Amazon does not rank books on what they are about. It ranks them on what shoppers type. Those are two different vocabularies, and the gap between them is where most self-published titles disappear — a brilliant novel filed under words no buyer ever enters.
What you receive
- Seven backend keyword slots filled with researched phrases, within Amazon’s 50-character limit per slot and without repeating words across slots.
- Evidence for each phrase: how often it is searched, how many titles compete for it, and whether the top results look like your book.
- A title and subtitle check — the strongest ranking words belong there, not buried in the backend.
- Category alignment test so your book appears in a chart small enough to reach.
- A one-page sheet you can paste into KDP, plus a record of what was changed and when.
How the research works
- Reader vocabulary. We collect the phrases used in reviews and in competing blurbs — the language buyers already use for this kind of book.
- Competition check. Each phrase is tested against the current results. If the first page is all bestsellers, the phrase is unattainable and we move on.
- Placement. Words are split between title, subtitle, description and the seven backend slots so nothing is duplicated.
- Measurement. Impressions and keyword-driven sales are compared after a fortnight, and the weak slots are replaced.
What makes a keyword worth using
A good phrase has three properties: real search volume, results your cover can compete with, and intent to buy rather than to browse. “Christmas romance” is enormous and unwinnable; “small town Christmas romance second chance” is smaller and reachable. The work is finding the phrases in the middle, where a new title can appear on the first page of results and a shopper who clicks is already looking for exactly that book.
What not to do
Stuffing the backend with other authors’ names, competitor titles, or claims the book cannot support is against KDP’s terms and can get a listing suppressed. Repeated words in every slot waste capacity. Irrelevant bestseller phrases buy impressions from shoppers who will never click, which lowers the click-through rate and damages the listing’s standing over time.
Measurement is the part everyone skips
Keywords are a hypothesis. Two weeks after a change we compare impressions, clicks and keyword-driven sales against the previous period, then retire the slots that produced nothing and promote the phrases that converted. Without that loop, optimisation is a one-off guess that slowly goes stale as the market’s vocabulary shifts. We also keep a dated record of every change, because a listing edited four times with no notes becomes impossible to diagnose, and a book that suddenly loses visibility usually has a metadata edit behind it rather than a mystery.
Pricing
| Tier | Scope | Price | Was | You save |
|---|---|---|---|---|
| Starter | 1 ebook | $20 | $30 | |
| Professional | 3 ebooks | $60 | $75 | |
| Bulk / Authority | 5 ebooks | $100 | $100 |
Frequent questions
How long until results appear?
KDP reviews metadata changes within 24–72 hours; ranking movement usually shows inside two weeks, though the effect compounds over a month.
Can you do this for a book already live?
Yes, and that is the common case. Existing listings often recover quickly because they already have sales history behind them.
Do you handle translations?
Yes — Amazon’s marketplaces differ, so the phrase list is rebuilt per language rather than translated word for word.
Get your keyword sheet — send the title, genre and a link to the listing.