It started with a problem.
Before the ISBN, booksellers and libraries identified books by title, author, and publisher — information that varied in format across different catalogs and systems. A bookseller ordering from multiple publishers used different ordering codes for each one. Gordon Foster, a statistician at Trinity College Dublin, proposed a standard book numbering system in 1966 at the request of the Booksellers' Association of Great Britain and Ireland. The International Standard Book Number system was adopted internationally in 1970 as ISO Standard 2108. The original system used ten digits: a group identifier for language area, a publisher identifier, a title identifier, and a check digit.
Historical sources below ↓Then something changed.
The ISBN did two things simultaneously. For the supply chain, it created a universal ordering and cataloging identifier that worked across publishers, distributors, and retailers. For the barcode era, it provided the content for the EAN-13 barcode on book covers — allowing point-of-sale scanning to identify any book in any store. A bookshop could scan returned stock and know exactly what to reorder. A distributor could process thousands of different titles without custom codes per publisher. The ISBN became an example of how a simple identifier standard can reduce coordination costs across an entire industry.
What a builder can learn.
Our interpretation: the ISBN solved the problem of identity across organizational boundaries. Each publisher had its own internal codes; each bookseller had its own catalog. The ISBN meant neither had to translate when communicating with the other. Any business that operates within a network of suppliers, distributors, and retailers faces the same translation problem. A shared identifier reduces the friction at every point of exchange, and the cost of creating one is usually lower than the cumulative translation cost it replaces.
A door still open.
A hypothesis worth testing: if a business mapped the number of times it translates a product or customer identifier when exchanging information with a partner — from invoice to delivery note to return record — and proposed a single shared identifier for one exchange, would that eliminate a measurable category of errors or delays?
How many times does the same order, product, or customer get re-identified by a different code before a transaction is complete?
Follow the evidence.
History is sourced below. Business lessons are editorial interpretations; future possibilities are questions to test, not predictions.
Prepared from primary-source histories. First edition: 25 Sept 2026.