Big data initially means the large data set which is not able to be analyzed, but later it was derivatized to the method to analyze huge amounts of data in order to gain great value.
This is a form which gradually gets attention, It's difficult to analyze these data and it's also difficult to store these data and it needs some unprecedented way, Now in China many companies use the open source Hadoop distributed data cluster to meet the needs of data statistics.
Since we can get segmented data with great value in huge data through data analysis, so many websites have big data strategy, especially in e-Commerce and video industry.
But the more challenging and representative among them should be the data mining in video websites, after all it's easy t have data mining in e-Commerce sites because no matter what users choose, they will eventually select a a product and then register, login and pay, they will unveil some important information in this process. But users on video websites do not register usually, how to do big data analysis without user information?
Recently, a Netflix homemade drama "House of cards" is concerned, it is said Netflix inferred the key elements of the success of the "House of cards" based on the data such as users like BBC Drama, the performance style of David Fincher. Ultimately it proved the conclusion based on big data analysis is convincing, "House of cards" gains popularity in the United States and more than 40 other countries, Netflix was a great success.
Who does best in big data?
Thus, we have a doubt, is there any video site doing big data? Who does the best?
Is it Netflix? At the technical level, I think it's not.
The reason why Netflix is successful is that its data can be used for data mining, Netflix is an online DVD rental site, they can know users prefer to BBC drama through their rental data, like the alternative expression of the social problems of David Fincher, so Netflix successfully launched the "house of cards".
YouTube does the best on anonymous user data mining, After I visited this magic video site, I am impressed by the precision of its recommendation algorithm and also the openness of its data, As long as you want statistics and analysis, YouTube will provide all data to you, and this should be attributed to rely on Google's strong technical support.
There are several video sites doing big data in China including Youku, Sohu; Ku6, Tudou, 56 etc; they also will become contenders for the video site big data.
Localized application of big data
The benefits of big data analysis is responsive, on-demand request. So in which area can Chinese video sites generate value?
Content recommendation is the most basic application, to provide users with more video to meet the appetite, increase the click rate and get more advertising exposure, It is the most direct form of gaining profit. This is the most simple and easiest to implement and is also a method commonly adopted by video sites, But depends on the degree of accuracy of the recommendation algorithm, many websites use simple words splitting, tags and other forms to recommend videos for users.
Replicable business model?
Are there other profit models verified based on big data in other industries? Are these models replicable in video industry?
Netflix already provides us a good example, to select directors and actors by analyzing users' preference, This approach can be perfectly copied to China. In 2013 Youku and Qiyi bet in the area of self-made drama. Two years ago, there were several successful self made dramas.
Another model is to copy Baidu's advertisement bidding model, by providing data analysis support for users and advertising marketing vendors,Let advertisers post advertisement by finding valuable videos themselves To create an eco-system which benefits the website and the industry.
Worthless video has no ads value, valuable video will be liked by advertisers. To increase the income of the original author through the content of the video itself can be described as a win-win rule; It not only encourages the enthusiasm of the original author, but also enable the video site continuously improve the quality of the content of the video.
Source : http://www.36kr.com/p/202522.html
