And Kelley Eskridge on The lessons of Amazonfail in terms of management issues.

From what I’ve heard it seems to have been a combination of their filtering policy (to exclude porn), content tagging (which is done on purpose, and possibly affected by miscommunication issues, but also with a long history of bias), and coding changes — the only part that can be described as a “glitch” — coming into effect to prejudiciably filter titles according to new tags or to tags that hadn’t led to filtering before. According to Amazon’s statement, 57,310 titles were affected.

Amazon still needs to apologise to the authors of the works they discriminated against, to recognise that there was a problem beyond the so-called “glitch”, in the system that led to the problem (the biases in tagging and filtering) and in the management of the problem, in responding to user complaints and in publicly addressing those complaints.

I can’t believe that amazon would go about the process of editing search results by hand, internally. You’d think that they’d sell preferential search responses at the very least, before they started censoring the list. And what’s more, the application of the censoring is very uneven. It hit academic books, quasi-mainstream cultural critique books that weren’t particularly radical (Female Chauvinist Pigs; I think some Camile Paglia’s stuff as well), some queer stuff was untouched.

Censorship is always uneven, particularly over large datasets, that’s sort of the point, but this is so particularly uneven that it seems unlikely that this is the result of a single monolitic ideological mission from within amazon.

There’s been some talk, particularly after the release of the “glitch” press release (see, above with regards to “public relations fail”) that some sort of public ranking feature that allowed users to flag books as being inappropriate, went awry and delisted things that had passed some threshold. This would account for the unevenness, for sure.

The questions, that I’d have for amazon, then, aren’t “why are you censoring queer folk’s books,” but rather:

- Why are you allowing anyone to censor anything of your search results? A well indexed, easily search able inventory is what makes amazon valuable, if search isn’t reliable, you won’t be as valuable nor sell as many books. What gives?

- Why are you letting this PR disaster continue on without providing information that would defuse it? More reliable information in this situation could only help matters for you and your customers. What’s up with this tightliped “glitch” press releases when you have blogs and twitter?

- What the hell happened with that customer service? Even if it were true, I can’t fathom wanting a customer service rep. to give that kind of answer. Seriously, yo?

Beyond that, my mind boggles.

I think I remember on Mark Probst’s original post to the effect that amazon.de has also been caught up in this, as well, but can’t find the link.