Non-fiction readers search for a problem they want solved, so the strongest keywords describe that problem in the reader's words. Series read-through matters less than in fiction; instead, judge ads by the book's own royalty, the paperback share, and any back-end value such as courses or services that you can actually measure.
Last reviewed by the InteliAds team.
Keywords that match a problem
Fiction readers browse; non-fiction readers usually arrive with a goal. "How to pay off debt", "anxiety workbook for teens", "sourdough for beginners" or "leadership book for new managers" each say exactly what the buyer wants. Your book's subtitle often already contains your strongest keywords.
List the problem, the outcome, and the audience, then combine them. Phrase match works well here because readers add small words around the core phrase.
Audience: "for nurses", "for beginners", "for small business owners".
Competing books: product targets on titles that solve the same problem.
Broad topics spend, specific how-tos sell
Single-word topics like "marketing" or "health" attract browsers and students, and other books with very different angles. They spend a lot and convert little. Specific phrases about a task or a reader type usually convert better because the reader can see your book is for them.
Check search terms often. Non-fiction auto campaigns can match on loosely related topics, for example a personal finance book showing for business accounting searches. Negate those early.
Watch for students and exam searches too. A book for practising nurses can match "nursing textbook" searches from students who need a specific course text. Those clicks rarely buy a general guide, so add them as negatives once you see them.
Format, price and KU
Non-fiction readers often prefer print, for workbooks, reference and books they want to mark up. If most of your sales are paperback, set your target from the paperback royalty after printing cost, which is usually a lower break-even ACOS than the ebook's.
KU still matters if you are enrolled, but for many non-fiction titles it is a smaller share than in genre fiction. Check your own KDP report: if page reads are small, your ACOS is a fair guide on its own.
Many non-fiction authors sell courses, coaching or services, and some run ads at a loss on the book to win clients. That can be sensible, but only with evidence. Track sign-ups that come from links or offers inside the book, over a few months, before raising your target ACOS to account for them.
Without that evidence, treat the book as a product that should pay for its own ads. It is easy to justify any ACOS with imagined back-end income.
Where InteliAds helps
InteliAds shows profit per book and per format, so a $24.99 workbook and a $6.99 ebook are not held to the same target. Search terms can be added as keywords or negatives directly from the search term view.
BidBot is the automated bidding in InteliAds. You pick a preset (Suggest only, Safe, Balanced, Aggressive or Aggressive+) and a target ACOS between 10% and 80%. Safe waits for 20 clicks before it raises a bid and leaves 48 hours between changes to the same target; Balanced waits for 10 clicks and 24 hours. You can also set it to lower bids only, and mark promo windows when it should not raise bids.
For non-fiction, a typical setup is a target ACOS just below the paperback break-even, BidBot in Safe or Balanced, and a Rule that adds search terms with 12 clicks and no orders in 30 days as negatives, previewed with Dry Run.
Write down the problem, outcome and audience your book serves.
Build phrase keywords from combinations of those, plus your subtitle words.
Product-target books that solve the same problem.
Calculate break-even ACOS per format; use the format that sells most to set the target.
Negate loosely related topics in auto search terms every week.
Only raise the target ACOS for back-end income you can measure.
Worked example (illustrative numbers)
Invented numbers. A budgeting workbook sells for $16.99 in paperback with a royalty of $5.10 after printing: break-even ACOS about 30%. The ebook at $6.99 has a royalty of about $4.80, break-even about 69%.
Seventy percent of sales are paperback, so the author blends the two: a target ACOS of about 40%. Search terms show "personal finance" spent $60 for 2 sales (ACOS above 100%), while "budget planner for couples" spent $20 for 5 sales. The author negates the broad term in exact match and adds the specific one as an exact keyword with a slightly higher bid.
Frequently asked questions
What keywords work for non-fiction Amazon Ads?
Phrases that describe the reader's problem, desired outcome and who they are. Broad topic words usually spend without converting.
Should I use the ebook or paperback royalty for my target ACOS?
Use the format that sells most, or a blend weighted by your sales mix.
Can I run book ads at a loss to get clients?
Only if you can trace clients back to the book. Otherwise, keep ACOS below the book's break-even.
Do series matter for non-fiction ads?
Less than in fiction. Judge most non-fiction books on their own royalty.
Amazon and KDP rules change; check the linked help pages before making decisions. Worked examples use illustrative numbers, not results from real accounts.