Estimation question

Estimate the memory needed for Nest cameras for the fiscal year 2020.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests structured estimation of cloud storage requirements using device count, recording behavior, and video compression assumptions.

State the goal: estimate the total cloud storage needed to support Nest camera video for fiscal year 2020.
Estimate active Nest camera installations: assume roughly 10 million active Nest cameras in use globally during that period.
Estimate recording behavior: assume each camera records an average of 4 hours of motion-triggered footage per day, since most plans use event-based, not continuous, recording.
Estimate storage per hour of video: assume compressed video at roughly 1 GB per hour at typical Nest streaming resolution.
Compute daily storage: 10 million cameras times 4 hours times 1 GB, about 40 petabytes of new footage generated per day.
Adjust for a typical 30-day rolling retention window, giving roughly 1.2 exabytes of stored footage at any time, then sanity-check against known cloud storage scale for comparable services.

Builds the estimate from concrete device, usage, and compression assumptions rather than guessing a storage figure directly.
Correctly assumes event-triggered recording rather than continuous 24/7 footage, reflecting how Nest's actual product works.
Incorporates a retention window explicitly, recognizing total storage depends on how long footage is kept, not just daily generation.
Sanity-checks the final exabyte-scale figure against known cloud storage scales for comparable consumer video services.

Assuming continuous 24/7 recording for every camera, drastically overestimating storage needs versus Nest's actual event-based model.
Computing only daily generation without applying a retention window, missing the actual total storage requirement.

How would this estimate change for users on a continuous-recording plan versus the standard event-based plan?
How would video compression improvements over time affect this estimate?