Estimation question

Estimate the number of Whatsapp chats happening in India on any given day?

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

Estimation and structured reasoning: can you build a clean top-down model with sensible assumptions rather than guessing a number.

State the approach: population of India, times WhatsApp penetration, times daily active rate, times chats sent per active user.
Assume India's population is about 1.4 billion, and WhatsApp penetration among smartphone users is roughly 500 million users (state as assumption).
Assume daily active rate of about 80% for a habitual messaging app, giving roughly 400 million DAU.
Assume an average of 30-40 one-on-one and group chats (distinct conversations, not messages) per active user per day.
Multiply to get an estimate, e.g. 400M x 35 = ~14 billion chats/day, and sanity check against message volume figures WhatsApp has published.
State what would refine the estimate, such as splitting urban vs rural usage or personal vs business accounts.

Clearly separates 'chats' (conversations) from 'messages' (individual texts), since the question is ambiguous and a strong candidate calls that out.
Uses round, defensible numbers and labels every assumption explicitly rather than presenting them as facts.
Sanity-checks the final number against an order-of-magnitude reference point, like known global WhatsApp message volume.
Shows the math cleanly rather than jumping straight to a final number.

Conflating messages and chats, which changes the answer by an order of magnitude.
Picking numbers with no stated reasoning, so the interviewer can't follow the logic.

Chapters of the AI PM course that teach what this question tests.