How would you detect unhealthy growth in YouTube

Problem Statement Description

Product context: Google is a consumer technology, ads, AI, and cloud company; its products include Search, YouTube, Android, Maps, Gmail, Chrome, Google Play, Workspace, and Google Cloud. YouTube is Google's video platform; its products include long-form video, Shorts, live streaming, subscriptions, YouTube Music, creator monetization, recommendations, and ads.

You are evaluating growth in YouTube at global scale, with particular attention to new internet users who may be coming online through low-cost smartphones, shared devices, prepaid data plans, emerging-market networks, or first-time Google accounts. The business is seeing usage growth, but the question is whether that growth reflects durable, valuable engagement or whether it is being inflated by low-quality, harmful, accidental, fraudulent, or unsustainable behavior.

In this metrics interview, define how you would detect “unhealthy growth” for YouTube without assuming that all growth in views, watch time, uploads, creators, or accounts is positive. Consider the different participants in the YouTube ecosystem: viewers, creators, advertisers, communities, and Google’s broader trust and safety obligations.

Your response should focus on metric definition, instrumentation, segmentation, and decision usefulness. The interviewer is looking for how you would distinguish healthy adoption from misleading growth signals, especially in a product where scale, recommendation systems, content quality, safety, monetization, and retention are tightly connected.

The experience should consider:

- What “growth” means for YouTube: users, sessions, watch time, views, uploads, creators, subscriptions, comments, revenue, or market penetration.

- Clear definitions of unhealthy growth, including spam, bots, clickbait, harmful content consumption, accidental plays, low-retention usage, policy-violating uploads, or advertiser-unsafe inventory.

- Denominators and normalization, such as per active user, per session, per creator, per geography, per device type, per traffic source, or per new-user cohort.

- Instrumentation needed to detect suspicious patterns across views, recommendations, search, comments, subscriptions, uploads, reports, removals, and monetization.

- Cohort and segment analysis for new internet users, including country, language, connectivity quality, signed-in versus signed-out usage, device type, and acquisition channel.

- Guardrail metrics around user satisfaction, retention, content quality, safety reports, creator trust, advertiser outcomes, and recommendation integrity.

- How to separate real product-market growth from measurement artifacts, policy changes, seasonality, abuse, viral spikes, or algorithmic distribution effects.

- How the metrics would help product, trust and safety, recommendations, creator, and monetization teams decide whether to investigate, intervene, or continue scaling.

Goal: Frame a metrics approach that can identify when YouTube’s growth is creating long-term user and ecosystem value versus when it is masking quality, safety, trust, or sustainability problems.