Most authors who try Amazon Ads and give up are not failing because of bad books or bad luck. They are failing because they went in without understanding the system — what it rewards, what it penalises, and what it requires before it starts working. Amazon Ads is not a tap you turn on and leave running. It is a compounding system: campaigns improve as they gather data, keywords prove themselves over weeks not days, and the authors who stay consistent through the early period of modest returns are the ones who eventually run profitable campaigns almost on autopilot.
This guide covers the system in full. For deeper treatment of individual topics, each section links to its dedicated article within this series.
How Amazon Ads Actually Work
Amazon Ads operates on a pay-per-click auction model. When a reader searches for something on Amazon — “cosy mystery British village” or “personal finance for beginners” — Amazon runs an instantaneous auction to determine which ads to show alongside the organic results. Every advertiser targeting that search competes. Amazon does not simply reward the highest bidder. It multiplies each bid by a quality score derived from the ad’s predicted click-through rate and conversion probability. A well-converting book with a professional cover and strong reviews can beat a higher bid from a poorly converting competitor.
You pay only when a reader clicks your ad — not when it is displayed. The amount you pay per click (CPC) is determined by the second-price auction mechanism: you typically pay just enough to beat the next-highest bidder, not your full maximum bid. In practice, book advertising CPCs in most categories run between £0.15 and £0.60, which is significantly lower than the Amazon-wide average. This makes book advertising more accessible than most authors assume — the barrier is not the cost per click but the conversion rate of the page the reader lands on.
When the reader clicks, they land on your product page. Whether they buy depends entirely on your cover, description, pricing, and reviews. The ad’s job is to deliver the reader to your page. The page’s job is to convert them. These are two separate problems, and no amount of ad spend fixes a conversion problem that lives on the product page.
What You Need Before You Spend Anything
There is a precise checklist of conditions your book should meet before advertising is likely to be profitable. Running ads without these in place is one of the most reliable ways to waste money on Amazon.
A professional cover. The cover is the first thing a reader sees in search results. If it does not immediately signal genre, quality, and professionalism, your click-through rate will be poor and your conversion rate lower still. An ad cannot compensate for a cover that fails to communicate what the book is. If your cover was made in Canva without genre design expertise, the best use of your current budget is a professional redesign — not ad spend.
A compelling, correctly formatted description. Your description has one job: convert a curious click into a purchase decision. Most author-written descriptions fail because they describe the book from the author’s perspective rather than generating the desire to read from the reader’s perspective. Read your description critically against the competition in your category. For the full description guide, see our complete guide to writing Amazon book descriptions.
At least 5–10 genuine reviews. Reviews are social proof, and Amazon’s algorithm uses review count and rating as quality signals affecting both organic ranking and ad relevance scores. A book with zero reviews will convert poorly on ad traffic — readers arriving via ads are often cold audiences who rely on reviews as a trust signal. Build your review base first through advance reader copies, newsletter subscribers, and review request programmes.
Competitive pricing. If the average comparable book in your category is priced at £3.99 and yours is at £6.99 without a clear reason for the premium, your conversion rate will be structurally poor regardless of how well your campaigns are built. Check the pricing landscape in your category before launching.
Optimised metadata. Your book’s backend keywords and categories directly feed Amazon’s automatic targeting algorithm. Vague or inaccurate metadata causes automatic campaigns to match your ad to irrelevant searches and drain budget. Get your metadata right first. See our KDP keyword research guide and category selection guide for the full process.
The Three Ad Types Available to KDP Authors
KDP authors have access to three advertising formats, each serving a distinct purpose. Most authors should start with one — Sponsored Products — and add the others as their catalogue and budget grow.
Sponsored Products are the core format: individual book ads that appear in Amazon search results and on other books’ product pages. Available to every KDP author regardless of catalogue size. They drive direct-response traffic — the reader clicks and lands immediately on your product page. The vast majority of your ad budget and attention should go here initially.
Sponsored Brands are banner-style ads displaying your author logo, a custom headline, and three or more books simultaneously. They require a minimum of three eligible titles and an Author Central page. Brand-awareness ads that work alongside Sponsored Products rather than replacing them. As of March 2026, Sponsored Brands include AI-generated imagery (free, via Amazon’s built-in image tool) and AI-powered prompts that answer shopper questions directly within the ad experience.
Sponsored Display extends your reach beyond Amazon search: it retargets readers who viewed your book but did not buy, reaches them through off-Amazon placements including apps, news sites, and Kindle devices, and uses lookalike audiences to find new readers who profile-match your existing buyers. For most new advertisers, Display is a later addition once Sponsored Products campaigns are working consistently.
Sponsored Products: The Foundation
Sponsored Products appear in three placement positions. Top of search placements sit above organic results — highest visibility, highest CPCs, highest click-through rates. Rest of search placements appear interspersed throughout results below the fold — lower CPCs, meaningful volume. Product page placements appear in the sponsored section on other books’ detail pages — often the lowest CPC and, for well-matched targeting, competitive conversion rates because the reader is already in buying mode for a comparable book.
All placements carry a small “Sponsored” label, but this rarely prevents clicks. The cover image is the dominant visual signal — a professional cover in the right genre competes effectively regardless of the label.
There is no upfront charge to set up Sponsored Products. You set a daily budget cap and per-keyword bids, pay only on clicks, and can pause or end campaigns at any time. Amazon moderates every campaign before it goes live — this typically takes 24 hours but can run to 3 business days. Build this into your launch timeline.
Automatic vs Manual Targeting
This is the most consequential structural decision in book advertising, and the answer is not one or the other — it is both, running simultaneously, feeding each other.
Automatic targeting campaigns let Amazon’s algorithm match your ad to relevant searches and product pages based on your book’s metadata. You do not specify any keywords. Amazon reads your title, subtitle, description, categories, and backend keywords and decides when your ad is relevant. The four automatic sub-types — Close Match, Loose Match, Substitutes, and Complements — can be bid on independently, letting you concentrate budget on the tightest-relevance matches while still gathering broader data.
The great value of automatic targeting is discovery: it surfaces search terms you would never have targeted manually — genre slang, comparative author searches, trope phrases. The cost is efficiency: it will always spend some budget on irrelevant searches until you have built a strong negative keyword list. This is not a flaw to avoid but a mechanism to manage actively through weekly Search Term Report review.
Manual targeting campaigns require you to specify exactly what to bid on — specific keywords with match types, or specific products by ASIN. Manual campaigns are more efficient than automatic for terms you have already confirmed convert, and entirely dependent on having the right list to begin with. The primary source of that list is your automatic campaign’s data. The two campaign types are a system, not alternatives.
The workflow connecting them is called harvest-and-scale: run automatic for 14+ days, pull the Search Term Report, find terms that generated sales at acceptable cost, add those to manual exact match, add them as negatives in auto to prevent the two campaigns competing on the same terms. Repeat weekly. This is the core engine of sustainable book advertising. See our dedicated automatic vs manual targeting guide for the full process.
The Campaign Structure That Works
For a new book, launch three campaigns running simultaneously from day one. Separating targeting types into distinct campaigns keeps data clean and budget allocation intentional — mixing types in one campaign is one of the most common structural errors in KDP advertising.
Campaign 1 — Automatic discovery. Targeting: automatic. Bid strategy: dynamic down only (Amazon can reduce bids for low-probability clicks but cannot exceed your set amount). Daily budget: £5–£10. Purpose: discover what readers actually search for when they find books like yours.
Campaign 2 — Manual keyword targeting. Seed with your primary genre terms, trope keywords, and comparable author searches using exact and phrase match types. Daily budget: £5–£10. As the automatic campaign matures, proven converters from the Search Term Report populate this campaign’s exact match list. Purpose: convert known high-intent searches efficiently.
Campaign 3 — Manual product targeting. Target the ASINs of the 20–30 most directly comparable books in your category. These readers are already interested in a book like yours — the context cannot be more targeted. Daily budget: £3–£8. Purpose: capture browser-mode readers at typically lower CPC than keyword advertising achieves.
Keywords: The Engine of Your Campaigns
Keywords are the search terms that trigger your ads. Getting them right — specific, relevant, and drawn from real reader behaviour — is the difference between targeted, profitable campaigns and expensive, diffuse ones.
Your Search Term Report is the most valuable keyword source available once your automatic campaign has run for 14+ days. These are not guesses about what readers might search — they are actual queries from readers who clicked your ad. Sort by orders descending: every converting term at acceptable cost is an exact match keyword candidate for your manual campaign. This is your primary harvest mechanism and the activity that makes book advertising compounds over time.
Competitor ASIN research via the Book Keyword Spy tool at KDP Rank Fuel by Vappingo shows every keyword that comparable books rank for. If a direct competitor consistently ranks for “small town cosy mystery female sleuth,” that term has proven demand from readers similar to yours — start targeting it immediately.
Genre and trope vocabulary matters increasingly in 2026. Amazon’s Rufus AI discovery assistant uses semantic matching — it reads your full listing and reviews to assess contextual fit, not just keyword strings. Fiction readers search extensively by trope: “enemies to lovers romance,” “forced proximity,” “dark academia,” “cosy fantasy,” “grumpy sunshine.” These trope terms often have lower CPCs than broad genre terms and higher conversion intent.
Amazon autocomplete reveals real search behaviour patterns. Type your primary genre term into Amazon’s search bar and capture every autocomplete suggestion. Each represents an actual pattern from Amazon’s user data. “Cosy mystery” autocompletes to “cosy mystery books,” “cosy mystery series,” “cosy mystery British,” “cosy mystery with recipes” — each suggestion is a distinct keyword with a built-in demand signal.
Bidding and Budgets
Three bidding strategies are available. Fixed bids use exactly the amount you set regardless of context. Dynamic bids — down only allows Amazon to reduce your bid when a click seems less likely to convert, but never exceeds your maximum — the recommended starting strategy. Dynamic bids — up and down allows Amazon to reduce or increase your bid by up to 100% — only appropriate for mature campaigns with extensive conversion history.
Calculate your breakeven bid before setting anything: (royalty per sale) × (expected conversion rate) = maximum profitable bid. If your ebook royalty is £2.00 and you expect 4% of clicks to convert, your breakeven bid is £0.08. Most book categories have CPCs of £0.15–£0.50, meaning you need 3–5% conversion for profitability. Strong covers, descriptions, and reviews drive conversion — invest there first if your conversion rate is low.
For daily budgets, start with enough to run at least 30 days without interruption. £5–£10 per day per campaign is a practical minimum for most book categories. If a campaign consistently hits its daily cap before the day ends and its ACoS is below your target, increase the budget — you are cutting off profitable traffic artificially.
Understanding ACoS
ACoS (Advertising Cost of Sales) = Ad Spend ÷ Ad Revenue × 100. If you spent £25 on ads that drove £100 in book sales, your ACoS is 25%. Your breakeven ACoS is your royalty expressed as a percentage of list price: a Kindle ebook at £2.99 on the 70% royalty tier has a breakeven ACoS of approximately 70%. A paperback at £9.99 with a £2.60 net royalty (after printing costs) has a breakeven ACoS of 26%. Know your breakeven before you set a single bid — it is different for every book and format.
Your target ACoS should sit below breakeven for an established book where you want direct profitability. During a launch window where you are prioritising sales velocity and organic ranking, running at or above breakeven ACoS for a defined 4–8 week period is strategically sound — you are buying position, not just sales. Define that window explicitly, set an organic ranking milestone you expect to hit by the end of it, and make a data-based decision at that point rather than running indefinitely at a loss hoping it will eventually turn.
What ACoS does not capture: any organic sales generated by the improved BSR and keyword ranking your ads produce. A campaign with 40% ACoS that is simultaneously moving your book to page one of “cosy mystery British” organic results is doing far more than the ACoS figure alone reflects. This is why TACoS is the more meaningful long-term metric.
Why TACoS Matters More Long-Term
TACoS (Total Advertising Cost of Sales) = Ad Spend ÷ Total Revenue × 100. Unlike ACoS — which counts only revenue attributed to ad clicks — TACoS uses all revenue: both ad-attributed and organic. This makes it the true measure of whether advertising is building a sustainable business or creating an expensive dependency.
The pattern you want to see: as a book matures through consistent advertising, TACoS should fall even while ad spend stays constant. Why? Because effective advertising improves your organic keyword rankings, which generates organic sales that lift total revenue without proportionally increasing spend. When TACoS is falling over time, your ads are compounding: buying direct sales and building organic position simultaneously. When TACoS is rising with flat or declining total revenue, you are propping up sales that would collapse without ads — a warning signal.
Practical benchmarks: a book in launch phase should expect TACoS of 20–35%. An established title with healthy organic traction should aim for 8–18%. TACoS above 25% on a title that has been advertising for more than three months suggests either a product page conversion problem or a keyword targeting problem — both of which require diagnosis, not just bid adjustments.
The Weekly Optimisation Routine
Consistent weekly optimisation is what separates compounding campaigns from stagnant ones. Amazon’s algorithm needs 7–14 days minimum to gather meaningful data — making changes faster produces decisions based on noise.
Search Term Report review (every 14 days). Download the 14-day Search Term Report. Sort by spend descending. For every term with spend exceeding 1.5× your royalty and zero sales: add as negative keyword. For every term with 2+ sales at or below target ACoS: add as exact match to manual campaign, add as negative exact to auto campaign.
Bid adjustments (weekly, for keywords with 15+ clicks). Keywords with ACoS more than 15 percentage points above target: reduce bid by 20%. Keywords with ACoS more than 15 percentage points below target: increase bid by 15–20% to capture more impression share. Keywords with fewer than 10 clicks: make no changes — the data is insufficient for a sound decision.
Budget review (weekly). Campaigns consistently hitting their daily cap with ACoS below target: increase budget by 30%. Campaigns consistently underspending with above-target ACoS: reduce budgets only if keyword cull has failed — the problem is usually keyword quality, not budget level.
Monthly structural review. Every 60–90 days: review whether campaigns generating poor ACoS are worth rebuilding from scratch rather than continuing to optimise. Check whether your product page metrics (conversion rate visible via the Advertising console’s product performance section) indicate a listing quality issue rather than a targeting issue.
Sponsored Brands: Building Your Author Brand
Once you have three or more eligible titles, Sponsored Brands serve a distinct purpose alongside Sponsored Products. Where Sponsored Products promote a single book at the point of purchase decision, Sponsored Brands promote you as an author — showing your logo, a custom headline, and three books simultaneously in a banner format at the top of search results.
A reader who sees a Sponsored Products ad for one book and a Sponsored Brands banner showing three books from the same author in the same session gets a very different impression than either ad alone. The catalogue display signals “this author has multiple books” — social proof that significantly increases the likelihood of a follow on your author page and return visits for future releases.
Since March 2026, Sponsored Brands include free AI-generated custom imagery via Amazon’s built-in tool, and AI-powered prompts that answer reader questions directly in the ad. The AI prompt capability means your description quality and review language now feed into what appears in your ad — shopper questions like “is this suitable for someone who doesn’t usually read thrillers?” draw answers from your product page content. Write your description with this in mind. Allocate 10–15% of total ad budget to Sponsored Brands once Sponsored Products are running efficiently.
Sponsored Display: Retargeting and Off-Amazon Reach
Sponsored Display serves two functions. The retargeting function shows your ad to readers who viewed your product page without buying — re-engaging them across Amazon pages, apps, news sites, and Kindle devices. Amazon data indicates that the majority of product page views do not result in immediate purchase; retargeting captures a meaningful proportion of these hesitant readers at lower CPCs than search-based advertising because you are not competing in the main search auctions.
The lookalike audience function finds new readers whose shopping behaviour resembles your existing buyers — essentially extending your proven audience pool. This is a later-stage strategy once you have enough buyer history for the lookalike model to work with.
For most new advertisers, Sponsored Display is a third phase addition. If your book is receiving fewer than 50 product page views per day, the retargeting pool is too small to generate meaningful volume. Build organic and Sponsored Products traffic to meaningful levels first, then layer in Display retargeting.
Goodreads Placements
Sponsored Products and Display ads automatically include Goodreads as part of Amazon’s advertising network. You cannot directly target or exclude Goodreads from the console — the platform handles this placement automatically. Goodreads audiences are warm (actively engaged readers tracking books and following authors), making automatic placements there potentially valuable. Because you cannot isolate Goodreads performance from your overall placement data directly, treat any visibility there as a network bonus. If your overall placement-level performance suggests poor conversion from certain placements, reducing overall bid levels will deprioritise lower-quality placements across the network.
AI in Amazon Ads in 2026
Amazon’s advertising platform has integrated AI at multiple levels that meaningfully affect book advertising. The Rufus AI shopping assistant — Amazon’s conversational discovery tool — now influences book discovery through semantic matching. When a reader asks Rufus “what should I read if I loved the cosy vibes of Richard Osman?” it does not simply match keyword strings. It reads your full metadata, reviews, and customer signals to assess contextual fit. Your description quality, review language, and category accuracy all feed into ad relevance in ways that did not exist before Rufus’s integration with advertising.
Within the Amazon Ads console, AI-suggested ad copy and headlines are available for Sponsored Brands. AI-powered prompts that answer shopper questions directly within ad placements went into general availability on 25 March 2026. These prompts use Amazon’s shopping data to generate answers to questions like “how long is this series?” or “is this appropriate for someone who doesn’t usually read fantasy?” from your product page content. The practical implication: descriptions that clearly answer common reader questions now perform better in AI-mediated advertising than descriptions that are vague or generic.
Use Amazon’s AI ad copy suggestions as starting points only. They do not know your book’s specific voice, its distinguishing details, or the precise reader it is trying to reach. Customise heavily.
A Realistic Timeline for New Advertisers
Unrealistic expectations are the leading driver of premature campaign abandonment. Most profitable book advertisers needed 60–90 days before their campaigns generated consistent returns.
Weeks 1–2: Learning phase. ACoS will be above target — often significantly above. The algorithm is calibrating. Make no bid changes. Gather data.
Weeks 3–4: First meaningful Search Term Report harvest. Begin the harvest-and-scale cycle. First bid adjustments on keywords with 15+ clicks. ACoS should begin to improve.
Months 2–3: Campaigns approaching or at profitability on direct ACoS metrics. Manual exact match campaign growing with proven converters. TACoS beginning to show the organic traction trajectory — or not, in which case a product page audit is warranted.
Month 3+: Consistent returns with 2–3 hours of optimisation per month. Organic ranking improvements reducing dependence on paid positions. The compounding flywheel becomes visible in the data.
The Most Common Mistakes
Advertising before the book is ready. No campaign structure compensates for a poor cover, weak description, or zero reviews. The sequence matters.
Running only automatic campaigns indefinitely. Auto is a data source. Without the manual campaign harvest cycle, campaigns never reach their efficiency potential.
Making changes too frequently. Daily adjustments based on 3–4 days of data produce decisions from noise. The 7-day attribution window means data from recent days is incomplete. Optimise weekly on 14-day windows.
Never adding negative keywords. A three-month-old automatic campaign without a single negative keyword is almost certainly wasting 30–50% of its budget on irrelevant searches. Negative keyword management is as important as bid management.
Pausing campaigns at the first sign of high ACoS. The learning phase produces above-target ACoS by definition. The most common failure pattern is a premature pause on day five based on incomplete data.
Optimising ACoS without watching TACoS. Cutting a campaign that appears expensive on ACoS but is building organic ranking destroys long-term value while appearing efficient in the short term.
The rest of this series covers every topic in depth. Start with Sponsored Products in full, then work through targeting, bidding, and metrics. Your ads bring the reader to the page — professional manuscript proofreading from Vappingo ensures the book they find earns the reviews that make every future campaign cheaper to run.
