Last reviewed by Robert Prime, July 2026
A healthy ad spend for a book is any spend that stays below your break-even advertising cost of sale (ACOS) once you count everything a reader is worth — the sale, plus page reads, plus the rest of the series. For a Kindle ebook on Amazon's 70% royalty tier, break-even ACOS lands around 67–68%; a sensible profit target for a standalone title sits nearer 20–35%. If you have a series with strong read-through, you can rationally spend far more on book one and still come out ahead.
Those numbers only mean something once you can calculate your own, so this guide shows the maths rather than handing you a magic figure. I've run advertising accounts across e-commerce for over 25 years and self-published my own book, and the authors who make ads pay are the ones who know their break-even cold before they raise a bid.
TL;DR — Key Takeaways
- ACOS = ad spend ÷ ad-attributed sales. ROAS = the inverse (1 ÷ ACOS). 25% ACOS is a 4× ROAS; 50% ACOS is 2×; 70% ACOS is about 1.43×.
- Break-even ACOS ≈ your royalty ÷ your price. On a 70% ebook that's roughly 67–68% (the royalty rate minus a small delivery fee). On the 35% tier it's about 35%. Paperbacks are lower, around 15–26%, because print costs eat the margin.
- Books are one of the cheapest, most efficient categories to advertise. Aggregated 2026 data (Ad Badger) puts the Books category at roughly a $0.38 average cost-per-click and 19% average ACOS, the lowest of any category.
- Standalone target: comfortably under break-even (≈20–35% ACOS). Series with read-through: you can tolerate 60–80%+ on book one, sometimes over 100%, because later royalties are unpaid-for.
- Console ACOS understates true profit for KU and series books because it ignores page reads and read-through. Use TACOS (spend ÷ total royalties) or blended ACOS instead.
- Kill a keyword after ~25–35 clicks with zero sales. Give a whole campaign at least 30 days before judging it.
ACOS and ROAS, Defined Once and Properly
Advertising Cost of Sale (ACOS) is the percentage of your ad-attributed revenue that you spent on ads. Spend £30 to make £100 in sales and your ACOS is 30%. Return on Ad Spend (ROAS) is simply the other way round, revenue divided by spend — so £100 from £30 is a ROAS of about 3.3×. Amazon's own definition confirms they're mutually inverse. Pick whichever you find more intuitive; just don't quote one thinking it's the other.
The quick conversions worth memorising:
| ACOS | ROAS |
|---|---|
| 20% | 5.0× |
| 25% | 4.0× |
| 33% | 3.0× |
| 50% | 2.0× |
| 70% | 1.43× |
| 100% | 1.0× (break-even on spend) |
Break-Even: The Only Number That Matters First
Break-even ACOS is the point where an ad-driven sale makes exactly zero profit, where your ad spend equals your royalty. The formula is refreshingly simple:
Break-even ACOS = royalty per sale ÷ list price.
To use it you need your real royalty, and KDP's ebook maths has a wrinkle: on the 70% tier, Amazon deducts a small per-megabyte delivery fee ($0.15/MB in the US, £0.10/MB in the UK) before paying you. So a £3.99 ebook with a roughly 1 MB file earns about 70% × (£3.99 − £0.10) ≈ £2.72, and its break-even ACOS is £2.72 ÷ £3.99 ≈ 68%. Work a few examples and a pattern appears:
| Book | Royalty per sale | Price | Break-even ACOS |
|---|---|---|---|
| $2.99 ebook, 70% tier (~1 MB) | ~$1.94–1.99 | $2.99 | ~67% |
| $4.99 ebook, 70% tier (~1 MB) | ~$3.34–3.39 | $4.99 | ~68% |
| £3.99 ebook, 70% tier (~1 MB) | ~£2.72 | £3.99 | ~68% |
| £0.99 ebook, 35% tier | ~£0.35 | £0.99 | ~35% |
| £9.99 paperback (after print cost) | ~£2.50–2.60 | £9.99 | ~25% |
The structural insight: on the 70% tier, break-even ACOS is always close to 70%, whatever the price, because the royalty is always about 70% of the price minus a few pence of delivery. Drop to the 35% tier (a 99p ebook, say) and break-even halves. Move to print and it falls further because printing costs are subtracted before your cut. If you want to check your own figures, our KDP royalty calculator does the arithmetic.
One UK note that keeps the maths clean: UK ebooks and print books are both zero-rated for VAT, so your royalty is calculated on the price the reader actually sees. In most EU markets, VAT is stripped out first, which lowers the effective royalty against the shelf price.
What "Good" Actually Looks Like for Books
Here's the encouraging part: books advertise well. Aggregated 2026 account data from Ad Badger puts the Books category at roughly a $0.38 average CPC and a 19% average ACOS, the lowest cost-per-click and lowest ACOS of any category they track, against an all-category average CPC of about $1.22. Clear purchase intent and low click prices make books forgiving to learn on.
That said, averages hide genre. Fiction, especially romance, tends to sit at the cheap end, with practitioner CPCs often in the $0.30–0.60 range. Competitive non-fiction and business books can run $1.50–3.00+ per click, because the audience is smaller and the advertisers deeper-pocketed. Your genre sets your baseline more than any benchmark does.
Don't chase someone else's ACOS number. Chase the gap between your ACOS and your own break-even.
For a standalone title, aim comfortably below break-even, a practitioner sweet spot of roughly 20–35% ACOS for an established book leaves real margin. During a campaign's learning phase, or at launch, a higher ACOS (even 70–100%) is normal and acceptable for a short window while you gather data and buy early rank.
Why Series and Kindle Unlimited Change the Rules
The advertising console shows you ad spend and ad-attributed sales. What it does not show is two things a reader is often worth far more than a single purchase: page reads and read-through.
If your book is in Kindle Unlimited, a borrow that gets read generates page-read royalties the console ignores. The KDP Select Global Fund per-page rate floated between roughly $0.0042 and $0.0049 across the first half of 2026 (it was about $0.004888 in May 2026). A 300-page book read in full earns around $1.30–1.45 in page reads on top of any sale — real money that makes a "losing" campaign quietly profitable.
If you write a series, a reader you acquire for book one may buy books two, three, and four with no further ad spend. That's read-through, and it transforms your maths. Suppose four books at £3.99 (about £2.72 royalty each) and 40% of book-one buyers read the whole series. The expected royalty from a single book-one buyer becomes roughly £2.72 + (0.4 × £2.72) + (0.4² × £2.72) + (0.4³ × £2.72) ≈ £4.54, more than the book's own price. Your true break-even ACOS on book one is then about £4.54 ÷ £3.99 ≈ 114%. In other words, you could spend more on advertising book one than it earns on its own and still profit across the series. (That 40% figure is illustrative, no reliable industry benchmark for read-through exists, so measure your own.)
This is why serious series authors deliberately run book one as a loss-leader. It's also why you should judge author ads with TACOS (ad spend ÷ total royalties) or a blended ACOS that adds page reads, not the console's headline number. Our Amazon Ads vs Meta Ads guide goes deeper on channel choice; the KDP Select vs wide decision affects how much of this page-read upside you can access.
Budgets, and When to Scale or Kill
Minimum viable budget. The practical floor for gathering usable data is around $10 per day (the range runs $5–15). Below that you won't get enough clicks to learn anything. Amazon's technical minimum is far lower, but a starved campaign just takes longer to teach you the same lesson.
How long before you judge it. Give a campaign at least 30 days, or one to two weeks of steady data at minimum, before drawing conclusions. Books have a longer consideration cycle than impulse products; early ACOS is almost always ugly.
When to kill. At the keyword level, the cleanest rule I've seen holds up: pause a keyword or target after about 25–35 clicks with zero orders. It's had a fair test and it's bleeding. Also cut targets whose CPC drifts well above your account average without converting.
When to scale. Raise budgets and bids on the keywords and campaigns sitting comfortably below your break-even — especially once you've factored in page reads and read-through. Scaling a profitable target is how ads compound; scaling before you know your true ACOS is how they hurt.
A note on 2026 platform shifts: Amazon has opened Sponsored Brands to single-book authors and now offers free AI image generation for author ad creative (it cites a 48% higher mobile click-through rate for Sponsored Brands ads that use a custom image over a plain product image, Amazon's own figure, so treat it as directional). Amazon CPCs have also crept up roughly 9% year on year, so last year's bids may under-deliver today.
Common Mistakes and How to Avoid Them
Judging KU or series books on console ACOS. You'll kill profitable campaigns because the console can't see page reads or later-book sales. Always reconcile against your total KDP royalties for the period.
Copying a headline ACOS target. "Keep ACOS under 30%" is meaningless without knowing the book's break-even. A 30% target is cautious for a 70% ebook and impossible for a 99p one. Calculate yours.
Starving the campaign, then blaming the platform. A £2/day budget on a competitive keyword gathers data so slowly you'll conclude "ads don't work" before they've had a chance to. Fund a proper test or don't run one.
Advertising a weak listing. No bid strategy rescues a book that doesn't convert. If your organic conversion is poor, fix the cover, blurb, and reviews first, see where to spend your budget for the order that works.
Frequently Asked Questions
What is a good ACOS for book ads?
It depends entirely on your royalty and whether you have read-through. For a standalone Kindle ebook on the 70% tier, break-even is around 67–68%, so a profitable target is comfortably below that — roughly 20–35% for an established title. Aggregated 2026 data puts the average ACOS for the Books category at about 19%, the lowest of any category. If you have a series or strong Kindle Unlimited read-through, a much higher book-one ACOS can still be profitable.
How do I calculate my break-even ACOS?
Divide your royalty per sale by your list price. For a 70% Kindle ebook, remember Amazon deducts a per-megabyte delivery fee first ($0.15/MB US, £0.10/MB UK), so a £3.99 ebook earns about £2.72 and breaks even at roughly 68% ACOS. On the 35% tier, break-even is about 35%. Our KDP royalty calculator will work out your exact royalty.
What's the difference between ACOS, TACOS and ROAS?
ACOS is ad spend divided by ad-attributed sales. ROAS is the inverse, sales divided by spend. TACOS (total advertising cost of sale) is ad spend divided by your total royalties, including organic sales and page reads. For authors with series or Kindle Unlimited, TACOS is the more honest measure because standard ACOS ignores the money ads generate that Amazon doesn't attribute to the click.
How much should I budget per day for Amazon book ads?
Around $10 per day is the practical minimum to gather usable data, though $5–15 is the working range. Below that you won't get enough clicks to make decisions. Give a campaign at least 30 days before judging it, and evaluate individual keywords on a "25–35 clicks with no sale, then pause" rule.
Can I run ads profitably with just one book?
It's harder. Without series read-through, every reader is worth only a single sale, so you must keep ACOS well under your break-even, which leaves little room. Kindle Unlimited page reads help. But in general, ads pay back best once you have two or more books; for a lone debut, organic channels and newsletter promotions usually return more per pound.
About the Author
Robert Prime is a self-published author, veteran e-commerce strategist, and the founder of publishing.co.uk. With over 25 years in digital business — including running the Amazon advertising agency MrPrime.com, he brings a practical, numbers-first perspective to self-publishing. After navigating the formatting and marketing of his own book, Google. Panic. Repeat., he built publishing.co.uk to help UK authors avoid the same pitfalls. He is co-owner of the LoveReading.co.uk network and a member of the Forbes Business Council.
