Amazon · 18 August 2026 · 7 min read

Metadata is the quietest lever on Amazon

Keywords, categories, subtitle, series field, A+ content and backend terms are the parts of a listing readers never see, and the parts the retailer reads first.

Amazon does not read a book. It reads a record about a book, and then decides which searches that record deserves to appear in. Metadata optimisation is simply the work of making that record accurate and competitive.

Most independent listings are not badly written. They are under-described: two vague keyword slots, a subtitle that repeats the title, an empty series field, and no A+ module telling the algorithm anything it does not already know.

The fields that actually move

Treat the seven backend keyword slots as seven distinct search intents, not one phrase chopped into pieces.

  • Seven backend slots, seven non-overlapping reader intents, no repeated words
  • Subtitle carrying genre signal a stranger can parse in one second
  • Series field completed and consistent across every title in the run
  • Two categories chosen for density, two more requested through support
  • A+ content written with searchable language, not decorative copy

SEO beyond the retailer

A book page also competes on the open web. A consistent title-plus-author string across Goodreads, the author site, BookBub and retailer listings is what lets a search engine consolidate a book into one confident result rather than five weak ones.

Review the record quarterly. Search language shifts, categories are restructured, and a keyword that carried a launch is rarely the keyword carrying month nine.

The Rittenhouse Reading List