Teasing out a single line...."Describing books accurately is difficult and almost no-one can do it at all well."
The foibles of modern publishing! This is an insider perspective...All books (aside from self-printed artbooks and zines) exist as data and meta data before they exist as a book. The "description" serves many masters including an AI one...if we place The Algorithm by its correct category. A book description must not only be marketing copy that appeals to a potential reader, it must also contain keywords for the data scrapers to process.
Even a pre-internet published title must have data loaded into the distributors database for it to be sold. There are a few master databases (usually owned and maintained by the distributor) all publishers use that all retailers pull from. Once a title contains the requisite data and is turned live, the retailers pick up the data feed and the book becomes a purchasable object via Amazon, Walmart, and every bookstore that picks up that data feed.
Amazon uses its algorithm to analyze key words in the description and other fields to make categorization and determine where it should place, rank, and display the book on a page. The Amazon Algorithm is a notoriously obscure function, tightly guarded to theoretically prevent gaming. Mostly because Amazon has already gamified the Algorithm for monetization. I can pay Amazon to consider certain elements of my data as more weighted to boost the book's placement. It becomes quite expensive so many publishers use the data fields--mostly the description--to front load as much meta data to positively influence the algorithm.
Were there bad book descriptions prior to the age of algorithms? Of course! But today's book descriptions are bad in a very specific way. This includes older books now available for online purchase. This also explains the frustration folks experience of searching for a something, a "Canadian feminist sci fi" and the display will show Margaret Atwood books as well as a bunch of other books that may have those three phrases in the meta data. Conversely, if one searches "the handmaid's tale" the Algorithm will show the book and any other product that uses those words in the description and metadata. The Algorithm is tuned for "discovery". Good for casual shoppers but annoying if you're seeking a specific thing. As Amazon has expanded beyond books, this process accounts for the frustration of seeing vacuum cleaners and a bio of Herbert Hoover when searching "Hoover."
A last comment. The other fallout to how the Amazon Algorithm and the marginal, shady edges of the publishing industry mess with readers is the overwhelming amount of older titles (and now its happening to newer books) that are read via OCR without review, analyzed for keywords then listed on Amazon with odd descriptions. If a reader buys this "book", the rip-off company uses print-on-demand technology to print the files and send the "book." Disappointing the reader with a cheap folio of the text and gobbledygook.
Teasing out a single line...."Describing books accurately is difficult and almost no-one can do it at all well."
The foibles of modern publishing! This is an insider perspective...All books (aside from self-printed artbooks and zines) exist as data and meta data before they exist as a book. The "description" serves many masters including an AI one...if we place The Algorithm by its correct category. A book description must not only be marketing copy that appeals to a potential reader, it must also contain keywords for the data scrapers to process.
Even a pre-internet published title must have data loaded into the distributors database for it to be sold. There are a few master databases (usually owned and maintained by the distributor) all publishers use that all retailers pull from. Once a title contains the requisite data and is turned live, the retailers pick up the data feed and the book becomes a purchasable object via Amazon, Walmart, and every bookstore that picks up that data feed.
Amazon uses its algorithm to analyze key words in the description and other fields to make categorization and determine where it should place, rank, and display the book on a page. The Amazon Algorithm is a notoriously obscure function, tightly guarded to theoretically prevent gaming. Mostly because Amazon has already gamified the Algorithm for monetization. I can pay Amazon to consider certain elements of my data as more weighted to boost the book's placement. It becomes quite expensive so many publishers use the data fields--mostly the description--to front load as much meta data to positively influence the algorithm.
Were there bad book descriptions prior to the age of algorithms? Of course! But today's book descriptions are bad in a very specific way. This includes older books now available for online purchase. This also explains the frustration folks experience of searching for a something, a "Canadian feminist sci fi" and the display will show Margaret Atwood books as well as a bunch of other books that may have those three phrases in the meta data. Conversely, if one searches "the handmaid's tale" the Algorithm will show the book and any other product that uses those words in the description and metadata. The Algorithm is tuned for "discovery". Good for casual shoppers but annoying if you're seeking a specific thing. As Amazon has expanded beyond books, this process accounts for the frustration of seeing vacuum cleaners and a bio of Herbert Hoover when searching "Hoover."
A last comment. The other fallout to how the Amazon Algorithm and the marginal, shady edges of the publishing industry mess with readers is the overwhelming amount of older titles (and now its happening to newer books) that are read via OCR without review, analyzed for keywords then listed on Amazon with odd descriptions. If a reader buys this "book", the rip-off company uses print-on-demand technology to print the files and send the "book." Disappointing the reader with a cheap folio of the text and gobbledygook.
Happy reading, everyone. Caveat emptor!
This is super informative and finally explains why I hate the Amazon search function! Thanks