There are millions of books on Amazon and almost nobody browses. Readers type something into a search box, look at the first screen of results, and buy from it. Which means the question of whether a book sells is largely the question of whether it appears when someone types the thing it is about.
That is a solvable problem, and it is striking how few independent authors treat it as one.
Amazon is a shopping engine, not a library
The distinction matters. A library catalogue tries to describe books accurately. A shopping engine tries to predict which result will be bought, and it learns from behaviour. A book that appears in results and gets clicked and bought will keep appearing. One that appears and is ignored quietly stops appearing.
The practical consequence is that the first weeks are self reinforcing in both directions, and that the metadata you supply is the only lever you have before any behaviour exists to learn from.
The seven slots
When you publish, Amazon gives you seven fields for search terms. Most authors fill them with single words, which is the most common and most expensive mistake in self publishing.
Each slot holds a phrase, and shoppers search in phrases. "Historical fiction" is a slot spent on a category so vast that ranking in it is hopeless. "World war two nurse novel" is a slot spent on a query with real intent behind it and few books competing precisely for it.
Three rules cover most of the value.
Do not repeat your title. Your title and subtitle are already indexed. Repeating those words inside the slots spends them on coverage you have for free.
Write what a reader would type, not what a librarian would file. People search by feeling, by comparison and by situation. "Books like" a well known title. "Gift for a new manager." "Cozy mystery small town."
Prefer specific over large. Ranking eleventh for a huge phrase produces nothing, because nobody scrolls. Ranking second for a small one produces sales, and those sales feed the algorithm that then lets you compete for larger phrases later.
The full seven slot approach, including how to research phrases without paid tools, is laid out in this KDP keyword strategy if you want the detailed version.
Categories are the other half
Categories decide which bestseller lists you can enter. The same logic applies and it is the same instinct that has to be resisted. A narrow category with modest traffic where your book can genuinely reach the top of the first page is worth more than a broad one where it sits on page nine forever.
Reaching the top of a small category also produces a visible badge, and that badge changes click behaviour on every other search result the book appears in.
The part that is checkable
None of this has to be guesswork, which is why the neglect is puzzling. Type your phrases into Amazon's own search box and see what comes back. The autocomplete suggestions are real queries that real people type, and they are free.
Then search your own book's phrases and see where it appears. If it does not appear at all, the terms are wrong, and they can be edited at any time from your publishing dashboard.
Why this matters more than it used to
Advertising costs on the platform have risen steadily and organic search has not. For an author without a budget, metadata is the remaining free lever, and it is one of the few in publishing that responds to an afternoon of careful work.
A good book with careless metadata is invisible. A modest book with precise metadata gets found. That is not how anyone thinks the world should work, but it is how this particular shop is arranged, and the authors who accept it early do considerably better than the ones who discover it in year two.
Hello, My name is John Smith, I am a freelance writer and I would love to build up my writing portfolio online.