Lengthening Instagram Stories' lifetime: decision metrics, guardrails and test design

Quick Overview

A product analytics question about whether Instagram should keep stories visible for longer before they expire, and which metrics should drive that call. It tests reasoning about effects on posters, viewers and the wider app, choosing decision metrics with guardrails, and designing an experiment where viewers and posters influence each other.

Lengthening Instagram Stories' lifetime: decision metrics, guardrails and test design

Company: Meta

Role: Product Manager

Category: Analytics & Experimentation

Difficulty: medium

Interview Round: Other

You are the product manager for Instagram Stories. A story currently disappears from viewers' feeds 24 hours after it is posted, and you are weighing whether to let stories stay up longer before they expire. Which metrics would you look at to make this decision? ```hint Who gains and who loses A longer lifetime affects the people who post, the people who watch and the rest of the app in different ways. For each group, think about what could improve, what could get worse and which other parts of Instagram compete for the same attention. ``` ### Clarifying Questions - What prompted the idea: data showing that many stories expire before followers see them, requests from people who post, or something else? - Which longer duration is under consideration, and would it apply to every story or be a choice the person posting makes? - Is the change aimed at everyone or at a segment, such as business accounts or people who open the app only every few days? - Does anything else change, such as how expired stories are kept in the poster's archive and highlights? ### What a Strong Answer Covers - The goal behind the change and a clear hypothesis for why a longer lifetime would serve it. - The likely effects on posters, viewers and the rest of Instagram (feed posts, highlights, ads), with metrics for each. - A short list of decision metrics and guardrails, with the result that would justify shipping. - An experiment design that compares more than one duration and handles the fact that posters and viewers affect each other. - Second-order risks, including what people become willing to post, how fresh the story tray feels and what it costs to store and serve stories. ### Follow-up Questions - The test shows more views per story but fewer stories posted per person. What do you recommend? - Viewers in the treatment group still watch stories from posters in the control group. How would you set up the experiment so the results are trustworthy? - Would you rather offer a longer lifetime as a setting the poster chooses? How would that change what you measure?

Overview: A product analytics question about whether Instagram should keep stories visible for longer before they expire, and which metrics should drive that call. It tests reasoning about effects on posters, viewers and the wider app, choosing decision metrics with guardrails, and designing an experiment where viewers and posters influence each other.

Meta logo
Meta
Jun 2, 2024
mediumProduct ManagerOtherAnalytics & Experimentation
0
0

You are the product manager for Instagram Stories. A story currently disappears from viewers' feeds 24 hours after it is posted, and you are weighing whether to let stories stay up longer before they expire. Which metrics would you look at to make this decision?

Clarifying Questions Guidance

  • What prompted the idea: data showing that many stories expire before followers see them, requests from people who post, or something else?
  • Which longer duration is under consideration, and would it apply to every story or be a choice the person posting makes?
  • Is the change aimed at everyone or at a segment, such as business accounts or people who open the app only every few days?
  • Does anything else change, such as how expired stories are kept in the poster's archive and highlights?

What a Strong Answer Covers Guidance

  • The goal behind the change and a clear hypothesis for why a longer lifetime would serve it.
  • The likely effects on posters, viewers and the rest of Instagram (feed posts, highlights, ads), with metrics for each.
  • A short list of decision metrics and guardrails, with the result that would justify shipping.
  • An experiment design that compares more than one duration and handles the fact that posters and viewers affect each other.
  • Second-order risks, including what people become willing to post, how fresh the story tray feels and what it costs to store and serve stories.

Follow-up Questions Guidance

  • The test shows more views per story but fewer stories posted per person. What do you recommend?
  • Viewers in the treatment group still watch stories from posters in the control group. How would you set up the experiment so the results are trustworthy?
  • Would you rather offer a longer lifetime as a setting the poster chooses? How would that change what you measure?