Recently, for the Internet Archive, I have been on the apparently quixotic quest to buy books from publishers. What gives this (sometimes epic) quest its quixotic flavor is that we are actually trying to buy these books, not subject them to 40 page license agreements. There are a couple of reasons for that, the most simple one being that we think libraries (like the Archive) have done perfectly well buying books and lending them out — for generations — and heck, it kinda seems to us like that model works.
I’ll also note, on this point, that for publishers, cutting licenses with each new distributor or retailer is a major pain in the rear end. Every platform expects something different, after all, and so today publishers must not only be experts on author and agent contracts, they have to figure out the whys and wherefores for increasingly complex, convoluted license agreements with a growing number of business partners for digital products whose own complexity grows with every month. That works well, huh?
But I digress.
One of the most common responses I get from publishers when I tell them I want to acquire their books for the Archive (once I explain it adequately) is a nicely-put response that buying books is certainly an intriguing idea, but “we’re not really set up for sales like this, have you tried our asking our distributor?” In other words, handling individual sales is a very painful, high threshold task, and publishers only want to accommodate “high revenue” arrangements.
Now the engineering part of me finds this a truly odd response, at so many levels. I mean, these are books, after all. Whether digital or print, these are ultimately consumer goods. They are not computer-controlled machine lathes. I could understand in part if the redirection to a distributor or retailer was because publishers dealt in great bulk with physical goods, and it just didn’t make sense to respond to individual consumers. That’s the paper towel model. I don’t make a habit of buying Bounty paper towels from the manufacturer; I get them at Costco or Target.
But ebooks aren’t like that. They are digital goods; I don’t need either Amazon or the manufacturer to ship them to me via UPS Super Saver. And as a consumer, from an engineering perspective, my ideal interaction with a publisher trying to vend, realistically, a small number of copies of a title in order to not have too many books cutting away from the profit of the very few books that get movie deals should be seamless and straightforward. As Brewster Kahle of the Archive observed to me, it should be rather more like going up to a vending machine and buying cans of soda. I want one of that, two of that, twenty of this. Okay, maybe not soda. But you get the idea.
In engineering land, what this would imply is … wait for it … an API. An automated interface that would permit the purchase of a book by any party (human or code) in whatever quantity they wished, in whatever format they wished, as long as whatever arcane territorial restrictions and contract clauses did not override the desire of the reader to part with their money to help make the author wealthier, happier, and in a position to write more books.
But its not like that. And one reason that buying books is not like buying towels, and there are no APIs, is that publishers have to spend extraordinary amounts of time preparing custom ONIX feeds, metadata bundles, and format packages for every distributor and retailer of note. Sometimes even multiple formats for a single retailer (looking at you, Amazon). Which leaves publishers utterly unable to take cash from readers, because they neither have the organizational slack, nor have they developed the expertise to write APIs, present title information in OPDS Catalogs, and augment web discovery via schema.org.
In other words, publishers spend the majority of their time on filling the supply chain, customizing the requisite data flow with every business partner, instead of focusing their engineering on what they are actually selling, which are books and – more important – the experience the reader gets when they read the book. And that’s kinda insane.
It’s like publishers are selling turbines for a new power plant. An incredible amount of customization goes into each opportunity for selling books, in large part because publishers have never stood together and told the retailers, “You’re getting EPUBs through an OPDS Catalog. Period.”
The other wrinkle with this is that if publishers worked this the right way, they would start to build relationships with readers. Publishers don’t actually have to sell direct to consumers to be able to touch or hold that relationship, although they could easily, once they standardized the supply chain and normalized product delivery. With set standards for purchase via an API, they could force an Amazon customer to come back to the publisher website (or repository) to obtain the title with a little train of useful information on the redirected URL.
Really, it could be so much simpler. All it takes is a set of URIs, an API, and the web. Oh, and resolve.
