Tuesday, May 15, 2012

Price Bias – A New Wrinkle In Amazon’s Ever-Changing Algorithms


OK, we’ve been keeping something from you. We suspected this early on, but it was such a major change to what we were used to seeing we wanted to be sure what we were seeing was actually there – and repeatable.

Price bias in Amazon’s algorithms.

Until recently, Amazon’s popularity lists have been price-agnostic. That changed on May 1. Ed Robertson has a nice rundown on the evidence. Huge thanks to Ed for the hours (and quantities of alcohol consumed) spent chasing down the figures.

I won’t repeat all the gory details here. Go read Ed’s piece for that.

What I want to discuss today is editorial in nature: Why Amazon might be changing the playing field yet again. The following is opinion and speculation only.

Let me be clear up front. In no way do I believe Amazon is changing its virtual display shelves in a way that is meant foremost to have any impact on indies. I’ve used the expression that we’re simply collateral damage numerous times, and I stand by that belief. Of course, whether you die by virtue of intended murder or unintended manslaughter, the result is the same.

So what’s up then? Weren’t Select titles that were going free and selling hundreds and thousands of copies post-free making Amazon money? Of course. But Amazon will sell books regardless.

Is Amazon reacting to a public outcry over the quality of the books being routinely put on the front shelves? Hmmm. I haven’t heard or seen such an outcry, and to make an impact, I would think it would have to be a tsunami of outrage to turn Amazon's ear.

What’s prompted the change, I believe, has all to do with the demise of agency pricing. When Amazon is again free to offer loss-leaders of bestsellers from the Big 6, they will need to make up the thin margins they’ll no doubt be earning per copy through quantity. It will behoove Amazon to ensure those books are kept in front of the readers, and keeping them high in the popularity lists will help accomplish that. Remember, Amazon routinely sends email blasts to customers that feature the top 5 or 6 titles in the pop list directly in the email and then links to the rest of the list itself, so the popularity lists feed Amazon's recommendation engines in large part.

One pricing bit we’re also looking at is the List Price of the ebook vs the actual Purchase Price. Our testing of List vs Purchase price is still ongoing. Of interest so far is to note that the Amazon imprints now seem to be offering a List Price of $9.99 while actually selling their titles for $4.99 and lower. That’s good marketing to make the customer think they’re getting a deal, but I believe it goes well beyond simple Marketing 101 strategy. If the price bias uses the List Price as its factor, then Amazon is giving its imprint books the same advantage in the algorithms as the Big 6 publishers will have, and the Amazon imprint books will keep getting pushed to the front shelves along with the Big 6 titles – and ahead of the indies.

Ahead, because “List Price” is not a tool the indies have to work with, and we’re clearly disadvantaged because of it should the theory about List Price prove out. We can change the price on a book and have the old List Price reflected with a slash through it, but only temporarily. Within a few hours or, at most, a few days, the new Purchase Price becomes the List Price.

What that means, of course, is that indie books that are lower priced have to sell a lot more copies to be promoted up the pages of the popularity lists than do books with higher Purchase Prices and/or List Prices. So the answer is to price everything higher, right? Not necessarily. First, a book has to have the perceived value of the price you want to sell it at for it to actually sell at all. Plus, if the algorithms are working off the List Price, that gives Amazon the luxury of pricing a book at any price and still enjoying better shelf placement. Say you price your new thriller at $4.99 to avoid the bias, and Thomas & Mercer is selling their newest thriller at $3.99 but with a List Price of $9.99. That T&M book will still be placed higher than yours unless you sell a substantial number more than the T&M title.

If this algorithm change persists – and who knows, it could be replaced in a month or two! – what I foresee when the agency-priced titles are wholesaled once again to Amazon is this: Customers will be able to purchase those pricier bestsellers (as ebooks) for less than what they’re being offered at now. Score 1 for the customer. In an effort to compete, indies will begin raising the prices of their books, and the 99c to $2.99 titles will be harder to come by. So customers will be paying less for one class of books and more for another class as prices begin to normalize within the $2.99 – 9.99 range that Amazon has been pushing all along.

Of course, there will always be outliers. Of course 99c books that can move 100,000 copies will still rank well and sell well. Of course indies who price their books above the value customers are willing to pay will not sell.

Of course there are factors other than price that will affect how well your book performs. If you concentrate on just one variable, such as price, to the exclusion of all others, you’ll likely be disappointed in the results. Price is still just another consideration in the overall equation.

On another note regarding the agency pricing scandal, check out David's Gaughran's Open Letter to the DOJ.

7 comments:

J. R. Tomlin said...

There is one way to show a price reduction on Amazon for indies and this is one reason I'm itching to get out of Select.

Keep prices higher on Amazon, lower them elsewhere and let Amazon price match. People have always done it for going free but it works for "price reduced" too. I can price my historical novels at $9.99 and price them on Smashwords at the current $4.95. Eventually (almost always) Amazon picks it up and shows it as a price reduction.

BELIEVE ME this is something I'm thinking about. At the moment, I'm playing around with raising the price on one of my novels to $7.99. I'll see what happens.

Phoenix Sullivan said...

JR: From what I've been seeing, the price reduction with the slash mark only holds for a few days at most before the reduced price becomes the regular price. Are you seeing the price reduction last longer?

Deanna Chase said...

Phoenix, I have not had that experience. I had a book price matched for almost a month (just changed it last week), and it always had the list price with a slash and then the discounted price.

Calee- Xist Publishing said...

Thanks so much for this! I had figured the change was so that when Amazon slashed a book's price (then raised it again) it would be sticky on the pop list.
The second half of your post has been confirmed in my data this month, though I hadn't figured out why. Our higher priced books have continued selling at their same rate (2.99 - 3.99) with the .99 titles almost falling off the map. It's a far cry from the 100s of sales the day after a free promotion, but even with decreased sales in April, our revenues were about the same--because the higher priced books made up for it. I'm seeing the same thing happen, through more amplified now and your theory seems to have support in our data. Thanks!!

Elizabeth Ann West said...

Technically, doing the price matching game on purpose is against the Terms and Conditions Agreement with Amazon you agree to when you publish through KDP. One of the provisions of the 70% royalty is that you agree to not set the price lower than what you set it as on Amazon.

Will Amazon enforce it, who knows. I'm just making sure authors are aware that taking those kinds of steps could ultimately lead to more consequences than what you expected, including losing your KDP account for violating the terms.

Amanda Brice said...

Phoenix, it's not just a temporary price reduction. Mine has been reduced for months now -- first to $2.99 and currently to 99 cents. Lovely red slash and all.

Amanda Brice said...

Of course, when I do a price match reduction, I reduce my royalty rate to 35% as opposed to 70%, so perhaps that's the differnce?