Wednesday, September 5, 2012

Fake apocalyptic book reviews


If you spend much time at sites that have user reviews of items that they purchased, Amazon to me being the classic one, it doesn’t take a lot of effort to figure out that many of the reviews are phony.  I always just figured that the authors threatened or bribed their family members to write good things about their book.

As it turns out, you don’t even have to bother your family members to say nice things about your self-published apocalypse-in-progress novel.  You can just pay people.

David Streitfeld, New York Times, 19 August 2011 (hat tip: Big Picture)

In tens of millions of reviews on Web sites like Amazon.com, Citysearch, TripAdvisor and Yelp, new books are better than Tolstoy, restaurants are undiscovered gems and hotels surpass the Ritz.Or so the reviewers say. As online retailers increasingly depend on reviews as a sales tool, an industry of fibbers and promoters has sprung up to buy and sell raves for a pittance.

“For $5, I will submit two great reviews for your business,” offered one entrepreneur…

A researcher showed one way that it could be done.


Myle Ott, Yejin Choi, Claire Cardie, Jeffrey T. Hancock, Cornell University, August 2012

 



Crowdsourcing services such as AMT have made large-scale data annotation and collection efforts financially affordable by granting anyone with basic programming skills access to a marketplace of anonymous online workers (known as Turkers) willing to complete small tasks.

 


To solicit gold-standard deceptive opinion spam using AMT, we create a pool of 400 Human- Intelligence Tasks (HITs) and allocate them evenly across our 20 chosen hotels. To ensure that opinions are written by unique authors, we allow only a single submission per Turker. We also restrict our task to Turkers who are located in the United States, and who maintain an approval rating of at least 90%. Turkers are allowed a maximum of 30 minutes to work on the HIT, and are paid one US dollar for an accepted submission.

The New York Times has a recent piece on Todd Rutherford, who worked in book marketing for self-published authors, and started a business based on fake reviews.  He quickly was taking in $28,000 a month. 

David Streitfeld, The New York Times, 25 August 2012 (also hat tip: Big Picture)

Before working for the self-publishing house, he owned a distributor of inspirational books. Before that, he was sales manager for a religious publishing house. Nothing ever quite worked out as well as he hoped. With the reviews business, though, “it was like I hit the mother lode.”

There is a good combination (LOL), self-publishing, inspirational, religious- sounds like Amway.

The Rhitholtz (Big Picture) piece notes a number of very technical ways to bias the real from the unreal reviews, but they example is directed toward hotel reviews, which doesn’t help me/us a whole lot here.  Any end of the world hotels we are going to be staying at hopefully have some sort of heavenly choir directing the way:  no online reviews required.

So how do you figure out the truth.  Oddly enough, the entrepreneur himself notes the method.  Look at the negative reviews.  Since most online reviews tend toward being overly positive, anything from a three down (out of 5) would be in the lower rankings.  I particularly find the three-star reviews helpful because they tend to comment on both the positive and negative aspects of a novel.  It also works for novels where the author has an existing online non-critical following who will tout his books with a religious mania regardless of their quality.  These are honest reviews, but not very helpful for those who are not drinking from the same pitcher of Kool Aid.

If you want to look at the positive reviews, one way to double check is to see how many other items the reviewer has looked at.  Most of the time the spam-reviews, or for for that matter the Kool Aid drinkers,  have very few reviews at all.  When they do have multiple reviews, they will always say the same thing about every book.  Note however, that the study above, which focused on hotel reviews, indicated that it was very difficult to distinguish between the real positive reviews, and the fake ones based on content alone.

4 comments:

Stephen said...

Ah, that explains much...

russell1200 said...

Stephen:

Yes it does. A few times I have debated commenting on the canned reviews during my own review, but generally the book's quality speaks for itself.

John D. Wheeler said...

Excellent post!! I enjoyed it immensely!! Everyone should read it!! I give it 5 stars!!

(Okay, where's my $1?)

russell1200 said...

John:

LOL- I am sorry but all contractors must submit a valid insurance certificate PRIOR to starting work on the contract.