The increasing usage of the Internet over the past few decades has become an opportunity for firms to collect user information to generate. One of the many rising concerns from this trend is the potential serious invasion of user privacy, where data breaches from firms, unauthorised usage of the data collected and/or even illegal data collection without consent and permission can compromise user privacy. To reduce such threats of personal information being misused, users are starting to adopt End-User Privacy Enhancing Technologies (PETs) to guard their personal information. PETs protect user information by introducing impurities in the data being collected in terms of measurement errors and/or missing values in the data. The management of cookies is one of the most familiar privacy-related issues affecting computer users, since Web browsing has become such an essential activity, and many information sites, services, and advertising companies on the Web use cookies to track user behaviour and collect personal information. This study aims to systematically find out how does measurement errors induced by PETs such as cookies erasers affects firms’ analytical performance in a context of predictive analytics such as predicting purchases. Our simulation experiments find that adoption of cookies erasers can impact the predictive capability in the context of purchase classification significantly up to a 50% decrease in performance. From these results, we are able to identify which group of consumers firms should put more emphasis on to mitigate the negative impact of the consumer adoption of PETs. The results provide a quantitative explanation for the extent of the impact of the adoption of PETs on firms’ analytical performance in the context of a purchase classifier and can provide some generalisation in other areas of analytics conducted by firms.
