Customer targeting: Your computer’s operating system may give you away.

Printer-friendly versionSend to friendPDF version

Orbitz, the US travel chain, varies the hotels offered to customers on its website depending on the operating system of their computer. Dana Mattioli discusses the implications of this new approach to targeting customers in the “Wall Street Journal”.
 
Individuals who access the Orbitz website looking for a hotel from a Mac computer are offered other, often costlier, options than those logging onto the website from a computer with a Windows operating system. This approach followed a data analysis of booking behavior. It showed that Mac users spend on average 20 to 30% more a night and are 40% more likely to book into a hotel with four stars or better. Prices of the individual hotels do not vary, but in some comparisons the average price over all hotels displayed on the first page of the search results was more than 10% higher.
 
The procedure is a good example of the benefits, and for some the potential dangers, of data mining. Apple users have a reputation as “big spenders” with justification. Forrester Research estimates the household income of Mac users at about US$ 99,000 compared to about US$ 75,000 for PC users. As with individualized news and information systems the obvious benefit is to be closer to the customer and anticipate wishes. The question of course, is firstly if customers like to be classified this way, secondly if they are aware of it, and lastly if such a classification has disadvantages. In this example everybody gets the same results sorting e.g. by price. In other applications however, customers might object if they are singled out especially if they are unaware of the criteria used. What is targeting for some, is perceived as discrimination for others.