Amazon uses three manual keyword match types, broad, phrase and exact, plus a separate automatic targeting option. The fastest route to profit is a staged workflow: discover new search terms with automatic and broad campaigns, validate the strongest ones with phrase match, then scale confirmed winners into exact match with tighter bids and better control over spend.
TL;DR:
- Broad match is useful for discovering demand but should be monitored closely to prevent wasted spend, especially since it now includes the former broad match modifier.
- Automatic campaigns serve as discovery tools, generating search-term data that should be reviewed weekly to identify high-converting keywords for promotion.
- Only promote search terms to manual matches after they have enough clicks to reliably assess their conversion rate and profit support, avoiding premature scaling.
- Organize campaigns with clear naming conventions and allocate budgets to favor discovery early, shifting towards exact match as keyword validation solidifies.
- Use negative exact and phrase keywords to eliminate irrelevant traffic based on real search-term data, reviewing and updating negatives monthly to adapt to seasonal changes.
Table of Contents
- Manual match types explained with concrete examples
- Automatic targeting explained: close, loose, substitutes and complements
- Negative keywords: cleaning up discovery traffic properly
- Discovery, validation and scale: the workflow that ties it together
- Campaign structure and naming that keep optimisation simple
- Reading search-term reports without losing a week to it
- How match types work differently across ad products
- How Osellpa operationalises the discovery-to-scale routine
- Common match-type mistakes and how to avoid them
- Put the discovery-to-scale workflow on autopilot
- Where this information comes from
- Sources
- FAQ
Manual match types explained with concrete examples
Each manual match type controls how closely a shopper’s search has to resemble your keyword before your ad can appear.
Exact match is the most restrictive option. Your ad only shows for the keyword itself or a very close variant, such as a plural, misspelling or minor word order change. If you bid on “leather dog collar” in exact match, Amazon may also show your ad for “leather dog collars” but not for “leather collar for dogs”. This precision makes exact match the natural home for proven converters.
Phrase match sits in the middle. Amazon can add words before or after your keyword, but the phrase itself must stay intact and in order. Targeting “t-shirt” in phrase match can pull in searches like “black t-shirt” or “t-shirt for men”, giving you some reach without losing the core intent, as Search Engine Land explains.
Broad match casts the widest net. Amazon can match synonyms, related terms and looser variations of your keyword, so “running shoes” might surface for “trainers” or “jogging shoes” too. This makes broad match useful for finding demand you hadn’t anticipated, though it needs closer monitoring to avoid wasted spend.
Amazon retired broad match modifiers some years ago, folding that functionality into standard broad match, so sellers no longer need to add plus signs to lock in specific words.
A quick reference for how each type behaves:
- Exact match: keyword or close variant only, tightest control, best for confirmed winners.
- Phrase match: keyword plus surrounding words, moderate reach, good for validation.
- Broad match: keyword plus synonyms and related terms, widest reach, best for discovery.
Automatic targeting explained: close, loose, substitutes and complements
Automatic targeting on Sponsored Products is a separate mechanism from manual keywords. Amazon decides which searches and product pages trigger your ad, using four distinct strategies, according to Amazon’s own targeting guide.
- Close match shows your ad for search terms that closely resemble your product.
- Loose match shows your ad for search terms that are more broadly related.
- Substitutes places your ad on product pages for similar items shoppers might buy instead.
- Complements places your ad on product pages for items that pair naturally with yours.
Close and loose match respond to what shoppers type into the search bar, while substitutes and complements respond to which product pages they visit. Treating all four as one undifferentiated bucket hides where your sales actually come from. Splitting bids or reviewing performance by strategy tells you whether your budget is winning search traffic or product-page placements, and lets you adjust each independently rather than guessing from a blended average.
Automatic campaigns earn their keep as discovery tools. Success here doesn’t mean a low advertising cost of sale on its own, it means surfacing search terms and product associations you didn’t already have in your manual campaigns, ready to be reviewed and promoted.
Pro Tip: Run one automatic campaign per product (or tightly related product group) so its search-term data stays clean enough to act on.
Negative keywords: cleaning up discovery traffic properly
Negative keywords stop your ads showing for search terms that waste spend without producing sales, and they come in the same two flavours as positive targeting: exact and phrase, per Amazon’s advertising FAQ.
Negative exact blocks a single specific search term. If “cheap dog collar” burns spend with no conversions, add it as a negative exact and nothing else is affected.
Negative phrase blocks an entire family of searches containing that phrase. Adding “cheap” as negative phrase stops your ad on “cheap dog collar”, “cheap leather collar” and every other search containing that word.
A simple three-step process keeps this manageable:
- Identify irrelevant or non-converting terms from your search-term report.
- Classify each one as a one-off (negative exact) or part of a wider pattern worth blocking (negative phrase).
- Apply the negative at the campaign or ad-group level where the waste is occurring.
Add negatives only after reviewing real search-term evidence, never pre-emptively, since a term that looks irrelevant can still convert. Review your negative list monthly, because seasonal shifts in shopping behaviour can turn a previously useless term into a valuable one.
Discovery, validation and scale: the workflow that ties it together
Amazon’s own guidance recommends combining automatic and manual targeting rather than treating match types as competing choices, using automation to surface demand and manual targeting to scale what works, as Amazon’s keyword targeting guide sets out. Broad, phrase and exact aren’t rival options, they’re stages.
- Launch discovery campaigns. Start with automatic targeting and a broad match campaign per product, using a comfortable daily budget so Amazon has room to test search terms.
- Let data accumulate. Run discovery campaigns for one to three weeks before making judgements, since match types need enough impressions and clicks to show a pattern rather than noise.
- Review the search-term report. Export it, filter for terms with meaningful impressions and clicks, and flag anything converting profitably.
- Promote winners to phrase, then exact. Add validated terms as new phrase-match keywords first, then move the strongest of those into exact match once they’ve proven themselves again, adjusting bids upward as you tighten control.
- Add negatives to the discovery campaign. Once a term graduates to manual targeting, add it as a negative exact in the automatic or broad campaign so you stop paying twice for the same click.
A term is generally worth promoting once it has produced enough clicks to judge conversion rate reliably and has converted at an advertising cost of sale that supports your margin. This threshold varies by category and price point, but the principle holds across catalogues: don’t promote on impressions alone, and don’t wait so long that you miss the sales a validated keyword could be generating.
This staged approach, recommended directly in Amazon’s keyword targeting documentation, is also where search-term discovery work pays off, since the quality of your promotions depends entirely on how carefully you read that report.
Campaign structure and naming that keep optimisation simple
A clean structure is what makes match-type strategy practical week to week, rather than a spreadsheet exercise you dread opening.
- Run one automatic campaign per product or tight product group, dedicated purely to discovery.
- Run a separate broad match campaign for the same product, also focused on discovery but under your keyword control.
- Run a phrase match campaign that receives validated terms as they graduate from discovery.
- Run an exact match campaign that holds only your proven, highest-confidence keywords.
Naming conventions matter more than they seem to. A structure like “ProductName - Auto - Discovery”, “ProductName - Broad - Discovery”, “ProductName - Phrase - Validation” and “ProductName - Exact - Scale” means anyone on your team can tell a campaign’s purpose without opening it.
Budget allocation should favour discovery early on and shift towards exact match as your keyword list matures. Initial bids typically run highest on exact match, since you’re targeting confirmed intent, moderate on phrase, and more conservative on broad and automatic, where Amazon has more latitude to decide what a click is worth to you. Adjust from there based on the bid optimisation approach that fits your margins.
Reading search-term reports without losing a week to it
The search-term report is the single most useful document in Amazon PPC, and a short recurring routine keeps it from becoming a chore.
- Export the search-term report for each active campaign on a weekly or fortnightly cadence.
- Sort by spend to see where money is going first, then by conversions to see where it’s working.
- Add converting terms from automatic and broad campaigns into phrase or exact match.
- Add non-converting terms with meaningful spend as negatives, exact for one-off waste, phrase for a pattern.
- Adjust bids on existing manual keywords: raise bids on strong performers nearing budget caps, lower them on keywords burning spend without sales.
Splitting performance by match type, rather than looking at a campaign as a whole, is what reveals waste. A broad match campaign with a high advertising cost of sale might be hiding two or three genuinely profitable exact-level terms buried among poor ones, and you’ll only see that by breaking the data apart.
Pro Tip: Set a recurring calendar reminder for this review. A consistent weekly routine catches wasted spend before it compounds over a month.
How match types work differently across ad products
Manual match types aren’t universal across Amazon’s advertising products, and assuming they behave the same way everywhere leads to wasted budget.
- Sponsored Products fully support automatic and all three manual match types, broad, phrase and exact, making them the natural home for the discovery-to-scale workflow.
- Sponsored Brands also support manual keyword targeting but centre on a different placement (brand and product showcases at the top of search results), so testing here usually follows, rather than leads, your Sponsored Products data.
- Sponsored Display relies on audience and product targeting rather than keyword match types.
Billing differs too: Sponsored Products and Sponsored Brands are cost-per-click, while Sponsored Display supports both CPC and viewable CPM, as Amazon’s own product overview notes. That difference matters for match-type strategy, since a CPC model rewards precise keyword control in a way a viewable-CPM model doesn’t. Amazon itself recommends starting new advertisers with Sponsored Products, since it pairs simple CPC buying with the full range of manual and automatic targeting, before layering in Sponsored Brands or Display.
How Osellpa operationalises the discovery-to-scale routine
Running this workflow manually every week is realistic for a handful of products, less so across a full catalogue. Osellpa automates the repetitive parts: pulling search-term data, flagging candidates that meet your conversion thresholds, and adjusting bids by match type as performance shifts. That reduces the wasted spend that sits unnoticed in broad and automatic campaigns while search terms wait for manual review. A useful way to validate the approach is running one product manually alongside one managed through automation and comparing advertising cost of sale after a few weeks.
Common match-type mistakes and how to avoid them
Sellers most often go wrong by leaving broad match unmonitored for months, promoting terms to exact match too early on thin click data, and forgetting to add negatives once a term graduates. Before promoting a search term, check it has produced enough clicks to trust its conversion rate and that its cost per sale supports your margin. Match types reward patience more than cleverness.
— Harry
Put the discovery-to-scale workflow on autopilot
Running this routine by hand works until your catalogue grows past a few products, at which point the weekly search-term review becomes the job rather than a part of it. Osellpa connects directly to your advertising account and applies the same discovery, validation and scaling logic automatically, flagging promotable terms and adjusting bids by match type without you exporting a single report. If you want to see where your current campaigns are leaking spend first, the free PPC bid optimisation report gives you that picture before you commit to anything, and full plans start at Launch, Scale or Advanced depending on your sales volume.
Where this information comes from
- Amazon advertising match types: confirms the broad, phrase and exact definitions and examples used above.
- Targeting with Sponsored Products: Amazon’s own breakdown of the four automatic targeting strategies.
- Amazon Ads FAQ: the source for negative keyword rules and match-type mechanics.
- Keyword targeting guide: Amazon’s recommended discovery-to-scale workflow.
Sources
FAQ
What is a match type?
A match type is a setting on an Amazon PPC keyword that controls how closely a shopper’s search has to match your keyword before your ad can appear. Amazon offers three manual match types, broad, phrase and exact, alongside automatic targeting, which Amazon controls itself.
Can you give an example of a phrase match keyword?
Targeting the keyword “t-shirt” in phrase match lets Amazon add words before or after it, so your ad can show for searches like “black t-shirt” or “t-shirt for men”, as Search Engine Land illustrates. The core phrase and word order stay intact, which is what separates it from broad match.
Which keyword is of the phrase match type?
Any keyword entered under the phrase match setting qualifies, for example “leather dog collar” set to phrase match. It will trigger for searches containing that exact phrase with extra words attached, but not for reordered or entirely different wording.
What is the difference between broad match and exact match keywords?
Broad match lets Amazon show your ad for synonyms and loosely related searches, such as “trainers” matching a “running shoes” keyword, while exact match restricts your ad to the keyword itself or a very close variant like a plural or minor misspelling. Broad suits early discovery, exact suits keywords you’ve already validated as profitable.
