Amazon PPC keyword match types control how a positive keyword can match a shopper query: broad explores related intent, phrase preserves a phrase or its close variations within a query, and exact provides the tightest matching—but still includes close variations. Running the same keyword in multiple match types does not automatically make you bid against yourself. The waste to prevent is duplicated work, unreadable reporting, and funding overlapping targets without a distinct purpose.
This guide focuses on manual Sponsored Products keyword targeting and moving useful search terms into deliberate targets. Platform facts were checked on 2026-11-04 against Amazon Ads’ match-type documentation. Automatic targeting groups and product targeting are separate systems, not additional positive keyword match types.
Amazon PPC Keyword Match Types: What Each One Can Match
A keyword is the target you add. A customer search term is the term reported for the traffic. The second is what you need to inspect before deciding that a match type brought the right buyer.
Amazon’s match-type help, checked 2026-11-04, says broad, phrase, and exact all account for plurals, misspellings, and translations. Its phrase and exact descriptions also mention acronyms and close variations. Exact therefore does not mean a byte-for-byte string lock.
| Positive match type | Matching boundary described by Amazon | Best working role | Common misreading |
|---|---|---|---|
| Broad | Terms can appear in any order; synonyms and related terms can qualify, and the keyword itself may be absent | Discover related shopping intent | Every word must appear somewhere in the query |
| Phrase | The phrase or sequence, including close variations, appears within the query | Explore modifiers around a useful phrase | Only literal copies of the phrase qualify |
| Exact | The keyword or sequence matches, including close variations | Give a proven intent separate bid and reporting control | Only one spelling and wording can receive traffic |
Amazon’s examples make the boundaries clearer. Its broad keyword sneakers can match queries such as basketball shoes and trainers. Its phrase keyword curtain rod can match extra long curtain rods. Its exact example allows outside lights to match outdoor lights (official help, checked 2026-11-04).
These examples describe eligible matching, not guaranteed impressions. Use them to understand why a row can appear, then evaluate the actual traffic your product receives. Do not build a supposedly exhaustive yes/no matrix for every synonym; Amazon’s broad matching also depends on keyword meaning and the advertised product’s context.
Give Broad, Phrase and Exact Different Jobs
Choose the match type from the question you are trying to answer.
Broad: What relevant language have we missed? Start from a product need, not a generic word barely connected to the listing. The output you want is a list of useful search terms and intent families. A broad target that discovers no new actionable terms may not deserve continued funding simply because it produces clicks.
Phrase: Which modifiers matter around this intent? Phrase is useful when the central product description is already relevant and you want to examine size, use-case, material, or audience modifiers. A successful phrase target should teach you which extensions fit the advertised variation and which describe a different product.
Exact: Which established intent deserves direct control? Add an exact target when the observed query is relevant and its performance justifies its own bid decision. Exact does not make a weak listing convert or erase the close-variation boundary. Continue reviewing its actual search terms.
You do not need all three match types for every keyword. Start with a seed map from Amazon keyword research, then add a target only when it has a defined discovery or performance job. Keep the advertised ASIN and variation visible in that map: a profitable term for a large pack may not work for a single unit.
Duplicate Keywords Are Not Automatically Duplicate Auction Costs
Amazon’s Sponsored Products targeting guide, checked 2026-11-04, explicitly says adding different match types on the same keyword does not make you bid against yourself. Its match-type help also says that when the same keyword has multiple match types with different bids, the highest-bid match type is used for a matching query.
That is a reason to inspect bids, not a promise that exact always receives traffic first. A broad bid above an exact bid can frustrate the routing pattern you expected. Do not assume a universal exact-over-phrase-over-broad priority rule regardless of bids and eligibility.
The operational costs of blind duplication are more concrete:
- Fragmented evidence: one customer query appears under multiple target rows, leaving each row looking less meaningful in isolation.
- Repeated maintenance: the same commercial decision must be applied to several targets and campaigns.
- Unclear objectives: multiple targets buy similar intent, but nobody can explain which is discovering and which is scaling.
- Misread bid changes: you adjust the exact target while most relevant traffic is still reported elsewhere.
Keep multiple match types if each serves a job you can measure. Consolidate when they merely duplicate administration. Separate campaigns can provide separate budgets, but they also add structures to maintain; splitting every seed into its own campaign is not necessary for a readable account.
For the wider campaign economics and objective choices, use the profit-first PPC playbook. Here, the unit of work is the query and its intended destination.
Build a Search-Term Migration Table Before Moving Anything
Export the search term report and preserve campaign, ad group, advertised product, targeting keyword, match type, customer search term, clicks, spend, and the sales or purchase metric you use. Keep the reporting dates and account type with the file.
Amazon’s search term report documentation, checked 2026-11-04, says only terms with at least one click appear. It also says customer search terms can be inferred in non-search contexts. A reported term is therefore not always a literal typed query, and an absent term is not proof of zero impressions.
First group the same reported query across its source targets for the same advertised product and comparable period. Review the combined commercial result, then inspect the contributing rows. Do not merge different variations or unlike attribution metrics simply because the query text matches.
Use this worksheet; the examples are hypothetical decisions, not campaign results:
| Observed row or pattern | Destination decision | Source decision | Verification |
|---|---|---|---|
| Relevant broad query with convincing performance | Add the query as an exact target | Keep discovery active initially | Confirm destination eligibility and traffic |
| Phrase target reveals a profitable modifier family | Add useful individual queries as exact; retain phrase if it still discovers | Preserve the productive discovery role | Review whether new relevant modifiers continue to appear |
| Same query appears under broad, phrase, and exact | Name the intended performance target | Review bids and the purpose of each source | Compare combined query economics before and after |
| Relevant query with weak economics | Review bid, offer fit, and listing conversion | Do not copy automatically | Reassess against the product’s margin |
| Incompatible size or use case | No positive destination | Send to the exclusion workflow | Check the scope of the proposed exclusion |
| ASIN rather than a word-based term | Evaluate as a product target | Do not paste it blindly into this keyword workflow | Use the product-targeting decision sheet |
For ASIN rows, move to Amazon PPC product targeting. For a query worth harvesting, record the original wording rather than immediately shortening it to a head term. The observed modifier may be exactly why the traffic fits your product.
Migrate in an Order That Leaves a Rollback Path
Create the destination before restricting the source. The purpose of harvesting is more deliberate control, not making a report look tidier by removing traffic.
- Validate the term. Confirm product relevance and review its economics over an appropriate attribution window. Use your own margin target; the ACoS guide explains that boundary.
- Check for an existing destination. Search the relevant campaigns and ad groups before adding another identical target. Update the migration ledger with its location.
- Add the desired positive match type. Amazon’s match-type help, checked 2026-11-04, says initially selected match types cannot be changed directly, but keywords with different match types can be added while the campaign runs.
- Verify delivery. Check target status, the advertised product’s eligibility, bids, available campaign budget, and subsequent search-term activity.
- Decide whether routing is needed. If the source still buys the intended performance traffic and that overlap prevents useful control, review a source-level routing change. If both targets have distinct productive roles, overlap alone is not a reason to remove one.
- Log and reassess. Record the source, destination, change date, reason, and rollback action. Compare equivalent periods rather than a complete historical source window against a fresh destination.
When routing calls for negatives, use the existing negative-keyword guide for match choice and scope. Avoid applying an account-wide exclusion simply because one discovery source needs steering. This article covers the positive destination and the verification sequence, not a second negative-match tutorial.
When an Exact Destination Receives No Useful Traffic
An exact target with no activity does not prove that harvesting failed or that the source should be blocked more aggressively. Work through the delivery chain.
| Check | What to look for | Why it matters |
|---|---|---|
| Destination | Correct product, active target, intended keyword and match type | A naming convention is not evidence of a working target |
| Campaign budget | Whether the destination campaign can still serve | An unavailable budget prevents the planned comparison |
| Bid relationship | Whether other eligible match types have higher bids | Exact is not an unconditional priority rule |
| Exclusions | Whether the destination is inside the affected scope | A routing edit can suppress the traffic you intended to preserve |
| Query boundary | Whether the observed traffic still fits the new target | Shortening a query can change what you are testing |
| Measurement | Comparable dates, products, and attribution metrics | A fresh incomplete period can misrepresent the result |
Do not change every item simultaneously. Correct a configuration issue first; then evaluate the commercial decision. Preserve the working discovery route until the replacement earns its role, or restore it if your routing edit removed useful coverage.
Also continue reviewing close variants after migration. Exact narrows matching; it does not remove the need to check whether a particular variant fits the advertised offer.
A Match-Type Review Checklist
- Every positive target has a discovery, modifier-learning, or performance role.
- Query reviews combine comparable rows before judging individual match types.
- Broad targets are assessed for useful discoveries, not reach alone.
- Phrase targets are assessed for relevant modifier families.
- Exact targets are reviewed for close variations and actual delivery.
- Destinations exist and work before sources are restricted.
- Each migration has a recorded change date and rollback path.
Frequently Asked Questions
Does exact match include only the exact words I enter?
No. Amazon’s help, checked 2026-11-04, includes close variations and gives the example of exact outside lights matching outdoor lights. Treat exact as the tightest positive match type, not a literal-string guarantee.
Should the same keyword run in all three match types?
Only if each has a purpose you can evaluate. Amazon says this does not make you bid against yourself, but you still need to manage bids, overlapping reporting, and campaign objectives. A small, purposeful target map is easier to maintain than automatic triplication.
Can I add a plus sign to require a word in Sponsored Products broad match?
Amazon’s current match-type documentation, checked 2026-11-04, describes the plus-sign broad match modifier under Sponsored Brands. Do not transfer that feature into a Sponsored Products targeting plan; choose an appropriate positive match type for this workflow.
Is automatic targeting another keyword match type?
No. Amazon’s help, checked 2026-11-04, distinguishes automatic targeting groups from manual keyword match types. If you need the initial campaign setup rather than positive-keyword routing, start with Amazon PPC for beginners.