Your question is Test for Measurable A/B Change. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Hopper gets recommendations wrong in 2 ways: it says to buy tickets and ticket prices drop, or it says to wait and prices rise. If data scientists develop an algorithm that reduces the later, how would you test for a measurable change, what would you control for, and what external data would you gather?