How Books Were Sold Before Amazon

In 1476, William Caxton set up England’s first printing press in the precincts of Westminster Abbey, and within a year was selling printed books to anyone who walked in off the street. This was genuinely unprecedented. Before Caxton, books were manuscripts — hand-copied objects of enormous labor and cost, owned by institutions (monasteries, universities, wealthy nobles) and largely inaccessible to anyone outside those networks. Caxton’s press could produce in a week what a scriptorium of monks needed months to copy. The price of books did not fall by 10% or 20%. It fell by something like 95% over the next century as printing technology spread and competition intensified.

What the Gutenberg press actually changed is frequently misunderstood. It did not create the idea of books. It created the economics of a book market — a situation in which books could be produced at costs low enough to sell to individuals rather than institutions, and where a printer who could identify what people wanted to read could make a substantial commercial return. This is the founding condition of the entire publishing industry: a technology of mechanical reproduction that makes intellectual content into a commodity with near-zero marginal production cost and an extremely uncertain demand function.

The early book trade was organized as an integrated model: the printer was the publisher and the bookseller, operating from the same premises. Caxton wrote some of his books himself, translated others, and commissioned still others from writers. He made editorial decisions, production decisions, and distribution decisions. He also bore all the commercial risk. This integration made sense when the technology was new and the market was small — the printer who could judge what would sell was the same person who decided what to print and what to charge. The functions hadn’t yet separated because the industry wasn’t large enough to sustain specialization.

The economics of early printing were genuinely difficult. Books required significant upfront capital: paper, ink, a press, type, and the skilled labor of compositors who set each page character by character. A print run of 500 copies — typical for the period — required all of this capital to be committed before a single book was sold. The early printer-publisher was therefore a venture capitalist making bets on which texts would sell, with the entire cost sunk before any revenue materialized. The high failure rate of early printing businesses reflects this risk structure. Many pioneering printers went bankrupt.

The institutional solution to this risk problem in England was the Stationers’ Company, which received its royal charter in 1557. The Stationers were a guild of London printers, booksellers, and bookbinders who obtained effective monopoly rights over printing in England — any book had to be registered with the Stationers’ Company to be legally printed, and Stationers’ members had exclusive rights to the texts they registered. This was, functionally, a cartel that controlled book production and distribution in England for over a century. The Crown liked it because it enabled censorship — any seditious or heretical text could be identified and suppressed through Company enforcement. The Stationers liked it because it eliminated competition and guaranteed profits to existing members.

The Stationers’ Company model is the first version of copyright in English history, though it wasn’t called that and it worked very differently from modern copyright. The intellectual property right resided with the publisher (the Stationer who registered the text), not the author. Authors had no legal rights to their works once sold. A writer who sold a manuscript to a publisher had sold it permanently — no royalties, no reversion of rights, no share in reprints. John Milton sold Paradise Lost for £5, with a provision for an additional £5 if the first edition of 1,300 copies sold out. It did. He received £10 for one of the most important works in the English language.

The Statute of Anne in 1710 is the actual founding document of modern copyright law. For the first time, it vested the copyright in the author rather than the publisher, with a term of 14 years renewable once. This was a revolutionary restructuring of intellectual property economics. It didn’t immediately benefit authors financially — publishers retained enormous bargaining power and authors typically sold their copyrights outright rather than licensing them — but it established the principle that the creator of intellectual content had a property right in it, which is the conceptual foundation of the entire subsequent system.

The publisher as a distinct commercial entity — separate from printer and bookseller — is largely a 19th century creation. As the book market expanded and the economics of specialization kicked in, the functions separated: printers who manufactured books, publishers who selected, edited, and financed them, and booksellers who sold them to the public. The publisher occupied the middle position and extracted value from both sides: negotiating down the price paid to authors and up the price charged to booksellers. The advance against royalties — a payment to the author before the book earns out, which the author repays from subsequent royalty earnings — became standard practice in the late 19th century. It aligned publisher and author incentives (both want the book to sell) while ensuring that the publisher bore the initial risk.

The pre-Amazon economics of trade publishing were built around a set of practices that look irrational from the outside but made consistent sense given the information structure of the book market. The key fact about books is that they are experience goods: you cannot assess the quality of a book before you read it, and reading it is the experience being sold. This creates a fundamental information asymmetry. The author knows how good the book is (mostly). The publisher knows after editing and production. The reader doesn’t know until after purchase. Everything in the economics of traditional publishing is a response to this asymmetry.

The returns system — which allowed booksellers to return unsold inventory to publishers for full credit — was economically bizarre from a standard retail perspective. No grocery store returns unsold milk. But books aren’t groceries. They’re experience goods with highly uncertain demand. The returns system enabled booksellers to stock broadly without bearing demand risk, which meant more books were stocked, which meant more books had a chance of being discovered by browsing customers. The cost was enormous: publishers estimated and managed their returns rates (typically 30-40% of shipments came back) and priced accordingly. The whole system was a mechanism for managing uncertainty about which books would find their readers.

The bestseller cross-subsidizing the midlist was the internal publishing equivalent of the same logic. Publishers made reliable profits on a small number of sure sellers — celebrity memoirs, established thriller writers, textbooks — and used those profits to finance a large midlist of books that were unlikely to be profitable individually but collectively maintained the publisher’s reputation as a serious editorial enterprise and occasionally produced the unexpected breakout. Publishers who abandoned the midlist in pursuit of pure commercial optimization consistently found their bestseller pipelines drying up within a decade: the commercial hits come from a talent pool that requires development and editorial investment, and that development is expensive and uncertain.

Amazon entered this system with a different theory of value. The traditional publishing system was organized around editorial judgment — humans deciding which books were worth making and using that judgment to signal quality to consumers. Amazon’s theory was that data and algorithm could replace editorial judgment, that reader reviews and purchase history and collaborative filtering could solve the information asymmetry problem better than human gatekeepers. Amazon also had a structural position that no previous book retailer had occupied: it was the dominant channel for book discovery and purchase, which gave it pricing power over publishers that the previous fragmented retail landscape had never concentrated in a single buyer.

Amazon squeezed publisher margins systematically and openly — the confrontations with publishers over ebook pricing in the 2010s were largely public. Amazon wanted 70% of ebook revenue. Publishers thought they were entitled to far more. Amazon had leverage because it controlled the dominant retail channel. The Hachette dispute of 2014, in which Amazon removed Hachette books from its recommendation algorithms and slowed delivery times during contract negotiations, was the clearest demonstration of this leverage: Amazon was willing to harm the reading experience of its own customers to win a margin negotiation.

Ebooks introduced an economic possibility that should, by standard economic logic, have dramatically reduced book prices: the marginal cost of producing one more copy of a digital book is approximately zero. No paper, no printing, no warehousing, no shipping. The minimum price at which a publisher can profitably sell an ebook should be very close to zero plus some allocation of editorial, marketing, and author costs. What actually happened is that publishers maintained ebook prices close to print prices, Amazon pushed for lower ebook prices than publishers wanted, and the primary beneficiary of the transition to digital was Amazon — which captured distribution margin that previously went to physical retailers and publishers.

The history of the book trade from Caxton to Amazon is the history of successive intermediaries solving the fundamental information asymmetry of experience goods — and extracting value commensurate with their structural position. The Stationers’ Company used legal monopoly. The 19th century publishing house used editorial reputation. Amazon uses algorithmic distribution control. The underlying problem — readers can’t evaluate books before buying them, so someone has to do the filtering and bear the risk of the filter failing — has not changed. What has changed is which institutional form captures the value of solving it. The answer is always whoever sits between the creator and the consumer, controls the channel through which discovery happens, and can extract rent from both sides of that transaction. The Stationers’ Company of 1557 and Amazon Prime in 2024 are not as different as they look.