U.C. Study Reveals Feedback Flaws at eBay
Saturday, January 13, 2007
Researchers wade into "feedback underworld"
"Some eBay users may not deserve the stellar reputations showcased in their member profile."
That terse sentence is the lead in a story that ran in Friday's San Francisco Chronicle reporting on the findings of a U.C. Berkeley study into eBay's feedback system. The story is the latest in a spate of news accounts detailing academic research into the flaws in one of the lynchpins the auctioneer's success.
The study finds what many eBayers already know too well: feedback can be fudged. How that fudging occurs may be news to you.
According to the paper, the U.C. study made a link between the virtually infinite number of listings for "seemingly valueless items" on eBay and schemes to inflate feedback. These nickel and dime items are listed "solely to manipulate feedback ratings. Buyers and sellers who complete a sale -- even if for only 1 cent -- can leave feedback for each other."
In many cases, sellers expressly spelled out in their listings that the buyer should leave a positive feedback for the seller, according to John Morgan, a business and economics professor who co-authored the study with Jennifer Brown, a graduate student. In return, the sellers said they would leave one for the buyer, Morgan told the paper.
The story "Skewed Ratings on eBay - UC Study Finds Some Manipulate Their Positive Feedback" also chronicles how the authors of the study were able to wade into the "feedback underworld" to buy a "positive feedback e-book." Not exactly the kind of "True Crime" grist that's going to rouse the ghost of Raymond Chandler, but the implications are serious and considerable.
In recent weeks, ATR has run a number of stories related to the issue of bogus feedback and manipulated reputation scores. On December, 17 we reported on Alice LaPlante's missive that asked if eBay's feedback system is "fatally flawed." And on January 5th we put you in touch with the NetProbe software developed by Carnegie Mellon University designed to data-mine for "false positives": the high feedback scores that result when crooks do e-business with themselves.
On behald of eBay, spokesperson Catherine England said the company has a large staff dedicated to fight this kind of fraud and takes in very seriously.
"Is any system 100 percent effective? No," England said. "But after 10 years, eBay has seen most of the ideas out there, and we have automated tools that keep our site safe."
Though all of this research is interesting and illuminates the ways in which eBay's feedback system - and online reputation programs more broadly - can be minipulated, keep in mind this nugget from eBay's own feedback tutorial:
"Each member may affect your score by only one point (positive or negative). However, they may leave you one feedback rating and comment for each transaction they have with you."
Sure, crooks and scammers invent new user I.D.'s, but the system really is designed to try to prevent a fraudster from building a bona fide rep. via an avalanche of positive feedback from a small group of accomplices.
"Some eBay users may not deserve the stellar reputations showcased in their member profile."
That terse sentence is the lead in a story that ran in Friday's San Francisco Chronicle reporting on the findings of a U.C. Berkeley study into eBay's feedback system. The story is the latest in a spate of news accounts detailing academic research into the flaws in one of the lynchpins the auctioneer's success.
The study finds what many eBayers already know too well: feedback can be fudged. How that fudging occurs may be news to you.
According to the paper, the U.C. study made a link between the virtually infinite number of listings for "seemingly valueless items" on eBay and schemes to inflate feedback. These nickel and dime items are listed "solely to manipulate feedback ratings. Buyers and sellers who complete a sale -- even if for only 1 cent -- can leave feedback for each other."
In many cases, sellers expressly spelled out in their listings that the buyer should leave a positive feedback for the seller, according to John Morgan, a business and economics professor who co-authored the study with Jennifer Brown, a graduate student. In return, the sellers said they would leave one for the buyer, Morgan told the paper.
The story "Skewed Ratings on eBay - UC Study Finds Some Manipulate Their Positive Feedback" also chronicles how the authors of the study were able to wade into the "feedback underworld" to buy a "positive feedback e-book." Not exactly the kind of "True Crime" grist that's going to rouse the ghost of Raymond Chandler, but the implications are serious and considerable.
In recent weeks, ATR has run a number of stories related to the issue of bogus feedback and manipulated reputation scores. On December, 17 we reported on Alice LaPlante's missive that asked if eBay's feedback system is "fatally flawed." And on January 5th we put you in touch with the NetProbe software developed by Carnegie Mellon University designed to data-mine for "false positives": the high feedback scores that result when crooks do e-business with themselves.
On behald of eBay, spokesperson Catherine England said the company has a large staff dedicated to fight this kind of fraud and takes in very seriously.
"Is any system 100 percent effective? No," England said. "But after 10 years, eBay has seen most of the ideas out there, and we have automated tools that keep our site safe."
Though all of this research is interesting and illuminates the ways in which eBay's feedback system - and online reputation programs more broadly - can be minipulated, keep in mind this nugget from eBay's own feedback tutorial:
"Each member may affect your score by only one point (positive or negative). However, they may leave you one feedback rating and comment for each transaction they have with you."
Sure, crooks and scammers invent new user I.D.'s, but the system really is designed to try to prevent a fraudster from building a bona fide rep. via an avalanche of positive feedback from a small group of accomplices.
posted by Chris and Frank at 6:45 PM
0 Comments:
Post a Comment