Some restrictions apply

by John Holbo on November 25, 2011

Sometimes Amazon makes me offers I find it quite easy to refuse:

{ 50 comments }

1

Harald Korneliussen 11.25.11 at 7:11 am

There’s nothing like academic books pricing. One thing I don’t miss about being a student.

2

Tim Wilkinson 11.25.11 at 7:53 am

And journal articles, of course. They really need to be made available free to all – I’m sure whatever necessary costs are incurred in the publication process could be defrayed by revenue from, say, institutional subscriptions to print editions, or a universities levy or public funds or something.

3

John Holbo 11.25.11 at 8:23 am

I suppose I should clarify. It’s a moderately expensive art book that some 3rd party seller is offering used for this ludicrous price. New copies are about $60. Go figure. But Amazon automatically suggests that you should buy a used copy and think about reselling it. Not really Amazon’s fault – or any academic publisher – that someone is asking the sun and moon for this particular title.

4

J. Otto Pohl 11.25.11 at 9:18 am

Come on a $1000 is surely just pocket change for a rich and successful academic like yourself. Don’t be a tightwad. ;-)

5

ajay 11.25.11 at 9:36 am

Absolutely. Just put in a bit of overtime on the consulting job at Fannie Mae and you’ll be fine.

6

tomslee 11.25.11 at 1:11 pm

Not really Amazon’s fault…

Isn’t it strange how accustomed we have become to interacting with algorithms? If the sentence was written by a human we’d be quite happy to label them an idiot (well, I would) and possibly berate them for wasting out time, but we are getting experienced at navigating the occasional stupidities demanded by automation and extreme scale.

7

Dragon-King Wangchuck 11.25.11 at 1:35 pm

What do they mean by “as little as”? Are there potential hidden fees in this transaction that might be incurred? Perhaps Amazon is acknowledging that their gift cards are inherently worth less than cash (i.e. you might not spend the entire amount of the gift card for some reason).

8

Zamfir 11.25.11 at 2:20 pm

@DKW, Amazon is encouraging you to use their second-hand book market as a kind of rental service. Buy a book, read it, then sell it on to get most of your money back. For many books this will cost you little more than shipping + plus some (fairly reasonable, depending on your POV) handling fees from Amazon.

Amazon itself isn’t the seller or buyer. They are providing the web-service and sometimes also the warehousing and shipping in return for a cut of the price , but they don’t set the prices and don’t take the risk of having unsold stocks.

So it’s not in Amazon’s advantage to have books listed for 999 dollar. They want lots of transactions (each time making a fee for them), not sky-high sales prices. That’s why they encourage you to resell the book again: it’s attractive for Amazon if the market is filled with amateurs who don’t need to make a living off of it, who are already happy if they can sell a book at a small “rental fee” loss instead professionals who aim for a profit large enough to cover shipping, fees and a bit of extra for the effort .

9

Dragon-King Wangchuck 11.25.11 at 2:31 pm

Zamfir,

I get the idea of what Amazon is doing, I was just questioning the wording. But I forgot about shipping – so yes, there’s a hidden fee. Thanks for the reminder.

10

John Holbo 11.25.11 at 2:51 pm

Just for the record, the book in question is “Lyonel Feininger: At The Edge of the World” (Whitney Museum). Judge for yourself whether it’s likely worth a $1000.

http://www.amazon.com/Lyonel-Feininger-Whitney-Museum-American/dp/0300168462/ref=sr_1_1?s=books&ie=UTF8&qid=1322232519&sr=1-1

11

Billikin 11.25.11 at 3:29 pm

John Holbo: “Just for the record, the book in question is “Lyonel Feininger: At The Edge of the World” (Whitney Museum). Judge for yourself whether it’s likely worth a $1000.”

Does he jump?

;)

12

Barry Freed 11.25.11 at 3:30 pm

I note there are two different sellers there. One is selling it for $999.99 while the other is shrewdly undercutting his competitor by selling it for $999.98.

13

Downpuppy 11.25.11 at 4:13 pm

In most states, the use tax would be about equal to the new book price.

14

ajay 11.25.11 at 4:18 pm

12: the Invisible Hand in action!

15

William Timberman 11.25.11 at 4:29 pm

The language of machines: If you wish to be defenestrated, press one. If you wish at any time to speak to an actual human being, just say agent, or press the pound key — and be prepared to listen to 47 minutes of canned music chosen by our sociologist-in-residence, and played by a manic synthesizer somewhere in the bowels of our Asian data center.

It’s amazing how quickly we’ve adapted to the ways in which computers haven’t really got us all figured out. If it hasn’t happened already, one day we may very well become what they anticipate us to be, and when the telephone asks us if we’d prefer to a) shit, or b) go blind, those will be the only two answers our nervous systems will have prepared us to give.

16

Salient 11.25.11 at 5:18 pm

Not really Amazon’s fault – or any academic publisher – that someone is asking the sun and moon for this particular title.

FWIW, it’s not the asker’s fault, either. Amazon’s internal system sometimes stratospheres the price when there’s only one seller and the book’s out of print. So you can try to sell a book figuring that $20 would be fair and Amazon advertises it for $645, apparently because it tries to base the asking price on Amazon’s own selling price, causing problems when no such selling price is known. This is a “known issue” that’s been around for years unfixed (as any collector of old obscure-yet-classic RPG books knows…).

17

Freddie 11.25.11 at 5:23 pm

And journal articles, of course. They really need to be made available free to all – I’m sure whatever necessary costs are incurred in the publication process could be defrayed by revenue from, say, institutional subscriptions to print editions, or a universities levy or public funds or something.

In fact most academic journals already operate at a loss, even with institutional subscriptions/

18

Henri Vieuxtemps 11.25.11 at 5:33 pm

Well, I imagine Bezos needs money to rebuild his personal spaceship. Don’t be too hard on him, you wouldn’t believe how expensive personal spaceships are these days.

19

Ginger Yellow 11.25.11 at 5:48 pm

I note there are two different sellers there. One is selling it for $999.99 while the other is shrewdly undercutting his competitor by selling it for $999.98.

Someone, I think Language Log, did a post about Amazon second-hand anomalies like this. They’re more common than you’d think and they seem to be driven by price-competition/arbitrage algorithms gone awry, sending the price of random books sky-high as the sellers’ algorithms ratchet each other up.

.

20

Lemuel Pitkin 11.25.11 at 5:58 pm

Amazon second-hand anomalies like this. They’re more common than you’d think and they seem to be driven by price-competition/arbitrage algorithms gone awry, sending the price of random books sky-high

The classic example.

21

Bloix 11.25.11 at 5:58 pm

No one is consciously trying to sell this book for $999.98 . SuperBookDeals is offering it new for $63.50 and used at $999.98. The other seller, “Quick & Easy Marketplace,” has all sorts of stuff for 999.99 (eg “Drawing Cats and Kittens”), so the price is obviously a default setting. Maybe SuperBookDeals has an algorithm that finds used books at other sellers and sets the price at a penny less. But there’s no human intelligence operating here- you’ve been fooled by a pretty stupid Turing machine.

22

Ginger Yellow 11.25.11 at 6:16 pm

That’s the one I was thinking of. Thanks, Lemuel. Not sure why I had it pegged as LL.

23

Ginger Yellow 11.25.11 at 6:18 pm

No one is consciously trying to sell this book for $999.98 . SuperBookDeals is offering it new for $63.50 and used at $999.98. The other seller, “Quick & Easy Marketplace,” has all sorts of stuff for 999.99 (eg “Drawing Cats and Kittens”), so the price is obviously a default setting.

More likely, based on the article Lemuel linked to, that it’s a cap and Quick & Easy is the overbidder on several of these under/over pairs. It maxes out at $999.99, and some other seller(s) is consequently topping out at $999.98

24

tomslee 11.25.11 at 6:19 pm

Bloix – Amazon is relaying an offer of the book for sale at $999.98, and if John H. put it on his credit card the sale would go through just like any other sale. Amazon’s algorithm designers are consciously trying to sell books and this is a result of their efforts, just as much as more mundane prices. To designate some offers as real and others as merely a “default setting” is to forgive algorithm designers for crappy algorithms.

On a brighter note, we can soon look forward to having a similar thread about offers over the phone, thanks to the truly scary IOVOX.

25

bob 11.25.11 at 6:28 pm

As has already been suggested, this is an artifact caused by a stupidly-designed sales algorithm. An interesting and even more extreme example may be found here
http://www.michaeleisen.org/blog/?p=358

26

Substance McGravitas 11.25.11 at 6:30 pm

The Feininger show’s gonna be in Montreal January through May. Go see if you can; it is excellent. I’d supply a link but CT seems a little eaty on Thanksgiving weekend.

27

Salient 11.25.11 at 6:59 pm

it’s a cap and Quick & Easy is the overbidder on several of these under/over pairs.

Wow, I figured the second-lowest price stuff was algorithmic, but it’s mind-boggling to discover some sellers’ algorithms are intentionally over-bidding, meaning it really is their algorithms’ quirks, not Amazonian internal quirks, at fault for driving up the prices. FWIW in my experience finding contact info for a seller and emailing them to ask them about the price leads to an apologetic-sounding reply note and a revision in the price within a couple days. But sheesh, I have a lot less sympathy for the sellers now.

28

tomslee 11.25.11 at 7:06 pm

As has already been suggested, this is an artifact caused by a stupidly-designed sales algorithm.

Vent alert: as has also already been suggested, this is not an artifact, if by “artifact” you mean “something not real”. It is absolutely as real as any other price on Amazon or anywhere else.

I don’t get why people are so quick to say it’s somehow an illusion. If it said $200 and someone paid it would that still be an artifact? Algorithms have output, they don’t have real output + artifact output. If real breathing people choose to use algorithms, then they remain responsible for all the output.

29

tomslee 11.25.11 at 7:11 pm

meaning it really is their algorithms’ quirks, not Amazonian internal quirks, at fault for driving up the prices

… of which, naturally, Amazon takes a rake as pointed out above. Again, Amazon is publishing the price, and to that extent is responsible for it. Amazon sets up a system designed for use by automated agents, it profits from it, it obviously doesn’t spend too much time worrying about overpricing (in this case ludicrously so, but what of the less spectacular cases?) ergo we should be able to at least laugh at it when it gives unexpected results.

30

Substance McGravitas 11.25.11 at 7:14 pm

Vent alert: as has also already been suggested, this is not an artifact, if by “artifact” you mean “something not real”. It is absolutely as real as any other price on Amazon or anywhere else.

The meaning has shifted thanks to computer geeks: artifacts are now actual byproducts of algorithms, just unwanted ones. Lossy JPG files, for instance, are commonly held to contain “artifacts” that look icky.

31

Gene O'Grady 11.25.11 at 7:18 pm

This isn’t really an Amazon problem. AbeBooks, which seems to have a better reputation that I wouldn’t say is deserved, is full of absurdly priced items — for instance, Loeb Classical Library volumes, all of which except for some real dogs that have been replaced, are in print new for $24.00, are frequently advertised as being available used for seventy or eighty dollars. Often from German dealers, for what that’s worth, whereas those dealers seem to be able to offer genuinely out of print German academic books for quite reasonable prices.

32

tomslee 11.25.11 at 7:19 pm

artifacts are now actual byproducts of algorithms, just unwanted ones.

– there are no byproducts of algorithms, only products.
– long-lived unwanted products in production systems reflect behaviour the owner is happy to live with, and according to links upthread this kind of price has been around for a while.
– The distinction between “wanted” and “unwanted” is being imputed to them by readers here, but I fail to see any evidence for the price being “unwanted” except in the narrowest of interpretations. It’s quite possible that an algorithm that gives occasional silly prices also gives slightly silly prices that people pay.

33

tomslee 11.25.11 at 7:21 pm

(The formatting was an unwanted byproduct of CT’s comment system; that was meant to be three list items, each introduced by a hypen.)

34

Substance McGravitas 11.25.11 at 7:22 pm

– there are no byproducts of algorithms, only products.

I have to agree, but at the same time, the word has been poisoned, and when they get a product they don’t want they say that an artifact has been introduced. If an artifact is really really big they might refer to the enormity of it.

35

tomslee 11.25.11 at 7:30 pm

This isn’t really an Amazon problem. AbeBooks, which seems to have a better reputation that I wouldn’t say is deserved, is full of absurdly priced items

Amazon bought AbeBooks for around $100m in 2008. I have no opinion on whether that price was a really really big artifact.

36

tomslee 11.25.11 at 7:53 pm

… not that you can tell that Amazon owns it from the AbeBooks home page. Amazon keeps its ownership of AbeBooks even lower profile than its ownership of IMDB.

37

Adrian Kelleher 11.25.11 at 9:17 pm

Wow. At that price I’d expect a box canvas print of Paul Ross.

38

Bloix 11.25.11 at 10:40 pm

I don’t believe that I said that the book is not actually for sale at $999,98, tomslee. I said that no one is consciously selling it for that price. SuperBookDeals has it for $999.98 and also for $63.50, which means that that they don’t expect anyone to by it for $999.98, doesn’t it?

As for Quick-N-Easy Marketplace, do you think that they expect anyone to buy “Drawing in Color: Cats and Kittens” for $999.99? They seem to have 66,000 pages of items, which means I presume that they have no items at all and that they buy stuff from elsewhere and have it shipped to their unwary customers from the original seller. How they set prices is unknown to me, but I doubt very much that there’s much in the way of human intelligence at work.

39

Salient 11.26.11 at 1:28 am

ergo we should be able to at least laugh at it when it gives unexpected results.

Definitely, no argument there (and btw I’m pretty sure Bloix wasn’t intending to suggest there’s a lack of liability, which seems to be what you’re on about; saying nobody intentionally set a price to $1k, as Bloix did, is different from saying the seller’s not responsible for setting a price of $1k).

Also, artefact really does mean something like ‘output that is not desired and not valued, but not worth fixing because the thing works ok enough to get by [and because the thought of another round of debugging gives you shudders].’

40

Dr. Hilarius 11.26.11 at 3:40 am

I’ve been a buyer, and via consignments, a seller on ABE for many years. The quality of the site has degraded. I can’t say if the decline is due to Amazon or is just some overall market feature.

The insane book prices come about several ways. Most are due to pricing software that checks Amazon prices and then sets a price based on the seller’s parameters. A lot of this is with “mega sellers” who have no relationship to book sellers. They are the people with handheld bar code scanners scarfing up dollar books at thrift stores, yard sales etc. The books go into huge warehouses with virtually no human input. Query one of these sellers about a specific item and you will most likely get a form email telling you that the books are in some remote, inaccessible fortress, so sorry, just buy and return if you’re not happy. I’ve given up on these sellers. I paid $43 for a scarce edition and received a battered paperback worth fifty cents at best. I got my money back but who needs the hassle.

The other source of insane pricing is “drop shipping.” The seller doesn’t actually have any books. You order a book and the seller tries to buy it from another seller who has priced it reasonably. The book might be shipped directly to you from the third party but usually is re-mailed by the original seller. If the wanted book isn’t available you will get an apologetic form email telling you of some listing error. None of this would make any sense but for the ability to automate every step in the process. Some sellers make money off of inflated postage and handling costs. The book is irrelevant. That’s why you see so many one dollar listings with $4.95 shipping and handling.

Some of these sellers go bust. Not an uncommon end point even for higher end book sellers. The inside joke: “How do you make a million dollars in the used book market? First, you start with two million dollars.”

Oh, nostalgia for the old days of cranky, eccentric book sellers in musty shops with indolent cats.

41

tomslee 11.26.11 at 2:08 pm

Bloix and Salient – points taken, and maybe I am overreacting, but what is bothering me is that commenters seem (to me) to be looking at the price from the point of view of the coder, and accommodating their expectations – sympathetically – to the cost/benefit calculations that coders make.

“My algorithm made me do it” excuses are not uncommon: they have been offered by Google for the Michelle Obama monkey picture and by Yahoo! for blocking emails about Occupy Wall Street, so they can be significant. Algorithms embody the cultural and other expectations of the coders or their employers (in these cases, of well-off 30-year old Californians) and those who make money from them need to take responsibility for the output.

In Salient’s sentence “output that is not desired and not valued, but not worth fixing because the thing works ok enough to get by [and because the thought of another round of debugging gives you shudders]” you are putting yourself in the position of Quick-N-Easy, and forgiving their laziness. Bloix’s description of Quick-N-Easy Marketplace is convincing that there is no human intelligence at work in the Q-N-E $999.99 price, but applies equally to all their prices, including the reasonable ones.

Personally, I see Amazon and Dr. Hilarius’s ““mega sellers” who have no relationship to book sellers” as having a net bad impact on the culture of books, and I’m not inclined to look at the world from their point of view or put in intellectual effort to learn how to interpret the foibles of their crappy coding, share such expertise with others, help to build up a culture that can navigate their algorithmic world, and so help Amazon more quickly put my neighbourhood bookstore out of business. And I think it is a political decision to make that intellectual effort.

42

Colin Danby 11.26.11 at 11:50 pm

All the arbitrage opportunities have been taken. Markets are efficient!

43

parsimon 11.27.11 at 2:28 am

So you can try to sell a book figuring that $20 would be fair and Amazon advertises it for $645

Salient up at 16, I’m 99.9% sure this is false. Perhaps this has been pointed out. Sellers’ prices aren’t upped by Amazon. Rather, the megasellers or dropshippers Dr. Hilarius describes in 40 are using automatic pricing software; in the case of insanely high prices, the software reveals that no other copies are currently available on Amazon, so they’ve arranged their settings such that an absurdly high price is put in place. A second megaseller’s software automatically drops that price by 5 cents, or 1 cent, or 50 cents. It is absurd.

Amazon’s trade-in offer (sell this book back) is more interesting: they really will accept your shipment to their warehouse of the book in reasonable condition, for that offered trade-in price. At least, at the moment. The trade-in price will change on an ongoing basis, depending on Amazon’s metrics regarding the book’s sales rank. Amazon isn’t really making money by buying semi-beat up books; notice that you will receive not cash but an Amazon gift card. It’s really very clever: this way you have to spend that money on Amazon.

44

parsimon 11.27.11 at 2:52 am

Further remarks on that last: this way Amazon also dissuades you from selling the book yourself (at more than their trade-in price), and scoops up used books that might otherwise go to the second-hand bookselling market. All while driving you back to spend your trade-in dollars on Amazon. These things are very much worth a loss to them.

45

Salient 11.28.11 at 5:06 am

Salient up at 16, I’m 99.9% sure this is false. Perhaps this has been pointed out.

Yep, I was wrong, a few folks explained the mechanism upthread — before that my only sources of information were a few sellers that I had contacted to ask about their price and happened to receive a reply from; they tended to say something vaguely apologetic, like ‘Amazon is quirky’ — it just never occurred to me that sellers would intentionally build an algorithm to sets their price above the highest previous price. I’m still not sure how obscure items with no other sellers get their prices set stratospherically, but presumably it’s due to the same seller-employed algorithms.

what is bothering me is that commenters seem (to me) to be looking at the price from the point of view of the coder

Well, ok, FWIW I’m definitely unnerved by the existence of automated sellers-of-everything advertising prices for items they don’t even have, and I feel quite a lot more irritated at sellers of obscure items who apparently blindly put stuff up for sale without having it in inventory (and the likelihood that the times I’ve been misleadingly told there was a ‘listing error’ it was actually a phantom product is beyond irritating).

46

Tim Wilkinson 11.29.11 at 5:19 pm

From the (very odd, un-human) point of view of pure profit-maximisation, an asking price like this doesn’t seem particularly odd to me – if the price is subsequently to be reduced after a certain amount of time unsold, anyway.

If you’ve got a book and, so far as your lo-cost research-lite tells you, it is the only copy in existence and you don’t know how much someone may be willing to pay for it, then you start off asking a very high price. Effectively a descending auction – potential buyers also don’t know who else is out there who wants the book and has loadsa money, so even if they know the price will gradually descend, at some point after it becomes affordable relative to their prefs, budget and price of alternatives they will, you reason, snap it up so that no-one else gets in first.

The issue of why, in LP’s example @20, the price went up is more tricky. Good reasons for this (which seem to be lacking) would I think have to include some info about the prospective buyers.

A classic case would be when there’s only one buyer as well as only one seller, and both know this – so they’re potentially in a stand-off of a familiar kind*. Making some kind of credible commitment to increasing the price (assuming the buyer would still prefer to buy at the new price rather than forego owning the the book altogether), with the implication that further price rises will follow, might break such a stand-off. Legend has it that Tarquin the Proud got on the business end of some of that treatment. (The standoff gets broken by appearing to be committed to a bold strategy that’s potentially locally irrational – which is not the same as convincing anyone that one is in any more substantial or general sense irrational, still less that one is stupid. Contrast the recent ‘Folk Theorem’ post).

Of course a lot here depends on what is and isn’t known about other pontential buyers and sellers, their reserve prices, strategies etc.

*Doing some trade is ‘positive sum’ (and win-win) compared to not doing any – so at that gross level of granularity, agreeing a price is a co-operative game. (Market types like to leave it at that of course, with everyone living happily ever after and no complaining.) But the cooperative game becomes a very tricky co-ordination problem – a stand-off – because once the feasible set of contracts is given, the subgame of bargaining for a specific price is a zero-sum – thus rival, competitive – game.)

47

Ginger Yellow 11.30.11 at 6:55 pm

Tomslee, I don’t understand why you’re getting so aggravated by the commenters (as opposed to people advertising things for sale they do not have for sale).

I’m not inclined to look at the world from their point of view or put in intellectual effort to learn how to interpret the foibles of their crappy coding, share such expertise with others, help to build up a culture that can navigate their algorithmic world, and so help Amazon more quickly put my neighbourhood bookstore out of business.

I’m really struggling to see how Amazon (or rather a seller on Amazon) failing to sell a colouring book for $1k is going to put your neighbourhood bookstore out of business. The only people suffering from the silly pricing products of these algorithms are the sellers themselves, and possibly anyone who accidentally purchases them at the inflated price.

48

tomslee 12.01.11 at 12:44 pm

GY: Yes, algorithms that gave fewer silly prices would probably do a bit better for sellers, but there would be a cost for that improvement. To the extent that we educate each other about how to navigate the quirks of these algorithms, we are subsidizing these sellers.

49

bianca steele 12.01.11 at 2:48 pm

Tom (this is the edited version of this comment),
I’m confused by your second paragraph @41. I’m not sure how I could say why in the space of a week’s worth of blog posts, much less a comment box, but I’ll give it a shot. If I have simply misinterpreted your words, and you meant something different from what I’m arguing against, just say that. If so, I apologize in advance.

First: Algorithms belong to the arena of math and science. Maybe you are saying something along the lines of, “science is politically tinged from the very beginning, and even Newton’s laws embody the political interests of Newton’s society,” or maybe you are saying something less far-reaching than that. Probably algorithms can be divided into general (how to calculate R), and content-specific (how to render a sphere in pixels), but neither of these is application-specific in the sense of being specific to economic analysis programs or videogames. They are neutral. Maybe you are saying something along the lines of, “engineers shouldn’t work for corporations, the way they shouldn’t work for defense researchers, because even the nuclear scientists have supported strong ethical practices for scientists and engineers whose work might be misused,” but in that case, the algorithms stuff seems irrelevant to the rest of the argument.

Second: Your argument sounds more like an argument about literature than about program code. Program code isn’t exuded from the programmer’s soul the way you could say poetic metaphors are. It’s a tool.

Third: As you well know, the number of people in an organization who make decisions about what product is going to be sold is small, by comparison to the number of people who work for the corporation. And they’re not always programmers themselves, in fact the stereotypical Silicon Valley start-up that’s founded by a computer science major turned marketeer is pretty unusual.

I am a little worried that this is going to be misunderstood, but so.

But you might as well blame Cosma Shalizi for basing a story on economic principles you disagree with as blame a programmer for implementing an economic algorithm that was invented by someone else.

50

Watson Ladd 12.01.11 at 3:14 pm

Tomslee, the only navigation taking place here is choosing to buy or not buy the book at the offered price. Its no different from any time you get an offer to buy or sell something. Unless you think we need special knowledge that the book isn’t worth that amount in the case amazon offers it vs. a used book seller, the argument doesn’t work.

Comments on this entry are closed.