Statistics Business Case

Dear all!

I wanted to post this business case since I know that it can bring value to our forum.
In order to apply for a job I got asked to solve a business case that goes as follows: There are 10 business units in a corporation. The business units all have their own expenses, however the administrative expenses are not accounted for each of them (there is no profit & loss per BU) but only a yearly total administrative expenses account i.e. a single expenses figure at the end of the year. We need to assign those administrative expenses to each of the BUs. For this, you get 5 different variables: amount of sales per BU/year, # of invoices per BU/year, # of employees employed in each BU, # of local purchases per BU/year, # of import transactions per BU/year.

With the information given we need to develop a formula that assign the yearly administrative expenses taking into account all of the variables. Do you have any idea how I can develop this formula?

I was thinking about a regression/correlation analysis that helps me first define if the variables make sense and are correlated, they are. The objective is to find a formula that assigns the administrative expenses within the BUs with these drivers of expenses: sales, transactionality, personnel employed, etc.

Any help or thoughts are greatly appreaciated, the deadline to post solutions is by wednesday evening


Last edited: