As previously mentioned, a challenge in measuring the influence
of hotel sales is that room sales data are not available to
researchers. We could not observe actual hotel bookings from
travel agent websites, as these are the private data of the individual
hotels concerned. Previous studies have used a variety of proxies to
infer product sales from other observed data. When selecting a
proxy for this study, we noted the policy of Ctrip (the online travel
agency from which we obtained the data) that only customers who
have successfully made online bookings are allowed to post a
review (one per transaction) after their stay. Therefore, the number
of published reviews on a travel website might be taken as a proxy
for sales of a hotel’s rooms through the travel agent during the
study period. Assuming customers have a constant probability of
posting a review on the travel agent website, the number of online
reviews for hotel i can be expressed as a linear function of room
sales of the hotel, where number Reviews ¼ / e sales. Essentially,
we can take the number of reviews as a proxy for the number of
bookings for the hotel and use the following regression model for
the analysis.