Estimate the share of Booking.com reservations that end up cancelled worldwide

You are interviewing for a product manager role at Booking.com. Estimate what percentage of reservations made on Booking.com around the world are eventually cancelled, and explain why the number lands where it does. ```hint Split before you size Cancellation behavior depends heavily on the rate the guest chose and how far ahead they booked. Estimate a rate for each group and weight them, instead of guessing one global number. ``` ### Clarifying Questions - Should the rate include only cancellations made by guests, or also those initiated by properties and by the platform, such as payment failures or suspected fraud? - Is the denominator all confirmed reservations made in a period, counted by booking date, and are no-shows and date changes excluded from the numerator? - Does "booking" cover every accommodation type on the site (hotels, apartments, vacation homes), with flights and car rentals out of scope? ### What a Strong Answer Covers - A precise definition of the rate: numerator, denominator, cohort and exclusions - A segmentation that actually drives cancellations, such as rate policy and booking lead time, with every input labeled as an assumption - Clear arithmetic that reaches a point estimate and a plausible range - The "why": guest, property and platform causes, including how flexible cancellation policies shape guest behavior - A sanity check from a second angle, and the inputs the answer is most sensitive to ### Follow-up Questions - Booking.com wants fewer cancellations without hurting conversion. Which lever would you test first, and how would you measure it? - How would your estimate change for a city market dominated by short-notice business trips compared with a beach destination booked months ahead? - A hotel partner complains that its cancellation rate through the platform is far higher than on its own website. How would you explain the gap, and what would you offer?