Amazon ads are the most accessible paid channel for a self-published book and the easiest place to lose a hundred pounds learning nothing. The difference is almost entirely in whether you did the arithmetic first.
Do the arithmetic before you spend
One calculation decides whether this is viable at all.
Your royalty per sale ÷ your cost per click = the conversion rate you need to break even.
Worked through: a $4.99 ebook at the 70% band earns roughly $3.40 after delivery. At $0.35 a click, you need a sale every ten clicks (a 10% conversion rate). That is achievable but demanding.
| Royalty | Cost per click | Break-even conversion |
|---|---|---|
| $2.00 | $0.30 | 15% (very hard) |
| $3.40 | $0.35 | 10% (demanding) |
| $6.90 | $0.40 | 6% (realistic) |
| $10.00 | $0.50 | 5% (comfortable) |
The table says something uncomfortable and true: cheap books are hard to advertise. At the bottom of the price range the maths barely works before you have sold anything. How to price an ebook matters here more than any targeting decision.
Two things sit outside the calculation and both argue for patience: read-through to later books in a series, and Kindle Unlimited page reads. If you have either, your real break-even is better than the table suggests, but do not assume it before you have measured it.
Ads amplify, they do not fix
This is the part that costs people money. Traffic multiplies whatever your product page already does.
Before spending anything, the page has to convert:
- The cover. Judged at thumbnail size against nine others: ebook cover dimensions, and genre convention is information rather than laziness.
- The description. A sales page, not a summary (see writing a book description).
- The price, in the band its category expects.
- Some reviews. A book with none converts poorly however good it is: how to get book reviews.
- The look inside. People open it. If the interior is rough, they leave.
A book failing on any of these will fail faster with ads. That is the honest sequence: fix the page, then buy traffic for it.
The campaign to start with
Ignore most of the interface. Start here:
- Sponsored Products, not Sponsored Brands or Lockscreen. It is the cheapest to learn on and the closest to intent.
- Manual targeting, keywords. Automatic is useful later for discovering terms, and it spends unpredictably while you are learning.
- A small daily budget. Enough to gather data, not enough to hurt.
- Low starting bids, raised slowly. Amazon will always suggest more.
- Thirty to fifty keywords, in three groups: comparable book titles, comparable author names, and the phrases a reader would actually type.
That third group is where most people go wrong. Broad category terms have enormous competition and vague intent; specific phrases convert.
Reading the numbers
Give it two weeks before judging anything. Daily numbers on a small budget are noise.
| What you see | What it means | Do |
|---|---|---|
| Impressions, no clicks | Cover or title is not landing | Fix the cover, not the bid |
| Clicks, no sales | The page does not convert | Fix description, price, reviews |
| Sales, above break-even cost | Working | Raise the budget slowly |
| Almost no impressions | Bid too low, or the keywords are dead | Raise bids, add keywords |
The metric to watch is ACOS, advertising spend as a percentage of the sales it produced. Below your royalty percentage, you are making money. Above it, you are buying rank and you should know that is what you are doing.
Cut keywords that spend without selling. Most of the work after week two is subtraction.
When not to bother
- A single low-priced book with no others. The arithmetic rarely closes.
- Before the page converts organically. Fix that first.
- With no budget to lose. The first weeks are tuition.
- A free lead magnet. Different job, wrong tool: lead magnet PDFs are distributed, not advertised.
Ads are best for a book that already sells slowly, in a series, at a price that leaves room. For a first single title they are usually not the highest-value thing you could be doing with the same hours.
The cheaper thing first
Before paid traffic, the free levers are usually unexploited: categories and keywords on the listing itself, which decide what Amazon shows you for at no cost at all (book metadata and categories), and the launch sequence that gets early momentum, in the book launch checklist.
And the structural one: the margin you are advertising against is set by how you sell. A direct sale keeps nearly all of the price instead of 70% of it, which changes every row in the first table (where to sell your ebook). Producing both files from one source makes selling in several places cheap enough to bother with: bukpress exports the typeset PDF and the reflowable EPUB from the same Markdown manuscript, so a direct storefront is one more upload rather than another formatting job.
Frequently asked questions
How much should I spend on Amazon ads for a book? Enough to gather data and no more while learning: a small daily budget for two weeks, then decide on evidence. The number matters less than having calculated your break-even first.
What is a good ACOS for books? Below your royalty percentage means profit on that sale. Higher can still be rational if you are buying rank or have read-through to later books, provided you know that is the trade.
Why am I getting clicks but no sales? The ad is working and the product page is not. Cover, description, price or the absence of reviews, in roughly that order of likelihood.
Should I use automatic or manual targeting? Manual to start, because you learn what converts. Automatic later, as a way of discovering keywords you would not have thought of.
Do Amazon ads work for non-fiction? Often better than for fiction, because the search intent is more specific and the prices are higher. Both of those improve the break-even.