contrary, it is entirely derived from the actual needs of customers. In fact, most big data analysis projects are directly driven by customers.

According to William Miller Jr., senior vice president and CIO at semiconductor chip maker Broadcom, this means that technicians should look ahead to demand and not just catch up to demand.

At Broadcom, IT and engineering teams work together in the same operating environment. To meet the interactivity needs of both parties, the company deployed corporate social networking technologies such as Jabber and Yammer to help engineers organize spontaneous teams as they happen, and then disband as soon as the task is completed. "This allows engineers to maximize productivity because they do not have to spend time slowly organizing and coordinating their work environment," explains Miller. "This is a very efficient process or organization."

Broadcom engineers also set up a private cloud environment, which ensures that staff can work flexibly at any time of the day, including the night. Broadcom has just acquired a new company a few days ago and expatriate employees find it hard to find that their familiar solutions and tools are fully integrated into Broadcom's work environment in just a single weekend - a remarkable example of flexibility. "It convinced us that the previous investments were worth the money," Miller told us.

Data visualization to meet practical needs

To be effective in big data investments, we must ensure that the information generated by the analytics is presented visually to staff. In the past, Liberty Mutual Insurance International received a quarterly report of up to 80 pages of operations each quarter, while executive teams spent hours reading and learning about it. Senior Vice President and CIO Mojgan Lefebvre Said.

The translation of this monologue into a list of key performance indicators needs to be done in stages, Lefebvre explains. First, decide which key questions need to be answered. Then, create a series of grouping reports. Finally, data visualization. The whole process was meant "to make data fit for real needs by tidying up." Lefebvre added that the conversion process often took more than a quarter of a second to get the report really working.

"Getting people to use data warehousing is a huge problem," Relich said. Especially in the retail industry this issue is particularly prominent, because buyers often have a very strong intuitive judgments. Guess's solution will first convert valuable data to the BlackBerry platform after formatting, then visualize it with a dashboard backed by Flash and eventually pass the results to the iPad app. Relich points out that aggregating sales data from each retail store can result in more efficient decision-making while reducing the number of planning meetings.

Real-time data visualization is not limited to the number itself. LIBO insurance company's on-site risk engineers can take pictures of the actual situation and upload to the database side. Here again, the picture will be inspected by insurance rate engineers again. According to Lefebvre, this mechanism allows the review cycle of premium payments to be reduced from the previous days to the present hours.

In the next five years, big data will gradually become part of an increasing number of CIOs. On the one hand, the popularity of business intelligence, so that enterprises have a full understanding of the importance of the data; the other hand, the rise of new applications such as social media, e-commerce, Internet of things, breaking the original value chain as a business wall, only Analyzing the data from all aspects of the original value chain can no longer meet the demand. They need to break the data boundary with the help of big data strategy and understand the panorama of a more comprehensive operation and operation environment.