Exception: First of all, the following data are written randomly, mainly to illustrate the usage of this model. If the screenshot involves infringement, please contact the author to delete it . Generally, we will make a series of reports for product operation in the company, one of which is used to monitor the trend of Siemens orders every day, as well as the trend of the chain, as shown in Figure 1 below: The blue curve is the number of users who paid successfully. business email list We saw that the number of users increased on July 12, and the number of users increased again on the 19th. It can also be seen from the orange line that No. 12 has risen by a step. So, the question is, we want to know what is the reason for the number of users on the 12th and 19th? Figure 1:
After the user searches, a series of refrigerators business email list will be displayed. If we monitor the successful payment from the marketing page to the end, the specific data is: the number of users on the marketing page - the number of users on the product selection page - the number of users who created orders - the successful payment users number. We can use these data to make a trend graph of the conversion rate, as shown in Figure 2 below: Through Figure 2, we can only see that the conversion rate of the first and third steps has decreased, and the second step has increased. In fact, we can also see that a data is that the UV of the marketing page has become higher.
But how do we know: Is it caused by the higher UV of the marketing page? Or did the conversion rate in the second step increase, resulting in an increase in the number of users who made the final payment? Model analysis: Because the total increase is mainly after the 12th and 19th. Therefore, in general, for this kind of abnormality, we will first communicate with the operation or product students whether they have done anything, and confirm whether it is caused by changes or normal fluctuations. For the specific abnormal analysis framework, see: "Growth Hacking-DAU Decline Analysis-Indicators" Anomaly Analysis Framework" After confirming, we can take the data of No. 12 and No. 19 as examples to see how to use this model. The specific data are as follows.