weblog/wEssays     archives     home
 

The Popularity of Consensus (or, how wonderful, you agree with me)   (October 14, 2006e)


One of the most popular features of the Internet is the sharing of users' reviews. I've relied on other consumers' opinions for wireless routers, DVD players, Las Vegas buffets, hotels in China and elsewhere--and I'm sure you have your own list.

So popular is this "share your opinion" format that it has spread to ranking the news (digg.com) and sex workers in illegal brothels. According to a recent San Francisco Chronicle series on the trafficking of young women into America, Diary of a Sex Slave, call-girls can be ruined by poor online reviews just as effectively as restaurants or other services.


Despite the appeal of "regular folks" reviews, the trend begs the question: are they actually accurate or useful? Or are they simply popularity contests? Are the top news stories on digg.com more important or more interesting than those chosen by editors at the Wall Street Journal or other major media publication? Are the rankings on digg.com "gamed" like the rankings on google.com? And if they can be (yes, they can), then exactly what is their validity?

As a reader and reviewer on Amazon.com, I've noticed an unsettling trend. On amazon, not only can readers/viewers/consumers write an online review, but other users can rank the "helpfulness" of each online review. I've noticed that unfavorable reviews are judged unhelpful, while rave reviews draw nearly unanimous votes of "yes, this was helpful."

You can go to my own reviews via my Amazon.com profile, or simply click on a couple of other reviewers' entire lists of reviews. Though this is only an anecdotal survey, it seems like even thoughtfully negative reviews draw far fewer user votes than glowing reviews, and at least half of those votes are rejections--that is, "no, this review was not helpful to me."

So what does this say about online reviews? That we approve of everyone who agrees with us, and vehemently reject opposing opinions. This is unsurprising, but there are some other trends visible in this online opinion gathering. The most "popular" reviews (those with the most user votes on helpfulness) are pushed to the top, where they gather even more votes. Those reviews with less votes get buried in back pages where they will languish unread. Since reviews which are somewhat negative receive less votes, they are shoved to the back of the queue along with mediocre ("unhelpful") positive reviews.

You see where this is going--we are drawn to reject opinions which differ from our own, and drawn to agree with the most popular opinions. In other words, we do not like views which challenge our belief systems, and despite our professed "objectivity," we are easily drawn into a herd mentality.

Have online reviews been helpful to me? It's a mixed bag. The technology reviews have been complete busts. One guy rates the product zero, the next, a five (top rating). Reviews are wrong about product features, and don't explain what might have caused the product to fail (like, computer is unplugged, computer is loaded with trojan horse viruses, etc.)

I have found film reviews on amazon to be generally more enlightening than those written by most professional reviewers. What's lacking in most reviews, of course, is the context of the film and a reasonable attempt to critique it rather than simply praise it. But a sharp critique will guarantee oblivion for your Amazon review, so the majority of reviewers write extremely positive reviews.

Amazon has some sort of algorithm which ranks reviewers (yet another ranking system), presumably by the number of their reviews and their popularity. With 22 reviews and 76 "helpful" votes out of 83 total, I am ranked number 39,561 of Amazon reviewers. (Whoopie.) "Top 10" reviewers have posted hundreds of reviews--how they do it is beyond my feeble comprehension.

Isn't the stock market a similar "user ranking system," in which "popular choices" are purchased and hence rise in value, and views opposite our own beliefs are largely dismissed? It pays to read what you don't agree with, and look at evidence which isn't swayed by the "helpfulness votes" of others: data and charts.

Curious about past entries? Either check my updated archives or use the nifty "search" box in the right sidebar.



For more on this subject and a wide array of other topics, please visit my weblog.

                                                           


copyright © 2006 Charles Hugh Smith. All rights reserved in all media.

I would be honored if you linked this wEssay to your site, or printed a copy for your own use.

                                                           


 
  weblog/wEssays     home