The most consequential structural decision in Amazon Ads for authors — how automatic and manual targeting work, why the answer is always both, and exactly how to connect them into a system that compounds in value every week it runs.
| 13-minute read | Beginner · Intermediate |
When authors first encounter automatic and manual targeting, the instinct is to treat them as alternatives — to pick one and run with it. This is the wrong frame. Automatic and manual targeting are not competitors; they are a system with distinct roles that only produces its best results when both are running simultaneously and connected through a weekly optimisation cycle. Understanding this changes how campaigns are structured, what you look for in the data, and how quickly your advertising becomes efficient.
The Question Most Authors Ask Wrong
“Should I use automatic or manual targeting?” is the wrong question. The right questions are: what is automatic targeting for, what is manual targeting for, how do they feed each other, and what does the optimisation cycle look like when both are running well?
Automatic targeting is a data source and discovery mechanism. It finds converting search terms you did not know existed and surfaces them in the Search Term Report. Manual targeting is an efficiency machine — it runs proven terms at optimised bids without the noise of Amazon’s broader automatic matches. One discovers; the other converts. You need both.
How Automatic Targeting Works
In an automatic targeting campaign, you set a bid (and optionally separate bids per sub-type) and Amazon takes full control of matching your ad to searches and product pages. Amazon reads your book’s metadata — title, subtitle, description, backend keywords, and categories — and uses that to decide which searches your ad is eligible for and which product pages are relevant to show it on.
You have no keyword list to build or maintain. You have no say in which specific searches trigger your ad, except through your negative keyword list (terms you explicitly exclude). The quality and specificity of your metadata directly determines the quality of Amazon’s automatic matches — more on this below.
Automatic campaigns are the only ad type that proactively reveals what your readers actually search for. The Search Term Report for an automatic campaign shows you the real, unfiltered vocabulary of your book’s audience — not what you guessed they search, but what they demonstrably searched before clicking your ad. This data is priceless and unavailable by any other means without buying it through ad spend.
The Four Automatic Sub-Types
Within automatic targeting, Amazon operates four distinct matching behaviours. You can set individual bids per sub-type, which allows meaningful budget control without giving up the discovery function entirely.
Close Match matches your ad to searches that are closely and specifically related to your book’s metadata — the tightest relevance filter. A cosy mystery set in a British village with “village mystery” and “amateur sleuth” in the keywords would match searches like “cosy village mystery books UK” and “amateur sleuth British detective fiction.” Close Match is typically the highest-converting automatic sub-type and should receive your default (or slightly above default) bid.
Loose Match broadens the relevance net to searches less directly related — tangentially connected genre terms, related subject areas, and broader category searches. The same book might match “British fiction cosy” or “light-hearted mystery books” under Loose Match — terms that might be slightly off the mark but can still find the right readers. Lower conversion than Close Match, useful for discovering peripheral audience vocabulary. Set Loose Match bids 25–30% below your default.
Substitutes places your ad on the product pages of books Amazon considers substitutes for yours — directly comparable titles where a reader is actively evaluating a purchase. This sub-type functions like a product targeting campaign — it reaches readers in active consideration mode for a book like yours, at the point they are already on a directly relevant product page. Conversion rates from Substitutes placements can match or exceed Close Match in many genres. Bid at your default or slightly below.
Complements targets product pages of books readers might buy alongside yours — adjacent rather than competitive. A reader on the page of a “how to write cosy mysteries” craft book might be shown your actual cosy mystery via a Complements placement. The relevance logic is indirect and conversion rates are typically lower — set Complements bids 30–40% below your default and review performance at 30 days to decide whether to maintain or pause.
Why Your Metadata Determines Automatic Targeting Quality
The quality of your automatic targeting results is a direct function of how well your metadata communicates your book’s genre, audience, and content. Amazon can only match what it can read. Vague, generic, or inaccurate metadata produces vague, generic automatic matches — and wasted spend.
Specific consequences: if your backend keywords are broad genre terms like “mystery” and “fiction” rather than specific terms like “cosy mystery British village amateur sleuth,” Amazon’s Close Match sub-type will produce much broader, lower-converting matches. If your book description does not include the trope vocabulary your genre readers use — “forced proximity,” “enemies to lovers,” “dark academia” — Amazon will not surface your ad for those trope-specific searches in automatic targeting. If your categories are incorrect or overly broad, Amazon’s Substitutes placements will target non-comparable books.
Before launching any automatic campaign, audit your metadata rigorously. See our KDP keyword research guide for the full metadata optimisation process. A well-optimised metadata layer makes automatic campaigns dramatically more efficient from day one and dramatically reduces the time to finding your first converting search terms.
How Manual Targeting Works
In a manual campaign, you specify exactly what to bid on — either specific keywords (with match types) or specific products (ASINs or categories). Amazon only enters the auction for searches or pages that match what you have specified. You have complete control over the targeting and full visibility into performance per target.
Manual campaigns are more efficient than automatic for proven targets because they contain no discovery overhead — there are no exploratory searches burning spend while you wait for signal. Every keyword or ASIN in a mature manual campaign has either been verified as converting or should be paused. The weakness is the reverse of automatic’s strength: manual campaigns cannot discover new terms. They can only perform as well as the list you give them. A manual campaign seeded with generic, unresearched keywords will underperform a well-maintained automatic campaign every time.
Manual Keyword Targeting
Manual keyword campaigns let you bid on specific search terms with defined match types (exact, phrase, broad). Building your initial keyword list requires combining several sources: your primary genre and trope vocabulary, comparable author search terms, Amazon autocomplete suggestions, and any converting terms you have gathered from research before launch.
The KDP Rank Fuel by Vappingo Book Keyword Spy tool lets you enter any comparable book’s ASIN and see every keyword it ranks for — effectively reverse-engineering the keyword strategy of proven bestsellers in your category. If three comparable books all rank consistently for “small town cosy mystery female sleuth UK,” that term has proven audience demand and belongs in your manual campaign from day one. This kind of pre-launch keyword research means your manual campaign is seeded with informed targets rather than guesses, and shortens the time to finding profitable terms considerably.
Amazon’s Rufus AI now influences which ads surface for conversational and semantic queries — “what should I read if I loved Agatha Christie but want something lighter?” This means keyword lists limited to exact-match genre terms may miss discovery opportunities that phrase and broad match types, combined with trope-specific vocabulary in your description, can capture. Build keyword lists that include the trope language your genre uses, not just the category-level descriptors.
Manual Product Targeting
Product targeting is the most underused element of book advertising and often the most cost-effective placement type available. When you target a specific ASIN, your ad appears on that book’s product page. When you target a browse category, your ad appears across all books in that category.
For ASIN targeting: build a list of 20–30 directly comparable books using the Competition Analyzer tool at app.vappingo.com. These should be books your target reader would genuinely consider buying instead of — or alongside — yours. Target their product pages and you reach readers who are already in consideration mode for a book in your exact space. The contextual relevance is as high as advertising gets.
For category targeting: select the 2–3 Amazon browse categories most precisely matching your book. Category targeting generates broader reach than ASIN targeting at typically lower average CPCs. Use it for building impressions volume while ASIN targeting drives more targeted, higher-intent traffic.
Product targeting campaigns should be separate from keyword campaigns. Their performance metrics read differently — impressions per ASIN vary widely, click-through rates differ from search placements, and bid optimisation is per-product rather than per-keyword. Mixing them in one campaign makes both unmanageable.
Connecting Them: The Harvest-and-Scale Workflow
The harvest-and-scale cycle is what connects automatic and manual targeting into a compounding system. Without it, you have two campaigns running in parallel that never improve each other. With it, every week of advertising makes both campaigns measurably better.
Step 1 — Run your automatic campaign for 14+ days. This is the minimum for statistically meaningful data given the 7-day attribution window.
Step 2 — Download the Search Term Report. Reports > Advertising Reports > Search Term Report, set to 14-day date range. This shows every actual search query that triggered an impression, along with clicks, spend, orders, and ACoS per term.
Step 3 — Identify harvest candidates. Sort by orders descending. Any term with 2+ orders at or below your target ACoS is a harvest candidate — a proven converter that belongs in your manual campaign at exact match.
Step 4 — Add to manual, negate in auto. Add each harvest candidate as an exact match keyword in your manual campaign with a bid reflecting its proven conversion value. Simultaneously, add it as a negative exact keyword in your automatic campaign. The negative prevents both campaigns bidding against each other on the same term — which inflates CPCs without benefit.
Step 5 — Identify negative keyword candidates. Any term with spend exceeding 1.5× your royalty (roughly £3.00 on a £2.00 royalty book) and zero sales is almost certainly never going to convert. Add it as a negative keyword in your automatic campaign.
Step 6 — Repeat weekly. Each cycle, your automatic campaign becomes more efficient as its negative list grows and proven terms are removed to manual. Your manual campaign gains proven terms at optimised bids. After 60–90 days, the difference in campaign quality compared to an unmanaged equivalent is substantial.
When Each Type Should Dominate Your Spend
Weeks 1–4 (launch phase): Automatic does most of the work. Your manual campaign exists with initial seed keywords, but the auto campaign is where most impressions happen and most discovery occurs. Budget split: 60% automatic, 40% manual.
Weeks 5–12 (optimisation phase): Manual grows as harvest cycles run. The auto campaign becomes more efficient as negatives accumulate. Budget begins shifting toward manual as proven terms in manual consistently outperform auto’s ACoS. Target split: 50/50.
Month 4+ (efficiency phase): Manual contains a well-curated list of proven converters and generates the majority of profitable ad-attributed sales. Automatic continues as a perpetual discovery engine at a smaller budget fraction. Target split: 30% automatic, 70% manual.
These ratios are guides, not targets. Let your actual performance data determine where budget goes. If your auto campaign at month 4 is still generating converting terms at good ACoS, maintain its budget. If your manual campaign is consuming its budget profitably and you cannot scale it further without adding unproven keywords, that is a product page or keyword ceiling issue, not a targeting type issue.
What Happens If You Run Only Automatic
An automatic-only strategy produces campaigns that plateau. After the first 30–60 days, the Search Term Report repeats itself — the same terms converting, the same terms wasting spend — without any mechanism for improvement. Without harvest cycles feeding proven terms to a manual campaign, you keep paying broad-match prices for terms that could be handled at exact-match efficiency. Without negative keyword management, the wasted spend fraction never shrinks. The campaign generates consistent mediocre results indefinitely instead of improving over time.
Many authors who report that “Amazon Ads don’t work” have been running automatic-only campaigns for months. The campaigns are working — they are generating data — but that data is not being used to improve anything.
What Happens If You Run Only Manual
A manual-only strategy can be profitable if your initial keyword research is excellent — but it has a ceiling and a discovery problem. Without an automatic campaign running alongside it, your manual campaign’s keyword list never grows beyond what you knew at launch. You miss the converting terms that readers actually use but that you did not know to target. Over time, without fresh discovery data, manual-only campaigns stagnate on their initial keyword list, which gradually loses relevance as the market shifts. The keyword list also ages: terms that were competitive at launch may become more crowded, and new trope and genre vocabulary that emerges in reader search behaviour never enters your targeting.
Practical Setup for Both Simultaneously
Day one setup for a new book should include all three of the following campaigns, launched at the same time:
Auto-Discovery. Automatic targeting, separate bids per sub-type (Close Match and Substitutes at default; Loose Match and Complements 25% below), dynamic down-only bidding, £5–£8/day budget.
Manual-Keywords. Your best initial keyword research in exact and phrase match, separated into ad groups by theme (e.g., one group for genre terms, one for author comparison terms, one for trope terms), dynamic down-only, £5–£8/day.
Manual-Products. 20–30 comparable ASINs plus 2–3 relevant categories, fixed bids starting at your keyword campaign default, £3–£6/day.
Total starting daily spend: £13–£22. Over 30 days this totals £390–£660 — a meaningful but manageable investment for learning what works for your book, before scaling what is working.
Bidding in Automatic Campaigns
Start with dynamic bids — down only across all sub-types. This gives Amazon the ability to reduce bids when clicks are unlikely to convert, but prevents it spending above your set amount. Set your per-sub-type bids relative to each other as described: Close Match and Substitutes at £0.30–£0.45 for most book categories; Loose Match and Complements at £0.20–£0.30. Adjust at the 30-day mark based on which sub-types are generating orders at acceptable cost.
Placement bid modifiers on automatic campaigns: only apply after checking your Placement Report. If top-of-search in your auto campaign is converting at significantly better ACoS than product pages, a small modifier (+15–20%) is justified. Do not apply modifiers in week one — you do not have enough data to know where your best placements are.
Bidding in Manual Campaigns
Manual keyword bids should reflect each keyword’s proven conversion value, not a single default applied across all terms. A keyword with 25 clicks, 4 sales, and 22% ACoS is worth bidding up — you can afford to capture more impression share at that conversion rate. A keyword with 20 clicks, 1 sale, and 65% ACoS should have its bid reduced by 20%, not paused immediately — one sale at 65% ACoS on 20 clicks is a small sample; give it another 14-day window at the reduced bid before making a final decision.
The bid ladder approach: adjust bids in 15–25% increments, not dramatic cuts or increases. A keyword at 60% ACoS against a 35% target should be reduced by 20%, not cut to half its previous bid. Dramatic cuts often drop a keyword out of auctions entirely rather than improving its ACoS.
The Targeting Mistakes That Cost the Most
Running automatic only and treating it as a complete strategy. Automatic discovers terms. Manual converts them efficiently. Without manual, you leave all the efficiency gains of proven exact match targeting untapped.
Never connecting the two campaigns via the harvest cycle. Running both campaigns without the weekly harvest-and-scale cycle means they operate in parallel but never improve each other. The harvest is the compounding mechanism — without it, there is no compound.
Failing to negate in auto when adding to manual. Not adding the exact match negative to the automatic campaign when promoting a term to manual means both campaigns compete for the same search — inflating CPC with no benefit. This is the single most common and most expensive omission in the harvest cycle.
Assuming all four automatic sub-types perform equally. Close Match and Substitutes typically outperform Loose Match and Complements by a significant margin for most book types. If you set identical bids across all four from day one and never check, you are paying top-of-search prices for bottom-of-funnel placements.
Abandoning automatic campaigns once manual is established. Automatic remains a perpetual discovery engine even when your manual campaign is mature. Reader search behaviour evolves — new tropes emerge, new comparable authors become breakout hits, seasonal search patterns shift. Automatic campaigns running continuously surface these evolutions. Stopping auto targeting to “save budget” cuts off the discovery pipeline and your manual campaign slowly ages.
For the Search Term Report mechanics that power the harvest cycle, see our dedicated Search Term Report guide. For the full campaign structure these targeting types sit within, see our complete Amazon Ads guide for authors.
The ads connect readers with your book — professional manuscript proofreading from Vappingo ensures the book they find meets the professional standard that earns the reviews your next campaign depends on.
