Diagnose a 10% Drop in Google Docs Document Opens

Quick Overview

A metric diagnosis question in which document opens for Google Docs fall by 10%. It tests clarifying the metric and timing, ruling out logging and seasonal effects, decomposing and segmenting the decline, ranking hypotheses with checks, and judging whether fewer opens is truly a problem.

Diagnose a 10% Drop in Google Docs Document Opens

Company: Google

Role: Product Manager

Category: Analytics & Experimentation

Difficulty: medium

Interview Round: Other

You are the product manager for Google Docs, and the metric that counts document opens has fallen by 10%. How would you investigate the decline, and what would you do once you know the cause? ```hint Decompose the count An open count can fall because fewer people use Docs or because each person opens fewer documents, and opens arrive through several entry points. Split the number along those lines before guessing at causes. ``` ### Clarifying Questions - What exactly counts as an open: any load of a document, one load per user per day, or only a load in the editor rather than a preview? - Over what period did the 10% drop happen, and was it a sudden step or a gradual slide? - Is the drop on the web, in the mobile apps or both, and in consumer accounts, business Workspace domains or education? - Has anything shipped recently in Docs, Drive or Gmail that changes how people reach a document? ### What a Strong Answer Covers - Pinning down the definition, size and timing of the drop before diagnosing - Ruling out logging, definition and seasonal effects before product causes - Decomposing opens into users, opens per user and entry points, then segmenting by platform, account type and region - Ordered internal and external hypotheses, each with the data that would confirm it - Judging whether fewer opens actually hurt users or the business, and the response that follows from the cause ### Follow-up Questions - Opens fell, but editing time and documents created are flat. Is this a problem, and would you change the metric? - The drop is concentrated in education accounts. What would you check next? - Suppose a new AI summary in Drive answers users' questions without opening the file. How would you weigh that against the lost opens?

Overview: A metric diagnosis question in which document opens for Google Docs fall by 10%. It tests clarifying the metric and timing, ruling out logging and seasonal effects, decomposing and segmenting the decline, ranking hypotheses with checks, and judging whether fewer opens is truly a problem.

Google logo
Google
Sep 13, 2025
mediumProduct ManagerOtherAnalytics & Experimentation
0
0

You are the product manager for Google Docs, and the metric that counts document opens has fallen by 10%. How would you investigate the decline, and what would you do once you know the cause?

Clarifying Questions Guidance

  • What exactly counts as an open: any load of a document, one load per user per day, or only a load in the editor rather than a preview?
  • Over what period did the 10% drop happen, and was it a sudden step or a gradual slide?
  • Is the drop on the web, in the mobile apps or both, and in consumer accounts, business Workspace domains or education?
  • Has anything shipped recently in Docs, Drive or Gmail that changes how people reach a document?

What a Strong Answer Covers Guidance

  • Pinning down the definition, size and timing of the drop before diagnosing
  • Ruling out logging, definition and seasonal effects before product causes
  • Decomposing opens into users, opens per user and entry points, then segmenting by platform, account type and region
  • Ordered internal and external hypotheses, each with the data that would confirm it
  • Judging whether fewer opens actually hurt users or the business, and the response that follows from the cause

Follow-up Questions Guidance

  • Opens fell, but editing time and documents created are flat. Is this a problem, and would you change the metric?
  • The drop is concentrated in education accounts. What would you check next?
  • Suppose a new AI summary in Drive answers users' questions without opening the file. How would you weigh that against the lost opens?