Amazon Ads Targeting: Keywords, Products & Geo (2026)

What Is Amazon Ads Targeting?

Amazon Ads targeting is the set of controls that tell Amazon who and what to match your ad against — keywords a shopper typed, products and categories a shopper is browsing, and (on Sponsored Display and Amazon DSP) audiences defined by who the shopper is. The single most useful reframe is this: targeting on Amazon Sponsored Products is a taxonomy of WHO and WHAT, not WHERE. Search advertisers arrive expecting to name a city or a radius. On Amazon, the levers are matching signals, not map pins — and that one correction resolves most of the confusion.

The honest mental model runs on two axes. The first axis is automatic versus manual: does Amazon pick the matches from your listing, or do you name them yourself? The second axis is keyword versus product versus audience: what signal are you matching on? Amazon's own targeting guide organizes it exactly this way — automatic targeting uses keywords and products similar to the product in your ad, while manual targeting lets you choose keywords or products yourself (Amazon Ads, A guide to targeting with Sponsored Products). Every other concept on this page hangs off those two axes.

Targeting is set inside a campaign, and each ad product exposes a different slice of the toolkit. See Amazon Sponsored Ads for how Sponsored Products, Sponsored Brands, and Sponsored Display differ. The rich audience and geographic controls Sponsored Products lacks all live in Amazon DSP and Streaming TV. MB Adv Agency treats this taxonomy as the foundation every account structure is built on — name the axes first, then choose the matching signal.

What Every Seller Should Know About Amazon Targeting

• Targeting is WHO and WHAT, not WHERE. Sponsored Products matches on keywords (what the shopper typed) and products/categories (what the shopper browses). Audiences and geography are Sponsored Display and Amazon DSP features, not Sponsored Products knobs.
• Automatic targeting has four buckets, each with its own bid. Close Match and Loose Match are keyword-side; Substitutes and Complements are product-side. You bid each group separately — bid up what converts, bid down what wastes.
• Automatic is a permanent discovery engine, not a beginner setting. Mature accounts run auto and manual together: auto finds converting terms and competitor ASINs, manual scales them, and the harvested terms are negated in auto so the two never bid against each other.
• Broad match is now semantic. Since 2025, Broad serves on synonyms and related queries even when your exact keyword is absent (Karooya, Dec 2025). Without a negative-keyword layer, Broad behaves like a discovery campaign, not a targeted one.
• There is no Google-style geo on sponsored ads. State, city, DMA, ZIP, and radius targeting (2 to 100 miles, up to 1,000 locations per line item) exist only in Amazon DSP (Adverio; Amazon Ads). If location genuinely drives your business, that is the reason to evaluate DSP.

Amazon Ads Targeting by the Numbers (2026)

Before choosing a targeting type, anchor on the platform-wide benchmarks that frame what “normal” performance looks like. Ad Badger tracks millions of bids across thousands of Sponsored Products and Sponsored Brands campaigns; the blended 2026 averages below are theirs, not mbadv data.

Read these as the gravity the whole page works against. Average CPC of $1.22 sets the floor every bid sits above; the 11.1% blended conversion rate is what Exact-match discipline pushes higher and unsupervised Broad pulls lower; and the 29.6% ACOS is the efficiency line that placement bid adjustments and per-group automatic bidding move in either direction. None of these numbers is a goal in itself — they are the backdrop that makes a targeting decision legible. A category running well below the blended conversion average is a candidate for tighter Exact and ASIN targeting; one running well above it absorbs wider Broad and category reach.

These are blended averages; the right target depends on which targeting type and which category you run. ACOS and ROAS are two sides of the same efficiency story — see Amazon Ads ACOS, ROAS, and metrics for how to read efficiency once a targeting structure is live. The targeting choices on this page are what move those numbers up or down.

Automatic vs Manual Targeting: The Reach-vs-Control Ladder

The clearest way to see the targeting tradeoff is conversion rate by mode. Tighter control concentrates spend on proven intent and lifts conversion; wider reach finds new demand at a lower conversion rate. Ad Advance's managed-account analysis of US Sponsored Products campaigns (September 2025) puts numbers on the gradient.

The ladder is not an instruction to abandon Broad and automatic — it is an instruction to use each for its job. Ad Advance also reported that automatic campaigns' share of total impressions jumped ~7 points beginning June 2025, reaching 40%+ of impressions in many accounts, with Loose Match and Substitutes gaining the most share. That makes per-group bid discipline more important, not less. Auto-targeting is only a discovery engine if you mine it — see Amazon Ads reporting and attribution for the Search Term Report that surfaces the keywords and ASINs to promote and negate. For broad-catalog skincare brands, the ladder is how a launch budget migrates from discovery into proven-converter scale.

Automatic Targeting: The Four Match-Type Groups

When you open an automatic Sponsored Products campaign's reports, spend is split across four rows: Close Match, Loose Match, Substitutes, and Complements. These are the four targeting groups Amazon matches your ad across, and each answers a different question. Close Match and Loose Match are keyword-side (what the shopper searched); Substitutes and Complements are product-side (what the shopper is browsing).

You can set a single default bid for the whole campaign or a separate bid for each group — and the separate bid is the entire point. Karooya's December 2025 ranking is directional: Close Match tends to deliver the highest conversion and lowest ACOS, Loose Match moderate conversion at higher volume, and Substitutes and Complements lower conversion (niche, but Complements can uncover incremental demand). No sourced numeric conversion figure exists for the individual groups, so treat the ranking as a bidding compass, not a guarantee.

Manual Keyword Targeting: Match Types and Negative Keywords

Manual keyword targeting offers three match types — Broad, Phrase, and Exact — in descending order of reach and ascending order of control. The standard structure bids Exact highest and Broad lowest. The 2026 wrinkle is that Broad match has become semantic: it now matches synonyms, related terms, and conceptually adjacent queries even when none of your keyword's words appear. That changes the negatives you need.

Read the match types and the negatives as one system, not two topics. Because semantic Broad now reaches conceptually adjacent queries, a Broad campaign without a negative-keyword layer is a discovery campaign — budget and read it as one. Negative phrase blocks a wasteful theme; negative exact blocks one specific query. Placement and bidding tune what the match types reach — see Amazon Ads bidding and budgets for descending-bid laddering. For baby product brands, where safety-adjacent searches drift fast, negative hygiene is the difference between Broad as a discovery tool and Broad as a budget leak.

The Five Ways to Tell Amazon Who to Target

The pillar's spine is this: there are five distinct ways to tell Amazon who to target — keyword, product (ASIN), category, audience, and contextual — and each matches on a different signal and lives in a different ad product. Keyword and product/category targeting are Sponsored Products and Sponsored Brands controls. Audience and contextual targeting are Sponsored Display and Amazon DSP controls. Confusing the two is the most common structural error.

Read the right-hand column as a routing map. If you need keywords or ASIN/category conquesting, that is Sponsored Products or Sponsored Brands. If you need audiences — in-market, lifestyle, interests, life events, or views/purchase remarketing — that is Sponsored Display or DSP. Each ad product exposes a different slice of this toolkit; see Amazon Sponsored Ads for the full split. For sports equipment brands with overlapping use cases, mapping each campaign to one row of this table keeps the account legible as it scales.

Product and Category Targeting: Conquesting and Shelf Defense

Product and category targeting is the lever search advertisers do not have. Instead of matching a query, you place your ad on a named ASIN or across a whole category. Target a specific competitor's detail page to conquest — your cheaper, higher-rated product on their listing — or target your own category to defend the shelf so a rival's ad is not the only one a shopper sees. Category targeting refines by brand, price range, star rating, and Prime eligibility.

Category economics differ structurally too, which informs where conquesting pays off. Ad Badger's 2026 category benchmarks show an inverse pattern — the highest-CPC categories tend to have the lowest conversion.

MB Adv Agency reads ASIN targeting as the precision instrument and category targeting as the reach instrument: conquest a named rival when you know exactly whose shopper you want, defend the category when the goal is presence. For watch brands and handbags and accessories brands, where a handful of rival ASINs own the consideration set, named-ASIN conquesting is the highest-leverage play on the page.

Placements: Top of Search, Rest of Search, and Product Pages

Targeting decides who you match; placement decides where the matched ad appears. Sponsored Products serves in three placements, and they are not equal. Top of Search is the prime real estate — the first row above organic results, the highest-intent and highest-converting slot. It captures roughly 67% of all Amazon ad clicks (PPCAssist) and converts two to three times better than Rest of Search (Autron). You bid into it with placement bid adjustments now reaching up to +900%.

The CTR gap between placements is wide. Sellerite's 2025 analysis of 10,000+ US, UK, and DE ASINs puts Top of Search sponsored CTR at 2.9–4.5%, Rest of Search at 1.2–2.3%, and product pages at 0.5–1.0%.

One caution: placement adjustments compound with dynamic bidding. A $1.00 base bid with a +900% Top of Search adjustment and dynamic bids up-and-down (which adds up to 100%) reaches near $20 for a single auction (ScaleInsights) — calculate the maximum theoretical bid before enabling both. Placement and bidding are two halves of the same lever; see Amazon Ads bidding and budgets for the full mechanics. For kitchen appliance brands defending high-consideration purchases, winning the Top of Search slot on branded terms is the placement worth paying the premium to hold.

Geographic Targeting on Amazon: The Correction

Here is the correction that most pages get wrong. Amazon does not offer Google-style geographic targeting on Sponsored Products or Sponsored Brands. There is no city, state, DMA, ZIP, or radius control on sponsored ads — they serve across the entire marketplace you advertise in, not a geo you draw. Adverio states it plainly: geo-targeting is not available for any type of Sponsored Products or Sponsored Brands ads. Granular geo lives only in Amazon DSP.

The strategic answer to “how do I geo-target Amazon ads?” is therefore: you do not, on Sponsored Products — and if location genuinely matters, that is the reason to move that budget into Amazon DSP. The DSP self-serve spend minimum was removed at unBoxed 2025, which makes that path more reachable than the old “DSP needs ~$50K” framing implied; it is still DSP, not Sponsored Products. Adverio notes Amazon briefly tested geo in Sponsored Display around Valentine's Day 2024 and withdrew it. MB Adv Agency states this plainly to clients: the honest answer to geo-targeting on Amazon is usually “that lives in DSP.” The full geo and audience toolkit is detailed in Amazon DSP and Streaming TV. For mattress and sleep brands and outdoor and camping brands with regional demand, DSP is where location control actually exists.

Audience Targeting Lives in Sponsored Display and DSP, Not Sponsored Products

The same correction applies to audiences. Rich audience targeting — in-market, lifestyle, interest, and life-event audiences, plus views and purchase remarketing — is a Sponsored Display and Amazon DSP capability, not a Sponsored Products knob (Amazon Ads targeting guide). Sponsored Display also renamed its product targeting to contextual targeting, so on Display you combine contextual (what the shopper browses) with audiences (who the shopper is). The remarketing layer — re-reaching shoppers who viewed your detail page or bought before — is the retargeting lever most brands underuse, and it is why budget that needs audience precision belongs on Sponsored Display or DSP rather than waiting for a setting Sponsored Products does not have.

Building the Targeting Mix: Auto vs Manual by Account Stage

The right targeting mix shifts as an account matures. A new account with no keyword history leans on automatic targeting to discover demand; a scaled account with identified winners leans on manual control while keeping automatic running as a discovery layer. The framework below is editorial — an MB Adv Agency starting point, not a sourced multi-account dataset.

Within the manual budget at the Scale stage, MB Adv Agency allocates the larger share to keyword targeting (Phrase plus Exact) and the remainder to product/ASIN targeting — the within-manual split is editorial guidance, not a sourced benchmark. The point that holds across every stage: automatic never turns off. Ad Advance's finding that automatic accounts for 40%+ of impressions in mature accounts confirms the discovery engine keeps earning its place. The whole loop — harvest from auto, scale in manual, negate the overlap — is the core of an optimization workflow; see Amazon Ads reporting and attribution for the report that drives it.

The percentages are deliberately round because they are a starting posture, not a prescription. A brand that launches with a strong existing keyword list compresses the Launch stage and tilts manual earlier; a brand entering a crowded category with no history holds the auto-heavy posture longer to let discovery run. What stays constant is the direction of travel — budget migrates from automatic discovery toward manual control as data accumulates, and automatic never reaches zero. MB Adv Agency revisits the split every time the Search Term Report produces a new cohort of proven converters, because each promotion into manual is what justifies trimming the automatic share.

What Changed in Amazon Targeting for 2026

Three shifts reshaped how targeting behaves in 2026, and each changes a default that worked in prior years. Semantic Broad match, a structural rise in automatic-campaign impression share, and a unified placement bid-adjustment ceiling together mean the old “set Broad and forget it” and “auto is for beginners” instincts now actively cost money.

Broad match went semantic. Karooya confirmed in December 2025 that Amazon's Broad match now uses semantic matching — your ad shows for synonyms and related terms even when your exact keyword is absent. CaptenAMZ corroborates that a Broad campaign without negatives behaves like a discovery campaign. The practical consequence: the negative-keyword layer is no longer optional hygiene, it is the control surface that keeps Broad from spending into adjacent intent.

Placement bid adjustments unified at +900%. Amazon extended the Rest of Search bid adjustment to match Top of Search and Product pages — all three now cap at +900%, for every targeting type and every bidding strategy. That gives sellers independent control of all three placements for the first time, and it raises the stakes on the compounding-with-dynamic-bidding math. Efficiency is read through ACOS and ROAS once these levers are live — see Amazon Ads ACOS, ROAS, and metrics. For home decor brands tuning seasonal demand, independent placement control is the lever that earned its own line item in 2026.

Three Misconceptions About Amazon Targeting That Cost Real Money

Three imported assumptions cause the most expensive targeting mistakes on Amazon: hunting for geo controls that do not exist, killing automatic targeting as if it were training wheels, and treating Broad match as if it stayed inside your keyword. Each is declaratively wrong, and each is backed by Amazon documentation or web-verified 2026 platform behavior.

Misconception 1: “Amazon has Google-style geographic targeting.” It does not, on Sponsored Products or Sponsored Brands. Geo-targeting is not available for any type of Sponsored Products or Sponsored Brands ads (Adverio). State, city, DMA, ZIP, and radius targeting (2 to 100 miles, up to 1,000 locations per line item) exist only in Amazon DSP (Amazon Ads). Treating Sponsored Products as if it had Google's radius controls sends advertisers hunting for a setting that does not exist. When location truly matters, the correct move is to evaluate DSP, not the sponsored-ads console.

Misconception 2: “Automatic targeting is a beginner setting you outgrow.” It is a permanent discovery layer. Its four groups continuously surface new converting search terms and competitor ASINs that you promote into manual campaigns (Trellis). Mature accounts run auto and manual at once — auto to find, manual to control — with harvested terms negated in auto so the two never bid against each other. Killing auto throws away the discovery engine.

Misconception 3: “Broad match keeps my ad inside my keyword.” It does not. Since 2025, Broad serves on synonyms, related terms, and conceptually adjacent queries even when none of your keyword's words appear (Karooya). A Broad campaign with no negatives is not tightly targeted — it is a discovery campaign, and should be budgeted and read as one. For skincare brands with crowded ingredient terms, that distinction decides whether Broad builds the account or drains it.

Frequently Asked Questions: Amazon Ads Targeting