many Internet practitioners are aware of the need to pay attention to the user, and the user is concerned about the user’s research. So, how to use the Internet in the end is how to do? What are the pain points now? What can improve the place?
recently participated in a forum on the application of game data, organized by the invitation of the game in Guangzhou to do operations, planning and data analysis of the people, talking about the topic of the game’s big data applications". After listening to a circle of people feel the sense of the data is still relatively weak, the basic operating data may not be able to do fine, let alone big data application. Not only is the game, probably the majority of Internet Co are so.
Status quo of
Internet application research
talk about the status quo, first of all want to talk about the concept of research. I think the current research and application of the Internet industry mainly includes the following two aspects:
Analysis and application of
1 Product data
this is the product of the Internet, because the Internet products can be a good record in user generated in the product, resulting in massive objective and detailed data, this is the envy of traditional industries. The perspective of the application of data may be divided into three levels Internet Co:
(1) is the basic analysis layer
based data includes both macro management data we often say, such as Talking Data Mobile Games data system, funnel model basically follow the Come-Stay-Pay-Leave routine, the data is almost a must see also the management of the concerned data, small firms may be obtained by connecting the alliance, Baidu Sinan and three party data services company SDK and the web page code; basic data also includes the product itself or the operation itself generated data, such as the level of the activity rate, click conversion rate and so on, this part of the data is often related to people in the face of business problems, looking for product development or the background data support departments to run, belong to the type specific analysis specific problems.
(2) layer is the BI center layer of
this layer of the company has been relatively rich product line, began to seek to build a unified database background to improve their own efficiency. But most of the company’s so-called data background, it may just be the first layer of some of the conventional data into a template, real-time or periodically automatically update data. A better application is to set up the company’s level of the label system, the particle size is accurate to the user, through the combination of tags to grab specific populations to analyze, which will be mentioned in the following. From another point of view, do the analysis of the data in these companies, certainly do not want to just stay in the level of operational data, will try to do a certain stage of data modeling, the core idea is actually associated with clustering.
(3) layer is the data application layer