In the era of data, the evaluation of advertising performance often produces many problems. The problem to be solved by Attribution Analysis is the text message service generation of advertising effects, and how the credit should be reasonably allocated to which channels. 1. What is attribution analysis? In the era of complex data, we are faced with a large amount of data generated every day and the complex consumption behavior of users. Especially in the Internet advertising industry, there are often a series of problems in the evaluation of advertising effects: Which marketing channels led to the sale? What is their contribution rate? What kind of user behavior paths are behind these contributions? How can we use the conclusions from attribution analysis to guide our selection of channel combinations with higher conversion rates?
The problem to be solved by Attribution Analysis is the generation of advertising effects, and how the credit should be reasonably allocated to which channels. Your first reaction may be: of course, which ad I clicked on, and then went to the product details page to make a purchase, all the credit goes to this ad! Yes, this is also the most popular analysis text message service method today, the most simple and crude single-channel attribution model - this method usually attributes a sales conversion to the first ( First Model) or last contact ( The last interaction model, Last Model ) channel. But obviously, this is an analysis method that is not rigorous and accurate enough. for example: Xiao Chen saw the Moments advertisement on his mobile phone and released the latest Apple mobile phone. During the lunch break, he swiped Douyin to see that there were Internet celebrities reviewing the latest Apple mobile phone.
When he was on the subway after get off work, he found that there were already friends. After receiving the phone, he was posting pictures, so he drank a cup of Jiang Xiaobai and went home to apply for funds with his wife. In the end, his wife approved to let him buy it on JD.com, which is guaranteed. So, how much value did each channel of Moments, Douyin, Moments, text message service and JD contribute to this transaction? - It's too difficult, I don't know Attribution Analysis Model Analysis Another example: The following figure is an illustration of the behavior sequence of an e-commerce user. The meaning of each letter is D-advertising space, Q-product details page, D-recommended position, M-purchasing goods. So, what is the contribution rate of the three advertising spaces Da, Db, and Dc to this user's purchase behavior? This question is relatively simple, and you will understand it when you read the article! Attribution Analysis Model Analysis We found that the reality is often complex multi-channel delivery.