Henry’s post was interesting. It reminded me of an anecdote passed along by an acquaintance, who shall go nameless.
The individual in question is involved in publication of limited run, high quality art books. You can’t do that if you can’t make significant profit, per unit. (‘Volume volume volume!’ doesn’t hack it if you lack volume.) Medium-length story short:
This person – small fish – backed out of a deal with medium-sized fish, when it became painfully clear medium-sized fish’s deal with Leviathan – Amazon – made the margins too small. That’s business. But, in the aftermath, small fish’s sold-out limited editions are Currently Out of Stock, Available Soon on Amazon, in perpetuity, at a steep discount. Which is nonsense. Those orders Amazon is taking will never be fulfilled. Now, Amazon isn’t taking anyone’s money on false pretenses. It’s not illegal, nor should it be – but it is odd – for me to offer to sell you an apple for 5 cents, even though I’ll never have that apple, so you’ll never give me the nickel. In almost all cases, Amazon’s uncertain promises of Available Soon are reasonable, and welcome to suppliers as well as customers, and that’s basically why they do it. But Amazon does not lightly cull old titles from the lists. This puzzled me for a long time. [UPDATE: per comments below, the following anecdote is misremembered. My book was completely unavailable for 3 years then, when the new edition came out, Amazon upped the old edition to from unavailable to ‘available in 1-2 months’, which was an interestingly warrantless surmise on their part. Presumably what happened is some 3rd party seller put some copies of the old edition up for sale, seeing the new edition available, and Amazon perked up about the old. Leviathan responds to motion and does nothing in response to no-motion.] On a personal note: it was only last month (!) Amazon stopped offering to take orders for the previous edition of Reason and Persuasion. That book was ‘available soon’ for years – more than 4 years I think? – after the previous publisher let it go out of print. I didn’t really mind, even after the new edition came out. New edition cheaper. But if it had been the other way around – if I had replaced a cheap edition with an expensive, newer edition – it would have been inconvenient. Buyers would maybe keep waiting for a cheaper Godot who’s never coming. Which gets me back to small fish. Small fish would like to make fresh runs of books. But it cannot be done at anything like the price point Amazon is promising. But everyone looks to Amazon to know what a thing ‘should’ cost. So small fish has a problem. If YOU saw a publisher offering an art book for sale for their site for $75, while Amazon says it’s ‘available soon’ for $45 – ‘expected in 1-2 months!’ – what would YOU do?
This isn’t a problem except for the small set of people in something like small fish’s peculiar economic position. (And there’s a quick fix that does quite a bit of good. Publisher posts on their site that ‘seriously, Amazon price ain’t real! $75 or nuthin’! Final offer!’) So this doesn’t crack the Top 100 of Potential Negative Externalities of Amazon Economics, even. But it’s a case of effective, algorithmic monopoly pricing with curiously distinctive, digital-age features. Little guy hit by backstroke from Amazon’s price-slashing in ways that might not be obvious.
Once your product is on Amazon, if you back out, Amazon gets to declare, forevermore, what it SHOULD cost. And probably everyone will believe Amazon.
{ 31 comments }
bruce wilder 01.15.17 at 2:28 am
Amazon does a profitable business providing a marketplace for third-party sellers willing to do third-party fulfillment. They keep the out-of-print books on the system as salients for these used book sellers. Your friend should offer his new product on Amazon with third-party fulfillment as rare out-of-print editions in mint condition. Then undercut his Amazon price on his own site for “new” editions. Confuse everyone by auctioning occasionally on ebay.
https://en.wikipedia.org/wiki/On_the_Internet,_nobody_knows_you%27re_a_dog
John Holbo 01.15.17 at 2:58 am
” Your friend should offer his new product on Amazon with third-party fulfillment as rare out-of-print editions in mint condition.”
Thanks, I think that is something like the actual plan at this point. To use Amazon but only the marketplace. I don’t really know this person, so I don’t really know.
“They keep the out-of-print books on the system as salients for these used book sellers.”
That I get. What I don’t get is the ‘usually ships in 1-2 months’ notice still there after 48 straight months of no product.
mjfgates 01.15.17 at 6:16 am
Probably nobody’s asked for the book. No requests, no ships, zero over zero is one, so it’s shipped within 1-2 months EVERY TIME amirite?
John Holbo 01.15.17 at 6:37 am
“No requests, no ships, zero over zero is one, so it’s shipped within 1-2 months EVERY TIME amirite?”
You are right, come to think of it.
Pavel 01.15.17 at 7:21 am
@mjfgates
I doubt Amazon’s programming is so utterly shoddy. No Orders != sum of shipping times / sum of orders shipped. Also 0/0 == undefined. I may be wrong; I’ve made even more embarrassing mistakes in my short programming career.
Manta 01.15.17 at 10:45 am
Obligatory xkcd.com/325/
aljabr 01.15.17 at 11:04 am
Zero over zero is not one, it’s undefined. Though somehow I doubt you’d ever see Amazon list that as the expected shipping date.
John Holbo 01.15.17 at 12:15 pm
Sorry, I was reading it loosely as a 100% not-fail-to-deliver-in-1-2 months rate. I wasn’t focusing on the actual division-by-zero calcuation, which is obvious not right.
Lee A. Arnold 01.15.17 at 1:07 pm
Division by zero of any number that is NOT zero is undefined.
But! zero over zero is NOT undefined. It can be any number: b x 0 = 0, therefore any b = 0/0.
I think this was pointed out in the Medieval period.
Tom Slee 01.15.17 at 3:32 pm
“Your friend should offer his new product on Amazon with third-party fulfillment as rare out-of-print editions in mint condition”
Following this advice, Friend’s job is no longer just producing and supplying a book, it is also becoming savvy about the Amazon ecosystem, managing their presence on Amazon, and gaming the Amazon experience — even while Friend doesn’t actually want to sell through Amazon. Maybe this is the best path for Friend, but let’s just acknowledge that not all people involved in high quality art books have this skill set, so the advice for those people turns into “give up” which would be a shame.
Tangentially related: I had a conversation with somebody a few months ago about a book I wrote, which is available on Amazon as an ebook but which they thought (incorrectly it turns out) was made available only with DRM. They objected to the DRM but accepted Amazon as a natural fact of life. My priorities are the opposite.
Pavel 01.15.17 at 6:08 pm
@Lee A. Arnold
In math yes, in computation it is considered undefined behaviour or more typically a run-time exception. This is because mathematical expressions typically have a state of endlessly delayed or “lazy” evaluation. You don’t need to know what 0/0 or b actually is, just that there exists such and such a relationship with such and such properties. In computation, those relationships generally cannot exist without evaluation, since in computation what you are doing is assigning instead of relating (b is literally the thing that 0/0 evaluates to). So when the language chooses to evaluate 0/0 it either has to settle on some random value (hence leading to undefined and non-deterministic behaviour) or throw up its hands and tell the programmer to do a better job (the exception).
The one exception (no pun intended) is when computer languages themselves use lazy evaluation. In that case, so long as the value b is not used in any further computations, the relationship can be retained indefinitely. However, the moment you actually need to use b as a concrete value, you will still evaluate 0/0 to undefined or an exception.
What I find fascinating is that math is able to exist in many cases as a completely lazily-evaluated system. A collection of relationships and transformations suspended in some aether that never actually resolve until such time that you need them to resolve.
George Painter 01.15.17 at 6:21 pm
Let us assume that 0*x = 0 for any number x. (Proof is about six lines). Also, let us assume that 0/0 = 1. We may write this as 0*(1/0) = 1. It follow that 0 = 1. We have a choice: either 0/0 is undefined OR 0 = 1. Perhaps your choice depends on you politics.
Patrick 01.15.17 at 6:23 pm
If he sells it as a mint condition limited edition out of print book, make sure he doubles the price.
Collin Street 01.15.17 at 7:34 pm
No orders means it doesn’t shift from the default.
J-D 01.15.17 at 7:56 pm
If division is defined as multiplication by the reciprocal, then the definition does not include division by zero, because zero has no reciprocal; and if what it means to divide by zero is not defined, then what it means to divide zero by zero is not defined.
derrida derider 01.16.17 at 12:10 am
I’m a bit intrigued by “Small fish would like to make fresh runs of books. But it cannot be done at anything like the price point Amazon is promising.” The cost of a new print run, even large format glossy ultra-high-res printing, would be a tiny fraction of the original setup and marketing costs, and certainly far smaller for low volumes than 20 years ago. Knowledge of that is no doubt exactly why Amazon is offering a lower price for the new run, so either small fish or Amazon is wrong in their printing cost estimate. In which case it’s not a market failure due to externality – just simply a mistake by someone.
Maybe small fish has confused AVERAGE cost and MARGINAL cost. $75 was a necessary price for the original print run to cover that setup and marketing, but now it’s sold out and unavailable a new print run at well under $45 should be a nice little earner. The fact that that will drive down prices of used copies of the first print run is neither small fish’s nor Amazon’s problem.
Factory 01.16.17 at 12:43 am
@Pavel
Another common way of computers representing n/0 is to return a NaN (not a number).
To most programmers this will translate to: ‘This code is screwed, provide the /0 cases with a number that makes sense to the users’.
Raven Onthill 01.16.17 at 1:11 am
“Amazon’s thing-like-a-contract continues to be deeply alarming, while their payment mechanisms continue troubling.” — Graydon Saunders, http://dubiousprospects.blogspot.com/2015/05/committing-book-again.html
Raven Onthill 01.16.17 at 1:25 am
And, damnit, it’s a market, not an ecosystem. Nothing squishy about it.
John Holbo 01.16.17 at 5:24 am
“Maybe small fish has confused AVERAGE cost and MARGINAL cost. $75 was a necessary price for the original print run to cover that setup and marketing, but now it’s sold out and unavailable a new print run at well under $45 should be a nice little earner.”
The books in question are pretty unusual items. Unique art-object design that goes beyond nice color printing. Also, small fish is determined to have production done state-side: ethical concerns about Chinese slave labor. Small fish knows the difference between average and marginal cost.
Guy Harris 01.16.17 at 5:47 am
Factory:
…except for the programmers who just let “NaN” pop up on Web pages.
Bill Benzon 01.16.17 at 12:12 pm
” Small fish knows the difference between average and marginal cost.”
Could almost be a line from Backstroke of the West. Probably needs a bit of rearranging, though. For example: “Average and marginal cost between the difference small fish knows.”
JimV 01.16.17 at 6:36 pm
sin(x)/x as x goes to zero = 0/0 =1; mathematically, whether 0/0 equals anything and if so what it does equal depends on the paths the numerator and denominator take to get to zero. As Lee Arnold said, it is possible for f(x)/g(x) as both f(x) and g(x) go to zero at the same x to be any number b (including b=0). E.g., f(x) = bx and g(x) = x. (Or f(x) = sin(bx), if that seems too artificial.) It is also possible for there to be no limit – no fixed number which the path is heading toward.
What happens in a computer program has been well and correctly stated here previously. However, if our universe is a computational engine, it seems to handle sin(x)/x at 0/0 without a glitch – if our mathematical models of it are correct. Like Zeno, I sometimes suspect we are over-complicating things with our insistence on continuous variables, although I grant they are often convenient for solving equations.
Sorry, I was tempted to comment on this yesterday and successfully resisted the temptation. Today I failed (but perhaps the moderator will not).
derrida derider 01.17.17 at 3:01 am
Oh well, if each book is an illuminated manuscript in intaglio gold leaf with a hand-tooled vellum cover then all bets are off. The economics of that have not changed.
To produce them she definitely needs slave labor (monks or Chinese convicts – makes no difference) and to sell them she needs a transnational elite engaged in competition for prestige goods and expecting personalised sycophancy from the seller (mediaeval Church or today’s plutocracy – makes no difference). The knowledge that the transaction required extensive kowtowing by lesser mortals is of course an essential part of the prestige aspect of the good.
Amazon’s really got nothing to do with it.
John Holbo 01.17.17 at 4:38 am
“The knowledge that the transaction required extensive kowtowing by lesser mortals is of course an essential part of the prestige aspect of the good.”
Well, it might be in part love of art. (Back of the envelope calculation.)
J-D 01.17.17 at 7:07 am
Jim V
There is a precise mathematical definition of ‘the limit, as x tends to zero, of the quotient of sin(x) and x’, but it’s a definition of the meaning of just that, not a definition of ‘the quotient of zero and zero’. To say ‘if you divide zero by zero you can get any number you like’ is to fail to provide a definition of what it means to divide zero by zero.
dsquared 01.17.17 at 11:31 am
sin(x)/x as x goes to zero = 0/0 =1;
“the limit of f(x)/x as x goes to zero” is not the same as “f(x)/x, evaluated at x=0”.
Bartholomew 01.17.17 at 7:01 pm
‘You know the formula m over naught equals infinity, m being any positive number? [m/0 = ∞]. Well, why not reduce the equation to a simpler form by multiplying both sides by naught? In which case you have m equals infinity times naught [m = ∞ x 0]. That is to say, a positive number is the product of zero and infinity. Doesn’t that demonstrate the creation of the Universe by an infinite power out of nothing? Doesn’t it?’
‘Well, began Lord Edward, and at the other end of the electrified wire, forty miles away, his brother knew, from the tone in which that single word was spoken, that it was no good. The Absolute’s tail was still unsalted.
Aldous Huxley, Point Counter Point (1928)
derrida derider 01.17.17 at 10:36 pm
John Holbo @25: Mate, aesthetes are far too small and impecunious a market. Ya gotta go where the money is. Best to sell them as coffee table books for guests to peruse on those tedious journeys in private jets. /trolling
It’s all a bit reminiscent of that scene in “Hannah and Her Sisters” where the rock star wants the neurotic painter to quote his prices by the yard in colours to match his new mansion’s walls. To be clear about this Woody is poking more (and crueller) fun at the pretentious artist than the vulgar rock star.
Lee A. Arnold 01.18.17 at 1:22 pm
Pavel #11: “run-time exception”
I like that — it’s a good & loaded phrase! We could almost define mathematics as an experimental science of symbolisms that reject run-time exceptions (up until they could be bracketed in the infinitesimal calculus). In the Medieval period there was a lot of disputes about how symbols should be made to represent the most basic operations — what’s allowed and what isn’t. We forget that math wasn’t handed to us as a pure prior understanding. More recently, “run-time exceptions” would make a good inroad to understanding Cantor, then the 20th-century intuitionists, and then Wittgenstein’s objections to Cantor. Why should denumerability characterize the “size” of an infinity? Because you can’t match non-denumerability one-to-one with denumerability? This sounds like a run-time exception, not a real “size”.
John Holbo 01.19.17 at 1:43 am
I just realized I am remembering wrong about my old edition. Not that it matters (and I don’t know how I forgot.) For several years 2012-15 the book was out of print hence unavailable and listed as such on Amazon. No 1 – 2 months usually available notice. (I don’t know how I forgot this because I had long and fruitless negotiations with the publisher over this significant barrier to sales!) The publisher let it go out of print. But then, after the new edition came out and enjoyed some significant sales, Amazon re-upped the old edition to the status of ‘usually available in 1-2 months’. That’s where it has been for the last 12 months. That was slightly annoying, per above – but not very annoying – and somehow my annoyance erased my previous 3 years of annoyance at my failure to see any Amazon availability (for which I blamed the publisher). Oh well. It was never a big deal. Perhaps it all has to do with the availability of 3rd party seller copies. If there are 3rd party sellers selling then Amazon gets hopeful and says ‘1 -2 months’ in the hopes that publishers will get on the ball.
Comments on this entry are closed.