Design an experiment question:

The Integrityteam wants to test the precision of new enforcement that is meant to target spammers on Facebook. This enforcement blocks all messages that happen between users who don't share any connections (mutual friends). The team wants you to design the experiment in a way that will provide conclusive results.

1. How would you design and set up this experiment?
2. What information, if any, would you need to properly determine the sample size and length of the experiment?
3. Which metrics would you plan to track in order to define the performance of the enforcement? Why did you choose these metrics?

This is such an open ended question not sure how to answer this.