Misunderstanding Book Returns - apps4work/co.a4w GitHub Wiki

Printing a copy of a book and then pulping it, is an obvious waste of money and resources. Others things being equal, it should be avoided.

"Book Returns" run at the order of 50% of the copies printed. If this could be avoided, other things being equal, there is a substantial savings in money and resources.

These two statements are not the same, and the clue is in "other things being equal". You can work out the difference by understanding all the mechanics of the book industry supply chain, but there is another way to understand what is going on.

Firstly it dangerous to assume that people don't know what they are doing. Well-established long-term surviving publishers run at returns of the order of 50%. They know that they are doing. Doing it is obviously not a fatal flaw in their industry. Otherwise they'd be dead. But they may not know why they are doing it, because evolution doesn't give out brownie points for knowing why, but only for doing what enables you to survive.

Secondly, let's look at the industry in a different way. Ask "What if it were free?". Suppose that the cost of manufacturing an additional copy of a book was zero. How many copies would a rational publisher produce? The answer is infinity and his return rate would be asymptotically close to 100%. Infinity is hard to deal with. So suppose that the cost was just above zero; at least enough so it was worth having procurement look at the grand total. They wouldn't order infinity, but they'd have to ask "when is enough". What would be the criteria to stop making copies even though the cost is negligible?

sidebar "What if it were free?" is an example of a thinking technique, from the general set called "Outlier Analysis". If you want to understand the effect of changing some parameter, plug in an extreme value and see you can figure what would happen, and then back off the extreme value and see if what you observe the extreme effect existing, probably less extremely, with the less extreme parameter. If is very useful in Physics. consider the vexing question: if you holding a heavy object while sitting in a boat on a lake, and you throw the weight overboard, does the level of the water in the lake go up or down or stay the same. You can do all the Archimedes stuff, or you can ask what would happen if the heavy object was a Black Hole. In about 2000, the question "what if it were free?" to contact your customer (forecasting email taking over from snail-mail). The typical executiove response to thquestion was to argue about whether email is actually free (it isn't; $2/million emails isn't free; snail-mail costs about $2/piece), but the correct response would have to predict Phishing attacks.

When do you stop making copies of your "Lost Cat" poster? When there is no where else to post them. When do you stop making more copies of your book, when there is no where to put them that has any chance of being sold. When the cost is tiny, but not zero, you put them everywhere you can where the probability of sale covers the tiny cost of the copy.

Publishers are behaving rationally in response to the economics of book manufacturing were the cost of manufacturing a copy is a small fraction of the retail price. The 50% return rate is a reflection of the number of places that exist that have a probably of a sell that justifies the (small) cost of placing a copy in there. The reason published appear to largely unconcerned about return rates, is that they actual don't care about return rates.

Of course [On Demand Manufacturing](On Demand Manufacturing) turns this upside down.

One of the problems of getting publishers to behave rationally (in their own interests) for On Demand books was their lack of understanding of why they have 50% returns, and, in particular, the management structure in place that reflects that reason. Of course they have someone who is motivated to avoid spending $infinity on copies of the book, but who is generally motivated to provide a large number of copies at least cost, without much concern about returns. That person looks at On Demand Manufacturing and rightfully scoffs at it. It is so absurdly expensive compared to "one more copy" in long run manufacturing, he is not even going to talk about.

The second problem is the guy that should be talking about it: the executives concerned with sales. They were concerned with the sales of the next Harry Potter or Stephen King or John Grisham. They are looking at sales of millions of copies of one title. Why should they be interested in the sales of title that might amount to a dozen copies in a year?

Publishers are skilled at selecting, publishing and promoting Best Sellers. Why would they spend their time on a title may sell a dozen copies in a year.

And why should they listen to a data-nerd who says that they should treat the sales of books as random numbers?

While publishers may be great at publishing and promoting a Best Seller, the available evidence suggests that they are appalling at selecting Best Sellers. A single anecdote is sufficient to disprove the hypothesis that publishers can predict book sales: John Grisham was turned down by 28 publishers before one agree to publish his. Many best seller authors have similar stories. A huge proportion of published books turn out to be sales dudes.

It is likely (through not yet provable) that the best mechanism for a publisher to select which books (of unpublished authors) to publish would be random selection. Which is insulting to those whose job it is to select them. Even though it is true.

For titles whose most recent print-run is now exhausted and the sales rate is small, it is provable that the best strategy is to treat sales as random numbers. That is insulting too.

And a dumb pricing mistake made that harder.

Suppose I offered you the change to play a game. You guess a number between 2 and 12, and I roll a pair of dice, and if the total matches your number, I will pay you $1. That's the whole game. You can play the game as many times as you want. What's your strategy? Do you need to work out the probably of guess the number right, in order to decide on your strategy? Do you need a strategy for choosing the "guessed" number? No. You just play the game as many times as you can. Sometimes you win $1; sometimes you don't win $1. There's no downside. You play the game a million times, you get some fraction of a million dollars.

But if I charged you to play the game, if I charged you different amounts for the number you choose, or paid out different amounts depending on the number that came up, or if I offered you many different patterns of charging, now you have a decision to make. Not only to you have to decide which choices are best, you also have to decide whether it is worth playing the game at all. Even if you could work out that it might be cash positive, you have have to guess whether its worth figuring it out. You have bigger fish to fry.

So the tactical pricing mistake was to make it a decision. To price it so that it might not be profitable, or so that it would profit only if the publisher "chose" the right titles. Of course it flattered the publisher to imply that he could make that choice, ignoring the reality that he could not. It insulted the publisher by implying that his extensive industry knowledge was no more valuable than random number generator, even though that was true.

"what if it were free?" - if cost nothing to make a title available "On Demand", you should roll the dice as many times as you can. It is not a decision. Sometime you will win and sometimes you will not win, but you will never lose.

Of course, making it cost nothing to make a title available "On Demand" was a considerable effort but really piggy backed on the movement of all books into digital format, perhaps spurred by e-books, but also a necessary part of modernizing the book industry. Making it cost nothing means making all the work be done by computers and not people. ("If we touch it, we lose").

Making it cost nothing, but pricing above nothing is a strategic blunder akin to Google charging for Web Searches.

Homework: how does this relate to the Garment Industry?

Is the cost of manufacturing the garment actually the factor that should dominate the industries thinking?
Isn't it closer to "free"?
Do we fill any shelf we can with product that has a chance of selling?
We don't pulp our returns, we pulp our margins.