Escalating a 5% Drop in YouTube Traffic: Validation, Triage and the Manager Update

You are a product manager at YouTube, and you learn that YouTube's traffic has fallen by 5%. How would you raise this with your manager? Describe what you would establish before raising it, what your first update would contain, and how you would keep your manager informed as the investigation goes on. ```hint Balance speed and certainty A 5% drop on a product this large may already be costing a great deal, yet it could also be a data glitch. What is the smallest set of checks that lets you tell your manager something reliable soon? ``` ### Clarifying Questions - What does "traffic" mean here: visits, searches, watch time or daily active users, and did all of them fall? - Did the drop happen suddenly, at a specific time, or gradually over weeks, and against which baseline is the 5% measured? - Is it concentrated in particular countries, platforms or parts of YouTube, or spread evenly? - Does the team have an incident process for drops like this, and does my manager already know? ### What a Strong Answer Covers - Quick validation that the drop is real (logging, data pipelines, dashboard definitions) and large compared with normal variation - Sizing and slicing: which metric, which segments, sudden or gradual - A short, ordered list of likely causes, internal and external, and how each would be checked - A clear first update: what is known, what is not, impact, owners, next steps and when the next update comes - Judgment on urgency and escalation, including when to recommend an action such as a rollback ### Follow-up Questions - Two hours in, you suspect a recent release but cannot prove it. Do you recommend rolling it back, and how do you frame that recommendation? - Your manager asks you to write the incident summary for leadership. What goes in it? - How would you change monitoring so a drop like this is caught and explained faster next time?