![]() |
|
|
|||||||
| Home | Register | FAQ | Classifieds | Arcade | Radio and TV | Members List | Search | Today's Posts | Mark Forums Read |
| Announcements & News For announcements of what's happening with the Maldiviandigital.com forum and important news. |
|
|
LinkBack | Thread Tools | Display Modes |
|
#1
|
||||
|
||||
|
Web sites that automatically customize themselves for each visitor so they come across as more appealing or simply less annoying can boost sales for online businesses by close to 20%, MIT research says.
These sites adapt to display information so everyone who visits sees a version best suited to their preferred style of absorbing information, say the four researchers who write about such sites in "Website Morphing", a paper being published this month in Marketing Science . So the site might play an audio file and present graphics to one visitor, but present the same information as text to the next depending on each person's cognitive style. Morphing sites deduce that style from the decisions visitors make as they click through pages on the site. "You need five to 10 clicks before you can really get a pretty good idea of who they are," says John Hauser, the lead author of the paper. He says over the past decade statistics have evolved to allow broader conclusions from less data. "You can infer a lot more from a lot less data by borrowing data from other respondents," he says. "When I first heard it I thought this couldn't possibly work." But it does. By using a sample set of users navigating a test Web site, individual businesses can set the baseline for what click choices on that site mean about the visitor. Over time with real potential customers visiting a live site, the morphing engine fine tunes itself to draw better conclusions about visitors' preferences and to serve up what pages most likely lead to a sale, Hauser says. The software is open source and available at MIT's Web site, but so far no one has created a commercial business to apply it to individual customers, he says. Such auto-customizing Web sites are less intrusive than the alternative - sites that visitors can manually customize, a time-consuming process that many visitors won't bother with, the researchers say. And they create the right Web site for maximum sales much quicker, Hauser says. They tested out their theory on 835 current U.K. broadband users or those interested in broadband who were paid to fill out questionnaires and navigate a sample BT Web site for broadband services. They concluded that if the site had perfect information about each person's cognitive style, it could boost sales 21%. Even with partially known cognitive styles, the site could boost sales nearly 20%, they say. If applied to active Web sites, morphing could add $80 million to BT's online broadband sales, the researchers say. Morphing changes the look and feel of sites, not just content, to suit cognitive styles. For example the balance of pictures and text can be adjusted to appeal to particular users. The BT experiment assessed subjects' cognitive styles based on four different cognitive-style characteristics, each having two options. The cognitive styles defined whether individuals were readers or listeners, impulsive or deliberative, visual or verbal and leaders or followers. This assessment was based on their click stream data. When visitors arrived at the site, they faced four options: comparing plans, visiting a virtual adviser, visiting a learning center and visiting a community of users. "We observe some number of clicks (say 10), infer probabilities for the visitor's cognitive style, then morph the Web site based on our inference of the visitor's cognitive style," the study says. "The visitor continues until he or she either purchases a broadband service or exits the Web site without purchasing." The visitors were then presented with sites that could vary based on three characteristics: graphs and pictures vs. text and audio; small or large amounts of information presented; focused content with a few options vs. general content with all options presented. Visitors rated as holistic received a morph presenting general information about BT, while analytic visitors received an analysis tool that manipulates information. Followers received a link to comments by another consumer; deliberative visitors linked to a page where they could post comments. The BT study found general trends among the learning-style groups. Readers and those rated verbal prefer focused content. Impulsive people tend to prefer smaller chunks of data. Analytic visitors prefer comparing options vs. receiving general information. Hauser says he has applied this methodology to Web sites for the Japanese bank Seruga. Rather than morphing the site to suit just cognitive style, his team morphed it to present information in formal or informal language depending on what visitors' click patterns revealed about their cultural preference. |
| The Following User Says Thank You to ahmedvisham For This Useful Post: | ||
AngelEye (27-05-2008)
| ||
| To Advertise Us | |
|
|
|
|