Identify User Interest in Group Video Calls Using Data

Quick Overview

Evaluates how to use one-to-one call data to identify demand for group video calls. Strong answers define the business goal, find likely adopters, reason about participant limits, measure incremental engagement, and design network-effect-aware experiments with quality and cost guardrails.

Identify User Interest in Group Video Calls Using Data

Company: Meta

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: hard

Interview Round: Technical Screen

### Group Video-Calling Feature Analysis #### Scenario Designing and analyzing a new group video-calling feature using historical one-to-one video call data. #### Key Questions ##### 1. Business Goal What is the business goal of this feature? ##### 2. User Identification & Data Requirements How would you identify which users are most interested in group video calls and what additional data would improve the analysis? ##### 3. Participant Limit Analysis Should we impose a participant limit on group calls? Explain how you would determine the optimal cap. ##### 4. Success Metrics & Measurement Which success metrics would you track after launch and how would you measure cannibalization versus incremental engagement? ##### 5. Experiment Design [Bonus] How would you design the experiment? #### Important Considerations ##### Hints - **Network Effect**: You need to consider the network effect when designing the experiment

Overview: Evaluates how to use one-to-one call data to identify demand for group video calls. Strong answers define the business goal, find likely adopters, reason about participant limits, measure incremental engagement, and design network-effect-aware experiments with quality and cost guardrails.

Community answers

Answer by Xiaoming

One-to-one call logs reveal heavy communicators, overlapping contact networks, and geographic clusters that indicate group call demand. Additional data on group structures, infra costs, and competitor practices refine product decisions. A participant cap is necessary; initial limit ~8–12 is reasonable. After launch, measure adoption, engagement lift, and quality, and run controlled experiments to distinguish incremental value vs. cannibalisation.

Answer by SS

User problem Messenger groups are currently optimized for asynchronous coordination (text, reactions). However, some group use cases require real-time, multi-party communication, such as: Event planning Urgent coordination High-traffic discussions that become hard to follow in chat I’d look for coordination stress signals in existing data: Message burstiness: messages per group per minute (tail groups, e.g., P75+) Repeated 1:1 calls among members of the same group in short windows Call waiting / missed calls when users are already on another call Heavy use of voice notes in large groups App switching shortly after message bursts (proxy for moving to another app)
|Home/Analytics & Experimentation/Meta
Meta logo
Meta
Jul 12, 2025
hardData ScientistTechnical ScreenAnalytics & Experimentation
176
0

Identify User Interest in Group Video Calls Using Data

You are designing and analyzing a new group video-calling feature for a large social or messaging app. Currently, you mainly have historical one-to-one video call data.

Constraints & Assumptions

  • The feature should create incremental real-time communication, not only shift users from one-to-one calls or messages.
  • Assume access to one-to-one call history, group chat metadata, invite behavior, device/network quality, retention, and experiment infrastructure.
  • Network effects matter: one user's treatment can affect other users in the same social cluster.
  • Participant limits and success metrics should account for product value, call quality, safety, and cost.

Clarifying Questions to Ask Guidance

  • Is the product already strong in one-to-one video calls, group chats, or both?
  • Are group audio calls available today?
  • What use cases matter most: family calls, work coordination, creator/community calls, or casual friend groups?
  • Are there hard constraints on participant count, device performance, or bandwidth?

Part 1 - Clarify the Business Goal

What is the business goal of the group video-calling feature?

What This Part Should Cover Guidance

  • Primary objective such as meaningful real-time communication, retention, reactivation, or competitive parity.
  • Secondary objectives and guardrails, including quality, reliability, abuse, and infrastructure cost.
  • A clear definition of incremental value.

Part 2 - Identify Interested Users

How would you identify users most interested in group video calls using one-to-one call history, and what additional data would improve the analysis?

What This Part Should Cover Guidance

  • Proxies such as frequent callers, repeated calls within group-chat clusters, sequential calls to several contacts, missed coordination, and dense social graphs.
  • Additional data from group chats, surveys, beta signups, device capability, network quality, and current substitutes.
  • Segmentation and scoring that distinguish likely adopters from users who would use any new calling surface.

Part 3 - Decide on Participant Limits

Should the product impose a participant limit, and how would you determine the optimal cap?

What This Part Should Cover Guidance

  • Demand distribution by intended group size, call success by size, quality degradation, device constraints, abuse risk, and cost.
  • Staged experiments or rollouts with different caps, if feasible.
  • A rule that covers valuable use cases while protecting reliability and user experience.

Part 4 - Measure Launch Success

Which success metrics would you track after launch, and how would you measure cannibalization versus incremental engagement?

What This Part Should Cover Guidance

  • Incremental active group callers, successful group-call sessions, call minutes, repeat use, retention, and invite/join funnel.
  • Cannibalization of one-to-one calls, messaging, and existing group surfaces.
  • Guardrails for dropped calls, join failures, latency, crashes, complaints, and cost.

What a Strong Answer Covers Guidance

A strong answer ties business goals, user targeting, participant caps, metric hierarchy, and experiment design together while accounting for network effects and cannibalization.

Follow-up Questions Guidance

  • How would you randomize an A/B test for a group feature?
  • If group calls increase total minutes but reduce message activity, how would you decide whether that is good?
  • Which user segment would you invite to a beta first?
Loading comments...

Browse More Questions

More Analytics & Experimentation•More Meta•More Data Scientist•Meta Data Scientist•Meta Analytics & Experimentation•Data Scientist Analytics & Experimentation