The advent of the big data age

Foreword: Recently read a lot of Big data analysis article, feel Big Data era coming. The Harvard Business Review also opened a new column called "Big Data". I have sorted out all the large data analysis articles which are collected and ready to record, this article is the most suitable for the first share. The article takes the case as the support, narrated the Big Data Analysis 4 basic points. After the article was recorded, Yoyo also recorded some thoughts on the analysis of large data according to the chat with the leader of the data analysis Group of the League of Heroes.

Http://blogs.hbr.org/cs/2012/10/getting_started_on_a_big_data.html

Inspiration:big Data strategy needs resources and capability.

The author begins by proposing that the collection and analysis of large data has become a new field of research on differentiated competitiveness. According to a recent Harvard Business Review article, the company needs 3 key competencies to support large data analysis: 1, locate and manage multiple data sources, 2, build up the ability to improve the data analysis model, and 3, the determination and management ability of the company to transform.

However, it is not an easy way to make large data analysis successfully. This process requires managers to constantly try to find a clear data analysis strategy. The author delves into several recent companies using large data analysis strategies to extract 4 basic points of a large data analysis strategy:

1, grasp the opportunity to deal with risks
Large data analysis brings a lot of opportunities to the company, such as the promotion of core competitiveness, innovation and so on. For example, insurance companies can use large data analysis to optimize policies; In the long run, insurers can even expand new insurance businesses with large data analysis. The key to success is to establish clear business goals and prioritize them at every stage of execution.

Large data analysis is a strategy that big companies must focus on. In one case studied by the authors, a traditional large retail company faces a new threat from a network retail company. This network company through the user's data analysis, obtains the detailed user data, and according to the data conclusion makes the seller easier to sell the goods more efficiently, thus affects the big company's sales and the market share. Large companies face such a threat, should focus on two points to solve: 1, to confirm their own market, 2, a reasonable investment in large data analysis to optimize the company's capabilities.

2. Ensure data source and data analysis capability
Before confirming a large data analysis strategy, you must ensure that the data source is available and that you have reasonable data analysis capabilities. For example, the data required for large data analysis may not come from a single data source, but rather multiple sources of data, sometimes requiring the support and collaboration of partner departments. After the data is obtained, the data will not produce the complete value without the data analysis by the specialized data analyst. Therefore, the complete data source, the effective data analysis strength, is the big data analysis necessary condition.

3, data analysis strategy and corporate strategy unification
When identifying the opportunities and resources for large data analysis, many companies are anxious to get their conclusions and execute them as soon as possible. This process is wrong. The data analyst needs careful planning and ample time to do large data analysis work.

In one case, a communications company has established an executive team to oversee the data analysis team to ensure that the strategy for large data analysis is consistent with the company's overall direction. The executive team focused on two questions: 1, how much competitiveness/influence do our brands have when they make a purchase decision? 2, what is the most important factor to consider when buying a user? is our product effectively explaining these factors to the user?

Based on the problem of the executive team, the data Analysis team obtains user data from multiple data sources and tries to draw a workable conclusion through analysis. For example, the sports channel and the pay channel are the key factors that users decide to buy. The data analysis team also found that when companies weakened telephony, users were unwilling to buy packages containing televisions, networks and telephones. The study also found that companies should add a mobile phone program to the package to meet user needs. The authors state that these conclusions were not found in traditional market research.

4, the importance of executive attention
The risks and opportunities of large data analysis are often only controlled by company executives. For example, a telecoms company's data analysis team found that two factors could lead to a company being criticized on the Web: 1, a network crash; 2, where users think the company is publishing misleading ads. As soon as the conclusion comes out, the company's network department and marketing department immediately blame each other. At this time, the company's senior management to coordinate and solve the problem.

The authors point out that executives must pay close attention and focus on large data analysis to make large data analysis bring value to the company.

——————————— below is the Yoyo and the Hero Alliance data analysis head chat record and thinking ———————————

First of all, lol players should understand that riot is Tencent's company oh lol. And congratulations LOL is already the most popular PC game in the world.

LOL's data analysis group head FEI from Los Angeles to Shenzhen on business, yoyo as a new LOL player (technically pretty rough), very happy to chat with Fei for a noon, the discussion focused on the data analysis of LOL.

LOL has many items such as heroic character, hero Skin, new hero of fixed cycle, game balance, commodity price and so on. The properties of these items are not arbitrarily determined. The data Analysis Group provides effective data analysis support for these key projects by using a reasonable data model, conducting business analysis and supporting decision-making in key departments.

To give a simple example, lol There are many heroes, the image of each hero design, the story set, supporting the skin, the company needs to invest a lot of resources to design production. How do you prioritize? Very simply, large data analysis can study massive user data, understand the user's favorite hero, and then focus on optimizing the most popular hero roles.

The price of a hero is also calculated from a rigorous data model, and this large data analysis is consistent with our intuitive understanding that cheap heroes are good controls and can attract new users of heroes. For example, 450 can buy the ice.

Overall feeling, riot company attaches great importance to the analysis of large data, and has done to the large data analysis conclusions to support important department decision, important Department Trust data analysis Conclusion Virtuous circle.

Finally think of their own now responsible for the Tencent open platform, cloud Platform product design work, which also has a lot of large data analysis opportunities. Developers to access high-quality applications to Tencent open platform, from the application of access to the final display of the Application Center, as well as in the management system and compass services, applications and users are constantly generating data. The value of this large data is very clear, can give developers, open platform, advertisers great value. A simple example, if a developer in-depth understanding of the use of their own users of the data, habits and other key factors, developers can be more targeted design applications, put ads, and promote their application. Finally achieve a win situation: the application of developers to get more users and active use, users get closer to their use habits of application.

Thank you for reading, welcome to the big data interested colleagues to discuss! ：）

Next Yoyo will find time to continue to write large data related articles.

Stay tuned.

：）