A5

.docx

School

University of Texas, Dallas *

*We aren’t endorsed by this school

Course

4352

Subject

Management

Date

May 5, 2024

Type

docx

Pages

3

Uploaded by AmbassadorMonkeyPerson711

Report
Assignment 5 Khushi Balar (kxb210017) Jindal School of Management, The University of Texas at Dallas ITSS 4352 : Web Analytics Professor. Vatsal Maru April 22, 2024
1. What would you advise Jaffer regarding the performance of the new data science algorithm? Ans: Utilizing the method below, the completion rate, click-through rate, and conversion rate for each model were determined based on the two-week result sets of Vungle's existing advertising model vs the data science algorithm model. Completion Rate = Completes/ Impressions Click Through Rate = Clicks/Impressions Conversion Rate = Installs/Impressions The calculated average of the completion rate, click-through rate, conversation rate, and eRPM for both models is given in the below table. Table 1: Average of both models Test A Test B Average Comepletion rate 0.892712518 0.887846759 Avg Click through rate 0.050677423 0.049239625 Avg conversion rate 0.004027151 0.003538106 Avg eRPM 3.3471 3.459 The campaign aims to raise the eRPM and conversion rate. The average conversion rate, a crucial metric for gauging the model's effectiveness, indicates that Test-A, the company's current promotional approach, is doing marginally better when looking at the average table. Test-A metrics for click-through rate and completion rate perform better than Test-B (data science advertisement model). But when taking into account Test-B, it performs better than Test-A in terms of daily average eRPM by $0.131 and $0.119. Jaffer should run the campaign for a little while longer to see if both models' findings are consistent, based on my analysis of the aforementioned model and the result set. Jaffer should also attempt to reroute more consumers to Test-B in order to comprehend the increase in the eRPM and conversation rate, as we are now sending 1 out of every 16 clients there. Redirecting more customers to Test-B could result in a greater discrepancy between performance compared to Test-A. Jaffer can make a decision based on the aforementioned recommendation for the data science model (Test-B), taking into account the difference in eRPM between the model and the current advertising model, as well as the cost of daily maintenance and updates. It is wise to approve the data science model based on the cost of updating it every day, as well as the eRPM and conversation rate if the ROI is more significant. If not, it is best to stick with the current model and try to enhance it by utilizing machine learning techniques to update it using historical user input and data sets, which reduces the cost of updating the model every day.
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help