(G)rave Amazon reviews

Published 16 February 04 10:45 AM | despos

An Italian newspaper reported it yesterday quoting this Amy Harmon's article from the New York Times. A quick search on Google proves that the rumor started spreading out way back. What's the point? Amazon book reviews are (sometimes? often? never? always?) a farce.

The article reports that a bug in the Amazon software (the Canadian site) revealed the identity of thousands of anonymous reviewers. In at least one documented case, the author of the book also wrote a five-star review.

"That anybody is allowed to come in and anonymously trash a book to me is absurd" 

This is the comment of John Rechy, one of the authors of such "homemade" reviews. There's more, though. There are significant clues leading to the conclusion that a sort of war is in course between competitors. Unknown reviewers--the NYT article reports--often appear behind the signature "a reader from New York".

Hey, wait a moment.

Have you ever read through the reviews of my Programming ASP.NET book? At this time, Amazon awards it with 4.25 out of 5. There are several fully documented five-star reviews with arguments. There are quite a few with less stars, but equally with arguments. Finally, there are a few with one star or two but NO arguments. And who's the author of (at least) one of them? Just "A reader from NY" :-))))

Quoting from that one:

I purchased this book because it is from MS Press. But this book really disappointed me. There is no logical flow . There is a lot of text in this book but it is very hard to keep your interest in this book. Sometimes I felt that even English wasn't good. I tried to read this book at different times of the day ,morning ,evening,before dinner, after dinner, on weekends ,during weekdays but nothing changed.I want to 'donate' this book to someone. My personal advice is 'don't buy this book unless you have read a few chapters at the book-shop'. In fact Jeff Prosise's book on .NET has better material on ASP.NEt than this book.

Respectable point of view--definitely not a review.

Probably, the problem (if we still want to call it this way) is in the free text that is allowed. A review must give certain structured info to the potential reader. It should be a sort of guided wizard, not just comments. It is too easy to "trash a book" w/o plausible reasons.

OK. Am I the author of one of those rave reviews for my book? No. However, I often answer the next question "Was this review helpful to you?". Of course, unargumented comments are not much helpful. To anybody. 

Comments

# mnrp said on February 16, 2004 06:15 AM:

Maybe a nice solution would be for Amazon to provide a little dedicated space for the author him or herself to write a about the book. Maybe even respond directly to comments (without being forced to leave a rating). How Amazon would limit input to the author I'm not sure, but it might fix the problem in most cases.

# SBC said on February 16, 2004 06:36 AM:

good points Dino - I think Amazon needs to rethink about how reviews are conducted. There appears to be three (may be more) categories of reviewers - one is the serious 'Gunderloy'-type & these are the non-anonymous "Top 1000" or so reviewers who do provide worthwhile reviews. The second is the average "happy to buy this useful book and so should you" type - these just want to spread the word that they spent their money on something worthwhile. I think I fall into this second category - most of my purchases contribute to my work and if they help resolve a problem or two then the book pays for itself almost instantly! The third category are the usual suspects - anonymous reviewers (with hidden agendas) and other heads who would just slam something for the sake of slamming.
PS - I thoroughly enjoyed your Applied-XML book, read my review at Amazon .. :-)

# G. Andrew Duthie said on February 16, 2004 09:50 AM:

"Maybe a nice solution would be for Amazon to provide a little dedicated space for the author him or herself to write a about the book. Maybe even respond directly to comments (without being forced to leave a rating). How Amazon would limit input to the author I'm not sure, but it might fix the problem in most cases. "

I do wish Amazon would do something like this. The main reason I don't respond to reviews that contain factually incorrect or misleading information is because I would have to choose a rating, and I'd prefer not to skew the results.

Amazon does offer the ability for authors to send in their information, and/or for authors or publishers to correct incorrect information in their book listings, but it's not an automated process, so it's kind of a PITA, and often not worth the effort.

I'm pretty much to the point of questioning whether the whole review feature is really all that useful to the buyer. It may be great for happy or unhappy customers who want to share their experiences with a product, but a potential buyer needs to be very careful about trusting the reviews, since there's no control over who can post, or what they can post.

# Scott Mitchell said on February 16, 2004 11:57 AM:

The Amazon.com reviews are like the American political process - the only people that take the time to voice an opinion are the 10% of people who have strong feels about the book (pro or con) and the people who stand to profit from a review (the author, with good reviews, and competitors with bad reviews).

# Grant Harmeyer said on February 17, 2004 07:45 AM:

I agree with Scott on this one. People that write the reviews for a book on Amazon are typically the people who feel strongly about the book in one direction or another. Also, it always appears that Amazon only shows reviews that will give the potential purchaser a balance of good and bad about the book. Some great, some good, some bad. I also don't know how to accomplish it, but the author should also have the opportunity to express their thoughts about their book.

I have never posted a review on Amazon, sorry Dino. But rest assured, I picked up your Programming ASP.NET book at my local B&N and it's my ASP.NET reference that rarely leaves my desk, and the examples are fine for someone who knows the language, but just wants to become intimate with the ASP.NET technology. Kudos.

# Kathleen Dollard said on February 19, 2004 10:34 AM:

The first review of my book thoroughly trashed it for not being the book they wanted. The guy was right, I don't cover Reflection.Emit in a book on application code generation. He was also correct that the text of the book is in VB.NET for reasons explained in the Introduction. However, we worked pretty darn hard to hand translate thousands of lines of VB.NET code to C# for the download because the book applies to both - and I do love C# programmers as much as I love VB.NET programmers. I just have so much code I had to stick to one language in the text. Also, anyone who has a binding problem should contact Apress, or myself and we'll get the problem fixed - that part worries me and if there are quality issues with the printer, let's get them fixed.

Whew! Thank you Dino! I feel so much better for having been able to respond SOMEWHERE to that review!

This morning there's another anonymous one star review :( Again, incorrect - The book has about 40 pages on the CodeDOM (out of over 700) and 15 of those are in an appendix. If you're doing Code DOM, this is pretty important stuff and I think worth the price of the book, but if you buy it as a CodeDOM book, you'll be disappointed.

I think the answer probably lies not in changing the system, but educating people in how to use it, and three small tweaks. The reviewers name link their other reviews. You can see what other books they ranked, and how they ranked them. If they normally due 17th Century English Poetry, and suddenly give a review to a .NET book, well... If they give everything five stars or one star, well.... People can figure it out if they know what to look for.

Which leads me to the tweaks I think the system needs. All reviews by the same person should be tracable to the same person or alias, allowing a consistent reviewing alias to be easily entered. I know that people will find ways to work around the system, but if all "normal" reviewers are encouraged to build a portfolio of reviews (even if its just two or three) then people can interpret their reviews. If they have only one review, well... people are smart. The second tweak is that author's should be allowed a few words to respond to every review. Limit it to 256 chars to keep us from waxing on if it seems important. But both Apress and I would really like to say "Please, please tell us if you have a quality issue. You were looking for a different book so I'm glad Amazon let you return it. Reflection.Emit is not an appropriate tool for application code generation, as discussed in Chapter 1."

I am no longer stressing over that unfair review becuase there are two newer reviews from people that actually read the book. My only regret is that people don't have a way to know that Jack Herrington (one of the reviewers) actually wrote the current other book on code generation - "Code Generation in Action" from Manning. The two books are very complimentery, and having Jack's backing for my book means a lot more to me than the "A reader from Southlake, TX" who misunderstands either application code generation or Reflection.Emit (or both) if he thinks they fit together. (or Redmond who seems confused about Code Generation creating code) That's the third tweak. I'd like people to be able to describe themselves and have that appear if you click the link on their reviews. It's valuable to know if someone is a twenty year programmer, or an MCSD, or whatever.

I know this is long winded, but if I'm honest this issue caused me to shed more than a single tear a week or so back.

# Charles Abboud said on March 8, 2004 05:17 PM:

I agree entirely with you Mr Esposito

I'm not an author but I've written one Amazon review (5 star) for Mike Gunderloy's 70-306 exam book and I felt he'd put such effort towards making it a class product and since I'd read every word and did all the exercises in it and consequently passed the exam I was irritated by what appeared to me to be unfair and at least (in my view) inaccurate criticism of a truly remarkable effort. I would say a negative review that is too vitriolic or does not make what sounds like a fair-minded argument about the book's shortcomings always says more about the reviewer than the book.

Regards

# Michael Brundage said on March 21, 2004 11:05 AM:

Scott's right, and so Amazon ratings never tell the whole story. However, some items (especially kids' toys) seem to break this rule, with hundreds of helpful reviews. IMO, any item with only five reviews might as well not be reviewed at all.

The Amazon ranking system is also strange, changing by thousands in each hour -- I guess you already know about http://www.junglescan.com ?

The real bummer (speaking from my experiences as an author) are poor-quality reviews by semi-professional book critics. I hope these are the exception rather than the rule, but programming books seem to get a lot of them.

You know -- reviews where the critic harshes on the subject itself (instead of your presentation of it) or other things outside your control as an author, or complains that X could have been ten times better (without any constructive ideas how, even if you follow up with him later), or rants or raves about the book without any real details (making you wonder if they read more than the back cover). As just one of many examples, I had one paid technical reviewer (a professional writer herself) say that I should have included a copy of the XQuery standard (what, 1200 pp.?) in my XQuery book (http://www.qbrundage.com/xquery/ ). That's just crazy talk!

I think as authors we just have to have faith in our readers, that they make these review evaluations too, so that poor-quality reviews (positive or negative) go straight to /dev/nul with little or no impact on actual sales. Otherwise, you'd go nuts over all the ridiculous stuff people write about your work.

Besides, your XML book is excellent, and everyone who's read it knows that, and passes that info through the grapevine. That's the review network that really counts, and it isn't written down anywhere. (I myself recommended your XML book just last week to a dev lead in my wife's group at MS, who wanted a good book on doing XML with C#. And once he's read it, he'll recommend it to others, and so on...)

# 杭州主机托管 said on July 6, 2004 03:58 AM:

~~

# Barbara Adams said on September 18, 2007 11:02 PM:

I'm an author, not in the field of most of you brilliant people here, (sustainable farming) but share your thoughts on the ability for any unknown to trash a book on a site as huge as Amazon, which acts like a worldwide bulletin board. Unlike smaller online forums or live book groups where word-of-mouth is legitimate and self-censoring because we can get to know the "reviewer," (Is she having hormone problems again today? There he goes again, complaining before he even finishes the first chapter), Amazon's ranking is kin to the McCarthy era. Nobodies can see their name or at least their reviews in print, and really have fun with it.

Science proves that brains learn and process information differently. Someone can even actually read a book and state that "it has no useful information" when that book is helping many people who process information in the manner the author targeted, yet they aren't the type to waste time reviewing on Amazon.

Will one of you brilliant people create a new online bookstore or talk someone else into it? Target the ethical customers, the people who want a society that supports new inventions and writers and authors and philosophers. Let the confused cheapies stay with Amazon.

Leave a Comment

(required) 
(required) 
(optional)
(required)