Although the implementation of the last-touch attribution scheme is relatively simple, 100% of the value is directly attributed to the channel that was last contacted before the purchase or conversion, regardless of how many touchpoints consumers have contacted during the whole process. So much happens before conversions that the model completely ignores the impact of the actions of the upper and middle parts of the funnel on conversions.
Therefore, our company integrates the first touch point attribution and the last touch point attribution, and calculates the complete shopping link behavior of the user after entering the first-level traffic entrance and then completing it.
The definition of first-level traffic inflow is mobile number list that it is impossible to jump between the various entrances. You can only jump by switching tabs or return to the initial position and click to enter again. In this way, we can perform sales attribution based on the outermost layer of the complete link of shopping, and also know the complete path of the user's shopping, and at the same time ensure that the sum of the sales of each entry pit after sales attribution is equal to the sales of the day.
Using this fusion attribution, it is also possible to know the conversion rate of the intermediate steps. For example, related recommendations on the event site page and product detail page, although the sales attribution of the e-commerce platform as a whole will not calculate the sales of each model on the event site page, nor will it calculate the related recommendations on the product detail page.
However, since we recorded the detailed path after the user entered the first-level traffic portal, we can also calculate the sales of each module for comparative analysis when we separately study the efficiency of the event site page and the product detail page. But remember not to mix it with the sales of the first-level traffic entrance, which will lead to duplication of sales attribution.