The development way of video website under large data pattern

Large data initially refers to data sets that are too much to be counted and analyzed, and later derived from--to analyze massive amounts of data in some way and method to get the service content that we need with great value benefits.

This is a form of gradual attention, analysis of the data is difficult to store the data as difficult, but also need to use some unprecedented way, now most of the domestic use of Hadoop, such as open source distributed data cluster to meet the needs of a variety of data statistics.

Because of the large amount of data can be analyzed with great value of the segmentation data, so many sites have a large data strategy, especially in the electric business and video industry.

But the more challenging and representative should be the big data mining for video sites, after all, the electrical business for large data still have a lot of advantages, such as the electricity business users choose anyway, eventually will select a product, and then register, landing, payment, virtually the most important information of their own. and video site Most users are not registered, how in this can not record user information how to do large data analysis?

A recent Netflix homemade play, the card house, has been a concern, and it is said to be a key ingredient in the success of the card house, based on data from Netflix: People who like BBC episodes, Fincher style, spacey performance. It turned out that the conclusions based on large data analysis were very reliable, and that the "card house" sparked a dramatic upsurge in the US and more than 40 other countries, and that Netflix had been hugely successful.

Big data who did the best?

From this we have a question, video industry someone is doing big data? Who's the best?

Is it Netflix? On a technical level, I don't think so.

Netflix's success lies in the fact that its data can be tapped, and Netflix is an online DVD rental website that, through direct user leasing data, can tell that users like the BBC's episodes, like Fincher's films on social issues, and Kevin Spacey in the portrayal of a variety of sinister villains of the superb acting, so successfully launched the "card House" this high-profile quality of the U.S. drama.

YouTube is the best thing to do in the big data on anonymous users, after I contacted this magical video site, not only to his accurate recommendation algorithm, but also to admire their openness to the data, as long as you want to statistics and analysis, YouTube will be almost all the video data to you, This should be attributed to the strong technical support they attach to Google.

All the big data information on YouTube is centered around a single video program, the first visit, the first share, the first comment, all in front of you, including the time and the URL, which is a kind of open-minded.

There are also several video sites for large video data: Youku, Sohu (data comparison), 酷6网, potatoes, 56 (provide some data), odd arts, music, Tencent (with statistics but not provided), they will also become the video site big data contenders.

Localized application of large data

The benefit of large data analysis is that it is responsive and available on demand; What are the areas that can generate value for the profit model of domestic video websites?

Content recommendation is the most basic application, to provide users with more consistent with the appetite of video, increase the click Rate, access to more advertising exposure, is the most direct form of profit. This is also the simplest, easiest to implement, but also the major video sites commonly used methods, but depends on the accuracy of the algorithm. A lot of simple use of cutting words, tags and other forms, for the user video recommendation is considered to be accurate recommended, and for its name "recommended algorithm" four words, we can not talk about what, after all, based on large data recommendation algorithm is not a simple project.

Secondly, based on the data, the user is screened according to the condition, and the profit content is promoted by different dimensions, such as the network promotion of the film and television works, the content preheating report, the recommended VOD of the paid video and so on. Through the subdivision of the user group, the result is higher, but the basic service is also quite varied. such as membership system, payment system, billing system, video content encryption method, and the function of fuzzy recommendation, all need a period of development, relative to invest more manpower and development costs.

Replicable business model?

Are there other industries that have been validated for their profit model based on large data, and that the patterns can be replicated and applied to video sites?

Netflix has provided us with a good example of how users ' preferences can be used to select directors and actors, and this is a perfect way to replicate domestically. 2013 Youku and Chimei are betting on homemade episodes. Two years ago, there are several successful micro-film series, such as "Adolescence" series, and later a few of the more popular "Godfather", "Upper", "mummy" is mostly to attract the attention of the eyes of the way to win users. If you can through large data analysis, the success of the real quality and taste to achieve radio and television standards of homemade drama, in the domestic also create a successful case, it will be possible to lay its dominant position.

Another case is the replication of Baidu's auction advertising model, by providing users and advertisers with large data analysis and support, let their own advertising, let the advertisers discover their own value of the video, put ads, generate an independent marketing of the ecological environment, will be conducive to the development of the website and industry.

Worthless video without the value of advertising, the value of the video by the advertisers have been competing for competition, through the content of the video itself to improve the original author's economic income, can be said to be the law of mutual benefit, not only to encourage the enthusiasm of the original author, but also can make the site's content quality of video constantly improve, forming an insurmountable value

(Responsible editor: Lu Guang)