Fermi Estimate of India's Nationwide Highway Toll Revenue per Day

Quick Overview

A product manager estimation case that asks you to size the total toll money collected in a single day across every toll plaza in India. It tests how you structure a Fermi estimate, pick defensible assumptions about tolled roads, traffic and vehicle mix, keep the arithmetic straight, and sanity-check the result.

Fermi Estimate of India's Nationwide Highway Toll Revenue per Day

Company: CARS24

Role: Product Manager

Category: Product / Decision Making

Difficulty: medium

Interview Round: Other

During a product manager interview at CARS24, you are asked to estimate the total amount of toll money collected in one day across all of the country's toll plazas. Take the country to be India, CARS24's home market, unless the interviewer says otherwise. Arrive at a single figure in rupees for a typical day, build it from parts you can justify one at a time, and show how you checked that the result is believable. ```hint Find what really drives the money Before picking any numbers, split the total into a few factors you can defend separately. For each one, ask which roads actually charge tolls, which vehicles actually pay, and how busy an ordinary plaza is across all 24 hours rather than only at rush hour. ``` ### Clarifying Questions - Should the estimate cover national highways only, or also state highways, expressways and city bridge or tunnel tolls? - Do you want a typical weekday, or an average day across the year including weekends and festival peaks? - Does "collected" mean fees actually paid after return-trip discounts, monthly passes and exemptions, or the full list price of every crossing? - Is a range acceptable alongside a single point estimate? ### What a Strong Answer Covers - A clear equation that splits the total into factors (such as plaza count, traffic per plaza and average fee), each estimated separately - Realistic scope: which roads are tolled and which vehicles are exempt - Traffic and fees modeled by plaza type and vehicle class rather than one flat rate - Correct arithmetic with consistent units across lakh, crore, millions and billions - At least one independent sanity check and a stated range ### Follow-up Questions - Which assumption moves your answer the most, and what data would you use to pin it down? - How would collections on a long holiday weekend differ from a normal weekday? - If plazas were replaced by charging per kilometre driven on tolled roads, how would you rebuild the estimate?

Overview: A product manager estimation case that asks you to size the total toll money collected in a single day across every toll plaza in India. It tests how you structure a Fermi estimate, pick defensible assumptions about tolled roads, traffic and vehicle mix, keep the arithmetic straight, and sanity-check the result.

CARS24 logo
CARS24
Jun 12, 2026
mediumProduct ManagerOtherProduct / Decision Making
0
0

During a product manager interview at CARS24, you are asked to estimate the total amount of toll money collected in one day across all of the country's toll plazas. Take the country to be India, CARS24's home market, unless the interviewer says otherwise.

Arrive at a single figure in rupees for a typical day, build it from parts you can justify one at a time, and show how you checked that the result is believable.

Clarifying Questions Guidance

  • Should the estimate cover national highways only, or also state highways, expressways and city bridge or tunnel tolls?
  • Do you want a typical weekday, or an average day across the year including weekends and festival peaks?
  • Does "collected" mean fees actually paid after return-trip discounts, monthly passes and exemptions, or the full list price of every crossing?
  • Is a range acceptable alongside a single point estimate?

What a Strong Answer Covers Guidance

  • A clear equation that splits the total into factors (such as plaza count, traffic per plaza and average fee), each estimated separately
  • Realistic scope: which roads are tolled and which vehicles are exempt
  • Traffic and fees modeled by plaza type and vehicle class rather than one flat rate
  • Correct arithmetic with consistent units across lakh, crore, millions and billions
  • At least one independent sanity check and a stated range

Follow-up Questions Guidance

  • Which assumption moves your answer the most, and what data would you use to pin it down?
  • How would collections on a long holiday weekend differ from a normal weekday?
  • If plazas were replaced by charging per kilometre driven on tolled roads, how would you rebuild the estimate?