Metrics question

There has been a sudden spike in partial order deliveries by Zomato. How would you solve it?

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 operational root-cause diagnosis of a quality metric spike, tracing the failure back to the specific point in the delivery chain.

Clarify what counts as partial: items missing from an order, whether from the restaurant's packing or lost during the delivery handoff.
Segment the spike by restaurant, delivery partner and region to see if it's concentrated or systemic across the whole platform.
Hypothesize restaurant-side causes: packing errors during high-volume rush hours, or a new restaurant onboarding without proper packing protocols.
Hypothesize delivery-side causes: bag mishandling, multiple order batching leading to mix-ups, or delivery partners confirming pickup without checking completeness.
Cross-reference the spike's timing against any recent operational change, a new batching feature, a packaging policy change, or an onboarding surge.

Separates restaurant-side packing errors from delivery-side handling errors as the first diagnostic split, since the fix differs completely for each.
Considers order batching specifically as a plausible cause, a real operational lever Zomato controls and could have recently changed.
Uses segmentation by restaurant and partner to isolate whether this is concentrated or platform-wide.

Treating all partial deliveries as one undifferentiated problem without splitting by restaurant versus delivery-partner cause.
Jumping to a blanket fix, like penalizing all delivery partners, before segmenting to find the actual source.

How would you design a checklist step to verify order completeness before pickup?
How would you compensate customers fairly while root cause is still being fixed?