I am trying to measure the impact of the launch of a staged test that occurred in various states over different months.But I am having trouble understanding the way lift is calculated in an existing approach.
Jan 2019 - State A; Feb 2019 - State B; March 2019 - State C
States A, B, C are my test groups, each measured separately against the control group which is rest of US excluding these three states.
I am currently measuring pre period : Three months before Launch; Post Period - Launch Month Till March 2019
Formula used: ((Post Lift % in State A - Pre Lift % in State A) - (Post Lift% in Rest of US - Pre Left% in Rest of US))*(Share of Volume from State A in Q1/Share of Volume from all of US in Q1)
Post Lift % is defined as lift from Jan 2019 through March 2019 vs. LIft from Jan 2018 through March 2018 and pre lift% is Lift from Oct 2018 through DEcember 2018 vs. Oct 2017 vs. December 2017. This period shifts based on Launch Date of the State.
My question is :
A) I don't understand why I would do this: (Share of Volume from State A in Q1/Share of Volume from all of US in Q1)
I tried looking up on the internet and I can only find posts which explain till Lift% without the Share of Volume calculation.
Appreciate your help.
(Share of Volume from State A in Q1/Share of Volume from all of US in Q1) + (Share of Volume from State B in Q1/Share of Volume from all of US in Q1) + ... = $1$
– AgentS Aug 02 '19 at 20:13