Large data applications have developed to a new stage

The advent of the "big Data" concept has opened a new era in data processing. Professor Xiaoyong, director of the University of Information, Chinese Institute of Computer Science, said Big data will bring three fundamental changes to business management (see the article "Big Data brings three fundamental changes").

These essential changes brought about by large data will extend the boundaries of management and the depth of management indefinitely. Business executives began to think about how to use large data to improve management efficiency. Although the concept of large data has been widely recognized for nearly two years, the commercial application model of large data has begun to escalate. We believe that in the past two years, the application of large data has undergone a rise from logical judgment to systematic thinking.

In the early days of large data applications, the application of large data is still in the use of large data to make simple logical judgments. For example, when many people use Google search for "influenza virus" this keyword, a wide range of geographical strong search, can let Google earlier than the major CDC predicted that the flu virus has begun to spread. However, these applications are only the initial exploration of large data capabilities.

In fact, with the transition from the initial logical thinking to the system thinking stage, the application strategy of large data has undergone the following three kinds of transformations:

Large data application transforms from initial logic judgment stage to system thinking stage

First, from the simple correlation of the search to the system survival relationship changes. The first phase of large data application provides people with a powerful tool for analyzing data relationships. When the various elements of the relationship are intertwined in reality, large data can be found in a state of chaos to identify possible relationships, pointing out the direction of management of this chaotic state; However, in the second phase of large data application, the search for simple correlation is only the beginning, how to use the system to survive the idea of using these relationships to become the key to success.

For example, while searching for the number of tweets associated with "the youth we will eventually die", it is 8 times times more likely to be discussed in small times than in the former. This data clearly foreshadowed the flop of the Little times. Le Vision decided to become the marketing side of the movie "Small Time" decisively. However, the data level can give the direct conclusion deduction, this is only the most basic application of large data, Weibo's attention rate is only a guarantee of investors return. Whether a movie can sell a box-office sale, the influence factor is various, the competition of other films of the same time, the promotion of various media and promotion means, even weather, school vacation Such factors may also affect the movie box office.

The continuous change of large data technology will not only discover the connection between the new elements and the existing elements, but also synthesize the various factors and view the relationship between each quality with the development of the system.

Second, from the problem of discovery to the change of interactive growth. In the first stage of large data application, people use large data to record and describe various variables and relationships of complex systems, and find problems through deep data mining to improve the efficiency of solving problems. In the second stage of large data application, the application of large data is more inclined to the deep cognition and understanding of the user, through the deep understanding to the user, constructs the product which conforms to the user's demand, and grows together with the user.

The case of the card house proves it well.

Netflix, the world's largest online pay video, online disc rental provider, has analyzed user habits through big data, using search techniques to observe user's viewing habits and found a seemingly unrelated "coincidence": Like Watching the 1990 BBC version of the card house. Audiences are also the fans of the famous director Fincher. They are also loyal fans of Oscar-winning Spacey. Netflix believes that a TV series that combines the three elements will have a much greater chance of success. So they invited Fincher to remake the card house and invited Spacey to be starring. Netflix spent 200 million of billions of dollars on a two-season new card house before any trailers or samples came out.

At the same time, because it is broadcast online, Netflix can easily pass a powerful database monitoring system, analysis of the "card House" after the line, the user where the pause button, how many users have seen a few episodes to give up, how many users replay and play the series again, this series of accurate data analysis, can provide references for future episodes.

When Netflix was making the series, it was entirely based on the interest of the user to analyze what the users of some common trait might taste, and then start making products based on their predicted user tastes and red. The production process of "card House" perfectly explains how large data can help enterprises to understand the user's needs in the interaction with the users, and help them to create the perfect products which meet the needs of users.

Third, from the established state analysis to the future transformation of ecological environment reconstruction. In the first phase of large data applications, large data is used as a static analysis of the user's current state. For example, large data can record the user's behavior, preferences, geographic location, and other real-time information, enterprises can use this data at the right time and place to push more accurate promotional information, better lock users. In contrast, in the second phase of large data applications, large data can help enterprises restructure user needs, create new ecological environment, and enable enterprises and users to grow organically with new business models.

Generally speaking, no one wants to publish their extremely private information on Facebook or Foursquare, but the Robert Wood Johnson Foundation of the United States (Robert Wood-Johnson Foundation) The PatientsLikeMe social network, which funded 1.9 million of billions of dollars, became an exception. In this open, chronic patient-specific social network, patients can measure their own illnesses, inquire about progress in treatment, and researchers can access their medical data. So far, nearly 200,000 users have created and shared their medical records on the platform.

"The gap between testing and medical care needs to be filled," Paul Wicks, director of PatientsLikeMe Research, said at the TED conference. When you have the right metrics and methods, you can do amazing things. "There are thousands of diseases, but few methods are available for patients to have a self-test," he said. Paul Wicks For example, multiple sclerosis patients can use a self-test form composed of seven questions, so far 30,000 users have participated in the completion. Paul Wicks hopes to create similar standardized self-test tools for other diseases.

Real Estate manager David Nors, aged 59, lives in the Virgin Is. Island of St. Chloe. He stumbled across the PatientsLikeMe website by searching online for information on new treatments for multiple sclerosis. Knowles has been with multiple sclerosis for more than 10 years and has participated in several patient community organizations. The website immediately attracted him. "No website has patientslikeme such detailed data. "By clicking on a symptom, you can see, ' Oh, there are 850 people who have this symptom, which is the therapy they're using," says Knowles. ’”

Knowles is particularly interested in a drug called his beads (Tysabri). One of his doctors had recommended the drug to him, but Knowles was skeptical about its side effects (affecting the brain, anxiety, fatigue). He found hundreds of people on the patientslikeme to take his beads to fight against the patients. After reviewing the results of their medication, he decided that for him the risks outweighed the rewards. So he took other therapies to talk to a doctor. "I think I'm in control of my medical care. "Of course I'll listen to neurologists, but now I'm listening to more advice from a team." Knowles praises PatientsLikeMe, saying it has "a wealth of treatment information from patients who have actually used these therapies." You can keep a close eye on the patient's response, whether it's three months, six months, or a year. ”

The profit model of the PatientsLikeMe website is to sell the user's information to the pharmaceutical manufacturer under the authorization of the user. Through these large and detailed records of user information, pharmaceutical companies can study the mechanism of various drugs on different patients and obtain sufficient information to develop new drugs.

PatientsLikeMe through in-depth analysis of user needs, a large number of users of data collection, for the patient, medical industry, pharmaceutical industry has provided a great value of data, these data in the whaley of the three sides, but also reshape the patient's way of treatment, the medical profession of the treatment program, and drug development and marketing programs in the pharmaceutical industry. Obviously, this ecological environment, which is reconstructed from large data, creates a new business model. This will be a business change that is worth looking forward to in the second phase of big data applications.

Zhou Jinchang is the managing partner of Deloitte China Technology, Media and Telecommunications (TMT), and Meng Zhaoli is the head of Deloitte China Technology, Media and Telecommunications (TMT) Excellence Center.