Subdivision analysis is the basis of data analysis, and the information value of single-dimensional index data is very low.
Subdivision analysis can be roughly divided into two categories, one is gradual analysis, such as: visitors to Beijing can be divided into Chaoyang District and Haidian District; The other is dimension crossing, such as new visitors who pay SEM.
02 comparative analysis
Comparative analysis is mainly to compare two related data indicators with each other, quantitatively explain and display the relative values of the scale, level and speed of the research object, and then find out the problems existing in different stages of the business by comparing the index data of the same dimension.
Funnel analysis
Conversion funnel analysis is the basic model for data analysts to conduct business analysis. The most common thing we do is to set the final transformation as the realization of a certain purpose, and the most typical thing is to complete the transaction. But it can also be the realization of any other purpose, such as using the app for more than 10 minutes at a time.
Synchronous packet analysis
Queue analysis is very important in the field of data analysis and operation, especially in internet operation, and it is especially necessary to carefully observe the retained situation. By comparing the retention of comparable groups of the same nature, this paper analyzes what factors affect user retention.
05 cluster analysis
Cluster analysis is simple and intuitive. Clustering in web analytics can be divided into: users, pages or content, and sources.
User clustering is mainly embodied in user grouping and user label method; Page clustering is mainly a grouping method of similar and related pages; Source clustering mainly includes channels, keywords and so on.
06 AB test
One of the main ideas of growing hackers is never to make a big and comprehensive thing. On the contrary, it is necessary to keep making small and refined things in order to verify them quickly. Quick verification, how to verify? The main method is AB test.
Buried point analysis
Only by collecting enough basic data can we get the required analysis results through various analysis methods.
Through the analysis of user behavior, it is subdivided into browsing behavior, light interaction, heavy interaction, trading behavior, clicking button browsing behavior and light interaction behavior. Self-service embedding technology can improve the effectiveness of data analysis because of its frequent use and simple data, and the required data can be extracted immediately, which greatly reduces the workload of technicians and needs to collect more information.
08 source analysis
When the traffic dividend disappears, we attach great importance to customers. How to label user sources effectively is very important.
In traditional analysis tools, channel analysis has only a single dimension, which requires in-depth analysis of the effects of different channels at different stages, and cross-analysis of source channels such as SEM paid search and users' regions to get detailed information of customers in different regions. The finer the dimension, the more valuable the analysis results are.
09 user analysis
As we all know, user analysis is the core link of Internet operation, and the commonly used analysis methods are: active analysis, retention analysis, user grouping, user portrait, user scrutiny and so on. User activities can be subdivided into browsing activities, interactive activities, trading activities and so on. Through the subdivision of active behavior, we can master the key behavior indicators.
10 form analysis
In form analysis, filling in forms is a necessary link for each platform to interact with users. A perfect form design plays a vital role in improving customer conversion rate.
When the user enters the form page, a micro funnel has been generated. Judging from the total number of people who entered and completed the form and the number of people who successfully submitted the form, how many people started to fill in the form and what difficulties they encountered when filling in the form will affect the final transformation effect.
For the content of data analysis ideas commonly used by data analysts, Qingteng Bian Xiao is here to share with you. If you have a strong interest in Internet big data, I hope this article can help you. If you want to know more about big data, the skills and information of data analysts, you can click on other articles on this site to learn.