Amazon Reviews
My own research suggests that GH is no more or less credible than Amazon's other "celebrity reviewers." Harriet Klausner, No. 1 since the inception of the ranking system in 2000, has averaged 45 book reviews per week over the last five years—a pace that seems hard to credit, even from a professed speed-reader. Reviewer No. 3, Donald Mitchell, ceaselessly promotes "the 400 Year Project," which his profile identifies only as "a pro bono, noncommercial project to help the world make improvements at 20 times the normal rate." John "Gunny" Matlock, ranked No. 6 this spring, took a holiday from Amazon, according to Vick Mickunas of the Dayton Daily News, after allegations that 27 different writers had helped generate his reviews.In the comments section of Metafilter was this:
I'm in the Amazon Vine program. Every month they send two free things, which I pick from a list of available items - books, health care, electronics, music, etc.. it's pretty nice honestly, these are things that would cost $50 or more had I to buy them, it isn't crap. The electronics in particular can be very high value items. In return, all I have to do is write reviews. It's unclear how Vine members are chosen, I'm ranked around 3500 but there are people ranker higher who were not chosen.
Everything this Salon article says is true, and worse. For example, I've seen self-published authors self-reviewing (bad books) with sock puppets giving dozens of 5-star reviews. I've seen reviewers gang up and help or hurt other reviewers. People getting paid to write reviews.
If reviewing for Amazon determines what you read, than it's gone too far. I review a lot of out of print obscure stuff that might get one helpful vote in 5 years - who cares, I review what I like to read. If your goal is to get helpful votes, than you must review new books for a number of reasons - first reviews get double the number of ranking points, and new books get a lot more people looking at them trying to decide to buy or not, thus more helpful votes.
The ranking algorithm was disclosed by Amazon a while back in the discussion forum - something like, the first 5 votes gets one point, 20 votes get another (all doubled if your the first reviewer). Unhelpful votes likewise remove points in similar block sizes (but never removes more than that review gained ie. 50 unhelpfuls and 20 helpfuls would net out to zero points)...oh on the ranking, the number of points caps out after 20 votes, so you have to keep writing new reviews to get more points, each review has a limited number of possible points.
posted by stbalbach at 7:01 AM on January 28
Also on Metafilter were these examples of gaming the system. Retaliation reviews and guerilla tagging.
2 Comments:
At 9:21 AM, Anonymous said…
I'm a fairly active Amazon reviewer and I'm involved in the Amazon discussion forum. Several of my friends are in the vine program and I know a couple of dozen members of the "top 1000" reviewers. I myself will probably join their ranks this year.
It's all true. Klausner doesn't write reviews so much as book-jacket synopsis filled with errors easily noticed by people who've read the books in question. There have been group reviewers posting as a single person, schemes and scams, and all kinds of rigging of the system.
The actual algorithm is far more complicated than the outline which was published. So far, none of the full-time review junkies has cracked it. But it can still be gamed, and is being gamed.
All that said, it's working for Amazon. They get millions of words of free content - billions perhaps - and that content helps sell product.
For myself, I'll keep doing it the old fashioned way - read a book or watch a movie - write an honest review - repeat.
At 8:55 AM, Anonymous said…
Generally, though I always tend to buy from the cheapest seller, I read reviews on a book or product from several different sources. While Amazon is becoming worryingly huge it still provides good customer service most of the time, and that's an important criteria to me, so nine times out of ten that's where I end up purchasing. Reviews always tend to be ambiguous anyway, with 40% adoring the product, 20% hating it, and the rest somewhere in the middle. More than 20% hating it generally is reason to give it a miss, though with books that's not necessarily the case. Plenty of research before buying is really the only yardstick.
Post a Comment
<< Home