Not All That Glitters: Beware Bad-Faith Book Reviews

Here's an interesting bit of information, worth pondering the next time you shop Amazon for a certification book (paper or electronic; they do them all). I've recently heard from a long-time cert writer that one particular title has come in for unusual coverage in its reviews on Amazon.

Basically, it looks like somebody is deliberately attacking that title by posting negative reviews, and then getting other people to give those reviews the Amazon equivalent of a Facebook "like." The actual text after such a review reads "Nn people found this helpful," where Nn is the count of people who endorsed the review.

For this particular book Nn varies in a range from 2-16, with most values 10 or higher. Endorsing a review simply requires pressing the "Helpful" button. Here's what that looks like:

How Reviews Work on Amazon

Anybody who wants to can post a book review on Amazon, but they must purchase a copy of the book that they review from Amazon. This makes any buyer what Amazon calls a "verified customer." Each buyer can post one (and only one) review of each book they buy.

A book buyer is free to say whatever they want, and to provide a star rating from 1 to 5 stars (more stars indicate a higher rating; fewer stars indicate a lower rating). Amazon does not police book ratings, and they don't make it easy for authors to contest or protest negative or unfair reviews.

I believe Amazon posits that heavily reviewed books will exhibit crowdsourcing behavior and tend toward the truth anyway, despite false positives and false negatives. FWIW, I agree with this philosophy.

On the other hand, as the author of more than 100 books myself, I have felt plenty of stings from negative reviews. IMHO, some of those negative reviews were unkind, incorrect, and unfair. That's life, right?

A Series of Unfortunate Events � or Not?

One fascinating element of this apparent smear campaign against this particular book at Amazon is a sequence of nearly identical book reviews posted one-per-day over a two-week period. Each such review means somebody bought at least an electronic copy of the book, if not a paper-bound edition.

All of these negative reviews are fairly short, and pretty sparsely detailed. I'm fascinated that Amazon's extraordinary AI capabilities don't extend far enough to look out for intentionally over-positive or over-negative reviews, or repetitions of same. I can only hope they'll get around to building such a filter sooner rather than later, knowing that they're unlikely to do much to limit potential sales of books (and other goods).

Ironically, one of the recurring criticisms in these reviews dings the book as "poorly written." The reviews that levy this charge, alas, do so using bad grammar, bizarre word choices, and occasionally impenetrable prose.

Another recurring item dings the authors for an incorrect acronym (which has already been corrected in the errata, and won't appear in subsequent reprints). One more repeat element claims that some particular Wi-Fi encryption scheme isn't covered in the book, when there's a multi-page section on that specific topic.

Interestingly the negative reviewers are mostly anonymous ("Amazon Customer"), provide only a first name, or go by common names with short review and purchase histories. Nearly all call the book out as insufficiently updated. They claim it's simply a repackaged version of the book's previous edition for the prior version of the cert exam it covers, and not really current.

However, this book has a brand new author team. They completely rewrote the title for the current version of the exam it targets, and recycled none of the old copy. In other words: reviewers' claims that the content is outdated and carried over from an old edition are simply untrue.

Careful examination of the reviews also show that another competitive title from a different publisher keeps popping up in reviewers' purchase history — and that other book garners glowing, ecstatic reviews. Could this simply be a coincidence? Probably not.

Even Reviews Must Be Questioned, It Seems

The Internet is sometimes characterized as "the net of a million lies" (from Vernor Vinge's brilliant sci-fi title A Fire Upon the Deep). There's no denying that one must ponder the source as well as the information when it comes to separating good book reviews from bad or misleading ones.

The lesson I'm taking away from this squalid little tale is that it makes sense to vet the reviewers, as well as what's being reviewed. I'm now much more inclined to pay attention to reviews from "Top Reviewers" and from those whose review histories show both knowledge of and interest in the topic being reviewed.

We all know the old Latin saw "caveat emptor" (let the buyer beware). I'd like to add another saw to that, namely "caveat lector" — let the reader beware. Don't say I didn't warn you, please!