Image: Illustration by John Ueland
More In This Article
-
What a Plant Knows
How does a Venus flytrap know when to snap shut? Can it actually feel an insect’s tiny, spindly legs? And how do cherry blossoms know when to bloom? Can they...
Read More »
In the beginning, Web site owners posted words and pictures on their pages. Today we refer to that primitive time as Web 1.0.
In the modern world of Web 2.0, though, the audience provides the material. Many of the biggest names on the Web fall into this category: Facebook, eBay, Craigslist, YouTube, Flickr, and so on. In each case, the Web site owner provides nothing but a forum for strangers to connect.
One of the most fascinating tributaries of the Web 2.0 river is the citizen review site. One Web site after another harnesses the collective wisdom of thousands of delighted or disappointed customers. Never again will you make a mistake by choosing the wrong vacation spot (TripAdvisor), restaurant (Yelp), movie (IMDB), car (Edmunds), contractor (Angie’s List), app (iTunes), book (Amazon), doctor (RateMDs) or malt beverage (RateBeer).
If you are that hotel operator, restaurateur, car dealership or whatever, the rise of the citizen review site is a sobering development. No longer are you on top of the mountain, blasting your marketing message down to the masses through your megaphone. All of a sudden, the masses are conversing with one another. If your service or product isn’t any good, they’ll out you. If you are a prospective customer, on the other hand, citizen review sites seem like gifts from heaven. These days if you go to a restaurant with slow service, it’s your own darned fault. You could have avoided that outcome by consulting the masses in advance.
It makes you wonder how relevant, exactly, the solo critic is anymore. I mean, if you read a movie review in the newspaper, well, you are taking your chances. Maybe the movie critic just broke up with someone or hated the movie’s director back in film school or just doesn’t share your taste. But when you’re reading the summarized assessments of 11,000 people who have seen the same movie, it’s much harder to go wrong. The kooks on either end cancel each other out, and the big middle ground gives you a pretty accurate assessment of the movie’s real worth. (On IMDB—the Internet Movie Database—High School Musical 3 earned 3.8 out of 10 stars from the 19,600 people who have voted. This reviewer feels that’s absolutely right.)
But what about the fake-review scandals that surface with alarming regularity? Yelp, TripAdvisor, Amazon and other sites have all endured accusations that phony reviews are poisoning their posts. Big dollars are at stake. No wonder merchants, using fake names, sometimes post positive reviews for their own products or companies or trash their competitors. (Internet wags have dubbed this practice “astroturfing.” Get it? Fake grass roots?)
There are some sneaky biases at work, too. Ever notice how many apps on the iTunes Store seem to score mostly either one- or five-star reviews? How could so many apps be so polarizing?
They’re not—it’s just that online reviewers are a self-selecting bunch. You’re more likely to review something if you’re fired up about it, one way or another; the vast, quietly contented multitudes generally don’t bother.
(For a while, Apple tried to address this problem by prompting customers to rate an app at the moment they deleted it. App developers cried foul. “You’re making our reviews skew negative,” they said, “by asking this question at the moment people are deleting our apps! If they liked it, they wouldn’t be deleting it!”)
How do we maintain the power of online reviews while minimizing the abuses? For starters, we can encourage voters to use their real names, as Amazon does. Yelp and TripAdvisor say they have staffers and software dedicated to zapping bogus reviews. Yes, it’s an arms race, but review sites know that their credibility is essential to their survival.
You can improve your fraud-detection skills, too. Sometimes you can just tell when a review seems overly enthusiastic. And often you can click reviewers’ names to see what else they have written. If there aren’t any other posts, that’s a red flag.
See what we're tweeting about
7 Comments
Add CommentRandom sampling. You want to randomly pick people who bought an item and ask them to review it, that's to avoid self-selection bias. Also, you want to let people know if the number of reviews is statistically sufficient large. Most cases I know of, it isn't even remotely close by. (A book sold 100,000 copies, how many reviews do you need to have a reliable judgement for how satisfied the average customer is. Statistics basics.) It is unlikely you'll get enough randomly sampled reviews on the huge amount of things to review. The conclusion you should draw is: don't trust online reviews.
Reply | Report Abuse | Link to thisAlong with the other suggestions mentioned, perhaps site owners should still have use for that skilled solo critic, but working for their site instead if only to set a recognisibably impartial benchmark review, which would by comparison make the hyperbole of bogus, self-promoting reviews stand out more clearly. Also whether reviewers use their own names, or not, a system for grading their reviewing abilities, after they have done a number of reviews of different products, may assist others in ascertaining the credibility of the reviewer, based on their past reviews meeting a median pass quality based on a favourable comparison with the majority review grade of the particular service, site or product etc. being reviewed. The only sure way to ensure the continuity of the online review is for a vast majority of individuals to try out places based on online reviews they read then report back with their own reviews of the online reviews to see if they were credible or false. That will bulid trust in the online review, so much so that even if an owner plugs his own site, his business will at least have to meet a high percentage standard of the plug he is pushing if he wants to continue in business, and not suffer from many bad reviews himself.
Reply | Report Abuse | Link to thisSorry for the typo:- it should read "recognisably" or "recognizably".
Reply | Report Abuse | Link to thisWhile crowd reviewing may average out to the 'real' and exclude outliers, how is one sure that the average opinion matches one's own tastes? Customer restaurant reviews can be snarky, especially when the reviewer has no experience with food preparation and restaurant management. 'Professional' reviewers usually visit a restaurant at least twice before writing a review, but peer, grass-roots reviewers can throw out an opinion at any time. For book reviews, I prefer professional reviews by people who know the author's work, or know the social or historical context of the book. Facebook has reduced us to simply 'like' or unlike without nuance and context.
Reply | Report Abuse | Link to thisheh...comments on an article about comments. Just had to comment: uh, no comment
Reply | Report Abuse | Link to thisAs a non-scientist, I'm curious as to why the author paints negatively the "solo voices of the untrustworthy". Isn't that an assumption? I.e. that these voices are untrustworthy? If someone had a habit of giving very high and/or very low scores to something (e.g. music or movies), and I found that their tastes co-incided with mine often, then I would be more likely to try something that they recommend which I might not have ventured to try that myself, as a result of my own biases.
Reply | Report Abuse | Link to thisSimilarly, the author seems to idealise the mainstream...they want to be in agreement with them. From memory, the movie "The dark knight" has one of the highest every ratings at IMDB...but I'm not THAT keen on it...same goes for "Titanic".
With music, sometimes if I listened to the mainstream, I would not find pleasure in albums that the self-appointed guardians of that music hate...e.g. in certain genres, some albums are reviled by the so-called 'real' fans of that band...and I've found that I LOVE those albums.
Sorry, I'm behind in my reading. Actually let my SciAm subscription lapse. Flog me.
Reply | Report Abuse | Link to thisHail, Mr. Pogue!! You made it through this entire article without using the botch "e-pinions"!!
BRILLIANT!!!!