KDP Metadata Kit

The problem

A book on Amazon is found through metadata, not quality. Readers search — they don't browse. If Amazon's index can't match your book to the words someone typed, your book doesn't exist to that reader.

Most KDP metadata advice online describes a system Amazon retired in 2023. Authors are still being told to email support for ten categories, and still stuffing seven keyword fields with variations of one term.

What it does Picks 3 browse categories, favouring deep winnable shelves over broad ones you can't rank in Fills all 7 backend keyword fields, each counted against the 50-character limit, each targeting a distinct search intent Enforces the repetition rule — no keyword wastes its slot on a word your title already indexes Writes an indexed title and subtitle, treating the subtitle as the discoverability asset it is Drafts a blurb structured by format, front-loaded for the "read more" fold Closes with pre-publish checks you can verify, not just assertions Built for 2026, not 2022

The reference file includes a section on advice that's now wrong, and covers what changed when Amazon layered semantic search over keyword matching — which shifts what actually works inside those seven fields.

Low-content books included

Journals, planners, notebooks, coloring and activity books get their own treatment. That buyer purchases an object, not a story: gifting queries dominate, specifications convert, and the cover carries the preview burden.

Who it's for

Self-publishing authors, low-content publishers, and anyone with a book already live on KDP that isn't being found.

What it does not do

It doesn't estimate demand. It can't tell you a niche's search volume, a competitor's sales, or whether a category is winnable in numbers — that needs live market data this skill has no access to. When you need those figures it says so and points you elsewhere, rather than inventing them. Amazon's rules also change: where the reference conflicts with your KDP dashboard, the dashboard wins.