Wednesday, 10 August 2005

It never does to bite the hand that feeds, but as a yet unpublished novelist I thought that I would mutter a bit about Amazon before it gets to the point where every post on Wenlock is an exhortation to buy my book (still no title, but the elephant remains a big selling point).

Before going into my mutterings I will note that there appears to be some debate about how significant Amazon's sales are. For instance Sarah Weinman offers some sales figures for a recent New York Times best-seller that suggest that of about 98,000 sales in a three week period Amazon were responsible for only 320, with Walmart's distributor, Anderson Merchandisers, being responsible for about 50% of the rest.

Weinman doesn't say where she gets her figures from, nor does she identify the book beyond saying "a NYT bestselling thriller writer's most recent book", nor does she say which three weeks the sales figures come from. I am therefore deeply suspicious of her "hard numbers" and "actual figures". The only place that is likely to have these sorts of sales figures for a three-week period is the book's publisher. It would not surprise me if these numbers represent the total number of copies shipped out by the publisher, and it would also not surprise me if they reflected a time period before publication date. Walmart have a "pile 'em high, sell 'em cheap" strategy and Anderson will buy in bulk and ship across the US to ensure that this can happen. In contrast Amazon are seeking all the time to minimise their stock levels to limit how much capital they have tied up in their warehouses. In the run-up to publication Anderson will buy large quantities, many of which may well be returned unsold in a month or two. Amazon will buy no more than they need to meet advance orders and likely sales over a few days.

A different view of Amazon sales comes from Chris Anderson, Morris Rosenthal and from MIT. This addresses the importance to Amazon of the "long tail" - those titles that are outside the top 100,000 carried by the big US bookstores (eg Barnes and Noble, Borders). The best summary of this work, with lots of links to previous published material is in Chris Anderson's blog, which is actually called "The Long Tail". I am grateful to Anne Weale for pointing me at this.

It would be wrong to set up Weinman and Anderson in opposition to each other. They are not discussing exactly the same thing. However I am not going to ignore Amazon as a cursory reading of Weinman's piece might encourage me to do. As Anderson shows, there's a lot going on in those numbers.

Which brings me to what I was going to say before I was diverted into all that analysis. I'm not going to moan about Amazon selling used books alongside new ones - I don't like it, but plenty of other people have said all there is to say on the subject. Instead I am going to moan about a new "feature" on Amazon (new, at least, in the UK). It appears with the words "Customers interested in this title may also be interested in:" The site explains this thus:

Sponsored Links are advertisements that provides for you. We receive Sponsored Links from Overture. When you click on a Sponsored Link, we get revenue. The selection of Sponsored Links that are displayed is based on keywords. For example, if you search for "Bruce Springsteen" or view pages about Bruce Springsteen, the Sponsored Links may point to sites that sell tickets to his concerts or provide information about him. Sponsored Links are always clearly labelled.

Generating additional revenue from Sponsored Links allows us to offer you lower prices--something we are dedicated to doing in every way we can.
Note that bit at the end of the second paragraph:
Sponsored Links are always clearly labelled.
Oh, really? The sponsored links are headed
Customers interested in this title may also be interested in:
This is distinct from the more familiar
Customers who bought this item also bought:
Customers who bought books by [your name here] also bought books by these authors:
See the clear distinction? Me neither.

Why am I so worked up about this (apart from needing something to blog about)? It is because I found the old, unsponsored links and recommendations so impressive. Did I tell you about the time Amazon came up with three book recommendations, and I had bought all three books (from non-Amazon sources) within the preceding four days?

One of things I really liked about Amazon was that these recommendations were genuine, in that they reflected the habits of other book buyers. Of course you had to aim off for best-sellers - whatever book you bought a couple of weeks ago you would have been told that customers also bought Harry Potter and the Half-Blood Prince, but once you got past that it was always worth following up other customers' preferences. This contrasted with the bricks-and-mortar chain bookshops whose displays and recommendations are paid for by publishers.

But now, in pursuit of higher profits - sorry, "lower prices" - Amazon have joined that tawdry crowd and slipped this little cuckoo into the nest. It's not there for many books yet, but it is creeping rapidly into the music section.

It's not as if the links are actually very useful. OK, so if I am looking for an album by Lindisfarne I might, just possibly, be interested in a holiday cottage near there, but if I am looking for the works of Boston I have more than a feeling that a flight to Massachusetts is unlikely to be top of my list. And why, for crying out loud, if I am browsing for an album by Yes (and this has been known) should I be interested in buying car insurance, even if it is from a company called "Yes Car Insurance"?

Luckily there is a feedback option for these links. I have been using it.


Simon said...

Take a look at Tod Goldberg's analysis of the cost-effectiveness of Amazon's sponsored links.

Stephen said...

Thanks for the pointer, Simon. $750 sounds pretty steep for that sort of promotion. When I get to that stage I will probably rely on word of mouth. I wonder how effective it is to be on one of those customer lists, particularly if it also includes a few high sellers.