Public comment Longwin: Big data is not an industry, it is difficult to form an industry

Our reporter slipped Ming fei Shanghai report

"The public comment itself is a big data site. "Public comments, co-founder Longwin, in an interview with the 21st century economic reporter, said that as of the third quarter of this year, the number of comments by the public has reached 42 million. It is also understood that the current public comment on the monthly active users more than 170 million, data service calls in Level 1 billion.

For the development of large data, Longwin said that a few years ago, large data in the high-speed development phase, technology is the peak period, and the business is in the desire to explode. In other words, the prior period must have the accumulation of data, and then can have follow-up development applications, if there is no data accumulation, there is no big data.

Unlike some of the market's views, longwin that big data is not an industry at all, because big data involves all walks of life. At the same time, he thinks the big data is difficult to form the industry, because professional big data mining companies do not necessarily know every industry, "there must be a chance, but definitely not so easy."

"I don't think the big data is an industry, just an application, and if it's useless for the company's business, users and businesses, big data doesn't make sense." "Longwin said.

In fact, the UGC and O2O models based on the public comments have different characteristics in large data collection, mining and application.

Openness is a trend of big data

"21st century": the public comment on the current large data model?

Longwin: The public comments on the content of the site (that is, data) are generated through the UGC, from the first user to upload the first merchant, the data began to produce. The first is the basic information of the business display, and then more and more users to comment on the process, the platform has accumulated business and user data.

One is about business address, dishes, environment, services and other aspects of the data, one is the user's consumption habits, and in the user comments process, there is a business data constantly adjust and improve the mechanism.

The data is richer when the general public reviews the trading business. At present, in the large data structure that the public reviews, the user's behavior log data accounts for the majority of the total amount, and the remaining is the transaction data.

From this point of view, the whole public comment itself is a large data site, based on user uploads, consumption accumulation, the large data presented.

"21st Century": UGC mode, how to drive user production content? How does the platform continuously optimize the integration of these data?

Longwin: No material rewards, mainly the user's psychological satisfaction, such as you are the first to find a restaurant, or your comments get how many praise.

To optimize the data, such as the 2003 station, only a user comments when there is no value, when the number of comments reached 25, on the more objective, by playing the form of fractions--that is, the calculation of a weighted average number, according to the user weight, star higher weight higher, and thus in the background to achieve scoring. This mode can be quickly replicated to all businesses.

Through the accumulation of nearly 12 years, the public comments in the business data structure, user data integration analysis and so on have certain rules.

"21st century": The public reviews what is the big data application?

Longwin: currently divided into two categories: one kind is the merchant pass, the promotion pass, these are the charge product, another kind, for example the populace reviews the index, provides the decision for the consumer, also has the catering industry vane, provides the profession development reference, simultaneously in the business aspect may provide the user personalized recommendation; reviews the Butler can help the merchant to analyze User characteristics. Now these products are free, the future is uncertain.

At the same time, the public reviews are also doing an open platform, the data to the third party open, on this basis, they carried out two development.

Moving towards openness is a big data trend.

80% from the mobile end

"21st century": compared to the PC-side, the advent of the mobile internet era, what will be the impact of large data?

Longwin: The first is the increase in user volume, data increase, the current public comment on the business more than 80% from the mobile end.

More importantly, the mobile end can obtain data information, such as geographic information, that is not available on the PC. For O2O Enterprises, this is very important. For example, you often in the vicinity of Xujiahui activities, then next to the merchant is more important to you, usually within the surrounding kilometer. And for Taobao such as physical electricity quotient, geographical location is not so important.

The time dimension also changes. For example, the query, the time of the PC query is relatively fixed, mobile end anywhere to check, the data call interaction requirements are very high.

"21st century": How do you view the current domestic development of large data industry? Can a new industry be formed?

Longwin: Many are shouting slogans, what hot shout what. I think, first of all to do a good job, have this ability, let its natural. And I don't think big data is an industry because it involves all walks of life.

Although there are a lot of big data on the start-up companies, but the formation of the industry is unlikely. The public comment is oneself accumulates, the excavation, but for the traditional profession, does not rule to rely on the specialized data company to enlarge the data. But each industry has the characteristics of each industry, the need to tailor tailored, specialized large data mining companies, how much do you know about the industry? But there's got to be a chance, but it's not that easy.

"21st century": In recent years, the traditional industry is also exploring the use of large data, what do you think compared to the internet, what is the characteristics?

Longwin: Large data are not internet and non-internet. The Internet through years of development, just make data accumulation more convenient, the user's every access is recorded on the server. Traditional industries, however, may have to develop a system to collect data manually.

Therefore, the combination of large data and the Internet is due to the natural advantage of collecting data, and the same data collected is not necessarily all useful. Now the traditional industry more and more attention to large data, they are in the collection, accumulation of large data may be invested more expensive, but in data mining is not necessarily poor, different industries have different characteristics, even if the Internet companies do not understand the traditional industry business model and characteristics, also not.

"21st century": for enterprises, how to position the role of large data?

Longwin: Large data accumulation and mining must first be valuable to the company's business, better know the user's habits, and how to meet the needs of users, to provide decision-making for operations. On this basis, then to meet the needs of business.

Do not large data for large data, with large data, sowhat? Or to make the user more convenient.

"21st century": compared with the physical electrical quotient, the O2O mode of catering such as the data produced is relatively thin?

Longwin: Catering industry has its own characteristics, consumption frequency, the amount of different people to make decisions. Food and beverage no warehousing logistics, I only need a serial number, consumers go to consumption, businessmen to the public review of the background verification confirmation, the data will be in the database. O2O has its own characteristics, useful data for other companies may not be so critical to us.

The challenge is to build the data model

21st Century: What challenges do you think the current development and application of large data are facing?

Longwin: The accumulation, the consummation, the iteration, will meet some challenges. Large data mining needs to have long-term precipitation, but the user behavior changes are very fast, quarterly, monthly, or even weekly, the product is in a change.

For example, milk, relatively small changes, can be long-term according to certain rules of accumulation of data, more stable. But the Internet industry product iteration is very fast, just precipitated one months of data, tomorrow your data type may be changing, so the challenge is not technology, but the key is how to build the data model. Milk products are stable, data models are stable, but the internet requires a lot of flexibility.

"21st century": Recently, the apple "pornographic door" incident also exposed a large data era of some security, privacy issues, how do you think of large data security?

Longwin: Security itself should not be a serious problem. On the one hand, the company's own integrity should not be a problem, we must protect the user's privacy, but do not exclude abuse; the second is the external attack, this is a purely technical problem, with the Internet, the problem exists, and not because of this problem the Internet will die. Large data security is mainly through internal management system and technology double protection.

"21st century": Will the public comment on the business of Internet finance based on large data?

Longwin: There are also banks interested in cooperation, because the public reviews have a good large data system, to understand the cash flow of businesses, operations, can be based on these data analysis of financial business, there is a possibility.