Diagnose rising Meta AI use in Instagram DMs alongside a drop in Discover search

Quick Overview

A metric diagnosis question for a Meta product manager: Meta AI usage is rising in Instagram direct messages but falling from the Discover page search bar. It tests data validation, attribution between entry points, internal and external causes, segmentation, and judging whether the shift is healthy.

Diagnose rising Meta AI use in Instagram DMs alongside a drop in Discover search

Company: Meta

Role: Product Manager

Category: Analytics & Experimentation

Difficulty: medium

Interview Round: Other

On Instagram, people can use Meta AI in direct messages and from the search bar on the Discover page. Over a recent period, Meta AI usage in DMs has gone up while usage from the Discover search bar has gone down. How would you investigate what is happening? ```hint Read the two moves together Before you split either number, ask whether this is one story or two: are the same people moving from one entry point to the other, and what has happened to total Meta AI usage on Instagram? ``` ### Clarifying Questions - How is "usage" measured (users, sessions or prompts), and is it defined the same way on both surfaces? - How large are the two changes, over what period, and did they start together or drift gradually? - Is total Meta AI usage across Instagram up, flat or down? - Has anything shipped recently on either surface, such as a search bar redesign or a new way to reach Meta AI from the inbox? ### What a Strong Answer Covers - Validating the data first: metric definitions, logging, and how usage is attributed when a conversation moves between the two entry points - Testing whether this is one shift (the same users moving) or two independent changes, using total usage and user-level tracking - Internal causes (launches, experiments, triggering logic, answer quality, bugs) and external causes (seasonality, news, regional changes), checked in a sensible order - Segment cuts by platform, app version, region, age group, user tenure and query intent to localize the change - A conclusion on whether the shift is healthy or harmful, and next steps that also protect regular search on the Discover page ### Follow-up Questions - You find that the same users moved from the search bar to DMs and total usage is flat. Is that good or bad, and what would you do with the search bar? - The drop is concentrated on one platform's newest app version. What do you check next? - How would you tell whether people are skipping AI answers in the search bar because they want regular search results instead?

Overview: A metric diagnosis question for a Meta product manager: Meta AI usage is rising in Instagram direct messages but falling from the Discover page search bar. It tests data validation, attribution between entry points, internal and external causes, segmentation, and judging whether the shift is healthy.

Meta logo
Meta
Mar 31, 2026
mediumProduct ManagerOtherAnalytics & Experimentation
0
0

On Instagram, people can use Meta AI in direct messages and from the search bar on the Discover page. Over a recent period, Meta AI usage in DMs has gone up while usage from the Discover search bar has gone down. How would you investigate what is happening?

Clarifying Questions Guidance

  • How is "usage" measured (users, sessions or prompts), and is it defined the same way on both surfaces?
  • How large are the two changes, over what period, and did they start together or drift gradually?
  • Is total Meta AI usage across Instagram up, flat or down?
  • Has anything shipped recently on either surface, such as a search bar redesign or a new way to reach Meta AI from the inbox?

What a Strong Answer Covers Guidance

  • Validating the data first: metric definitions, logging, and how usage is attributed when a conversation moves between the two entry points
  • Testing whether this is one shift (the same users moving) or two independent changes, using total usage and user-level tracking
  • Internal causes (launches, experiments, triggering logic, answer quality, bugs) and external causes (seasonality, news, regional changes), checked in a sensible order
  • Segment cuts by platform, app version, region, age group, user tenure and query intent to localize the change
  • A conclusion on whether the shift is healthy or harmful, and next steps that also protect regular search on the Discover page

Follow-up Questions Guidance

  • You find that the same users moved from the search bar to DMs and total usage is flat. Is that good or bad, and what would you do with the search bar?
  • The drop is concentrated on one platform's newest app version. What do you check next?
  • How would you tell whether people are skipping AI answers in the search bar because they want regular search results instead?