I have received a very ill formulated question that I have a hard time wrapping my head around - so I hope you can help!

INTRO: Top-beer is a local brewery in US, where you work as a digital marketing manager. To better target the customers, Top-beer decides to use email marketing for selling. The basic idea is that Top-beer sign contracts with the advertising agency. The latter will send out the Top-beer advertisement emails to its subscribers. After a series of meetings with you team, you have the budget planning. The cost of each advertisement email should not greater than 2-cent USD. A successful click from each advertisement email is estimated to bring 8-cent USD for Top-beer. Now, you need to shortlist several agencies to approach.

  1. Based the scenario above, please create a decision-tree model and discuss why your model is valid. Based on your proposed decision-tree model, calculate the break-even point.

  2. To better design such email marketing campaign, you asked the advertising agencies to share their historical data of email marketing and they agreed. Which data will you ask for from the advertising agencies? Why? Please provide detailed description of the data you request.

  3. Suppose all your requested data given, please discuss how you will use such dataset to shortlist the qualified advertising agencies. Please specify each analysis step and the associated STATA commands.

We have to use logistic regression.

I have made a decision tree of "click" and "no-click" which I guess will also be the dependent variable for the logistic regression.

However, I do not understand how this can be used to shortlist agencies as click - no click will be on a consumer level and thus I will be able to tell whether a consumer will click or not.

Do I misinterpret the question - or would you do the same?

Thank you in advance!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.