There are several definitions for both personalization and customization that vary per scholar. A definition given by Blom (2000) in the article describes the process of personalization as one that "changes the functionality, interface, information content, or distinctiveness of a system to increase its personal relevance to an individual." Systems that automatically personalize information acquire information about user behavior data in overt and covert ways. In the over method, data is gathered directly from questions to the user asking for information such as name, gender, phone numbers, zip codes, etc. In the covert method, the systems go "undercover" by placing cookies in user browsers and observing online behavior. You may have experienced such a method yourself when logging into a website that greets you by your name and presents information (such as weather or movie listings) based on your location or even sends you deals on products/services based on previous purchases on the web.
Customization, however, involves an active user rather than a passive one. Unlike personalization that tailors content for you, customizable systems put the user in the "driver's seat." It features "a number of affordances that allow users to make changes to the form and content of the interfaces," according to the article. Users can customize from something as simple as font or a color scheme on Web Changes to more advanced customization such as selecting what news or features of a website you'd like to know more about. The user is able to define their own desires and needs on the web for themselves unlike personalization that does the defining for them.
Preferences for Customization over Personalization Increase with Usage
In a study analyzed within the article, non power users showed more positive attitudes toward web content provided by the personalized condition rather than the customizable condition and control condition (featuring websites not influenced by customization or personalization). The study defines "power users" as "highly self-motivated learners who commit greater effort to discovery and experience frustration if restricted or given little learning autonomy." Nonpower users were defined as users that "lack the expertise and interest in adopting newer technologies and interface features." Power usage was determined by asking the participants a series of questions dealing with their technology usage.
After exposing participants to a Google Website featuring SIP, UIC, or the "control" conditions, participants were asked "how well each of the following words described the news content on the site: accurate, believable, biased, clear, comprehensive, factual, fair, informative, important, objective, persuasive, sensationalistic, and well-written." The items were rated on a Likert-type scale between ‘‘describes very poorly’’ and ‘‘describes very well." The data found that UIC > SIP for power users, and SIP > UIC for nonpower users. This two-way interaction provided support to the "H2" hypothesis that predicted a moderating effect of power usage.
User Preferences for Customized and Personalized Information Correlate with Level of Privacy
After the first study was completed, a follow-up study was conducted that eliminated the control condition and added a privacy manipulation to the experiment. After the first session where the first experiment was replicated, a second session took place that included the privacy manipulation. Participants exposed to the low-privacy condition read a statement that said "The News Website that you are about to browse MAY USE your browsing information to provide the services you’ve requested," and that the site ‘‘may share aggregated non-personal information with third parties outside of the company.’’ Participants in the high-privacy condition were instead exposed to a statement that read ‘‘The News Website that you are about to browse WILL NOT use your browsing information to provide the services you’ve requested,’’ with an assurance that said the site ‘‘will not share aggregated nonpersonal information with third parties outside of the company.’’
After surveying the participants with a similar ten-point scale following the experiment, it was found that perceived control and convenience increased with power usage in the low-privacy condition with UIC, but decreased with power usage in SIP. For high-privacy participants, perceived control and convenience increased with power usage with no significant differences between UIC and SIP. This can be explained by the need power-users have to feel in control that they can do with customization rather than personalization. However, for nonpower users, they showed more control under the personalization condition which implied that the low-privacy manipulation may have adversely affected their interaction with the website. For the high-privacy condition, there was no significant difference. Power users did show a gradual preference for SIP over UIC which may be as a result of increased convenience. Strangely, nonpower users prefferred UIC over SIP, suggesting that the high-privacy assurance actually serves as an enables positive attitudes toward the site’s ability to let users act as "active content gatekeepers."
No comments:
Post a Comment