Agoda: measuring the net effect of cancellation charges on bookings and revenue

You are a product manager at Agoda, an online travel agency where travelers book hotels and other stays. Your task is to measure the impact of cancellation fees, the charges a traveler can incur for cancelling a booking. Walk through which outcomes you would track, how you would separate the effect of the fees from everything else that moves bookings, and how you would put a rough number on the net effect. ```hint Look past the cancellation itself List every party and every stage of the booking journey where a fee could change behavior, then decide which effect the business actually cares about before you pick a headline metric. ``` ### Clarifying Questions - Is the decision about introducing a fee on bookings that are free to cancel today, raising an existing fee, or auditing the fees already in place? - Who sets and keeps the fee: the property through its own rate terms, Agoda, or a split between them? - Over what period should impact be judged, given that many stays are booked weeks before check-in and a booking's fate is known only then? - Does the scope cover every market and property type, or a particular segment such as one region or one kind of stay? ### What a Strong Answer Covers - A success measure built on net value (completed stays and revenue per visit), not the cancellation rate alone. - Effects mapped across the booking funnel and both sides of the marketplace: conversion, rate choice, cancellation timing, no-shows, fee income, property partners and traveler loyalty. - A causal design (a randomized test where the policy can be varied, otherwise staggered rollout or comparison markets) that handles seasonality and the long lag between booking and stay. - A rough sizing of the net effect with every assumption labeled, a break-even point and a sanity check. - Guardrails for support contacts, payment disputes, repeat bookings and partner satisfaction. ### Follow-up Questions - Your test shows fewer cancellations but also fewer bookings. How do you decide whether the fee is worth keeping? - Travelers in the test can see a rival site offering free cancellation on the same hotel. How does that change your design or your reading of the results? - How would you measure a loss of loyalty that will not show up within the few weeks of a test?