Quantify Latent Demand for Group Video Calling Feature
Overview
This question evaluates a data scientist's competency in product analytics, causal experimentation, metric definition, and trade-off analysis for product launch decisions, with emphasis on network-aware thinking and estimating latent demand from observational signals.
Meta
Aug 4, 2025, 10:55 AM
Data Scientist
Technical Screen
Analytics & Experimentation
0
0
Scenario
A consumer messaging app is preparing to launch group video calling. You have access to rich data: message logs (timestamps, thread IDs), friend/interaction graphs, 1:1 call logs, and user surveys.
Task
Design an analytics plan that:
- Quantifies latent demand for group video calling using existing data (messages, friend graphs, surveys).
- Defines the primary success metrics for launch.
- Proposes an experimental design to evaluate impact while handling network interference.
- Recommends a data-driven initial cap on the number of participants allowed in a group call, considering user coverage versus engineering/quality costs.
Hints
- Consider interaction clustering (e.g., communities, group threads), A/B tests with guardrails, retention/attach rate, percentile analysis of desired group sizes, and the trade-off between engineering cost/quality and user coverage.
Solution
Login required
Show