Amazon PPC match types are not a quality ladder. They are a control dial.

Match types do not decide whether you make money — they decide how much of the auction you can see, and how much control you keep over which shoppers you pay for. Broad, phrase, and exact are not a quality ladder where exact is "best" and broad is for amateurs. They are three different trade-offs between reach and control, and the seller who treats them as a ranking instead of a toolkit ends up either blind to demand or bleeding spend on terms that never convert. The match type sets the funnel. The bid still decides the profit.

The three match types, stripped of folklore

Every keyword you add to a manual Sponsored Products campaign carries a match type, and the match type tells Amazon how loosely it is allowed to interpret your word. That is the whole mechanism. It governs the width of the net, not the price of the catch.

Exact match shows your ad only for the keyword and its very close variants — plurals, misspellings, minor word order. Tightest control, narrowest reach. Phrase match shows your ad when the shopper's query contains your keyword as a phrase, with words allowed before and after but the core sequence intact. Broad match shows your ad for any query Amazon judges related — synonyms, related terms, and increasingly, semantic neighbours that share no words with your keyword at all. Widest reach, least control.

Note what is missing: broad match modifier, the old +keyword syntax, is gone — Amazon retired it years ago and folded its behaviour into phrase. If a 2021 guide tells you to use it, that is your signal the rest of the guide is stale too.

Broad match is not what it was — and that is the real 2026 story

The single most important thing to understand about match types today is that broad match has quietly stopped being a keyword setting and become a semantic one. It no longer just adds synonyms and plurals; it now leans on Amazon's own intent model to serve your ad against queries that share meaning rather than words. Used well, that is a discovery engine. Used carelessly, it is the fastest way to spray budget across queries you would never have chosen — the exact complaint sellers raise when a tightly-themed keyword starts matching to barely-related searches.

This is why "just use exact, it's safer" is bad advice dressed as caution. Exact-only is safe the way never leaving the house is safe: you stop wasting money and you also stop discovering the long-tail terms that quietly convert. The job is not to avoid broad. It is to run broad as a sensor with a conservative bid and a disciplined negative list, and to harvest what it finds into tighter control. That is the same logic behind running auto and manual campaigns together, permanently: a discovery layer and a control layer doing different jobs, not a beginner stage you graduate out of.

Negatives are match types too — and they are where the money is

Most sellers think of match types as a positive-targeting feature and forget that negatives carry match types of their own. Negative exact suppresses one precise query; negative phrase suppresses a whole family of queries containing a phrase. The choice between them is the difference between sniping a single wasteful search term and shutting down an entire seam of irrelevant demand.

This matters because broad and phrase only earn their reach if a negative list keeps them honest. A broad keyword with no negatives is not discovery — it is an open tab at the auction. The cheapest profit lever in most accounts is not a clever new keyword; it is the negative list nobody maintains, applied at the right match width so it blocks the waste without starving the discovery you actually want.

The EU wrinkle: match width behaves differently per language

Here is what US-centric match-type advice never tells European sellers: broad and phrase match depend on Amazon's understanding of shopper intent in each language, and that understanding is uneven across marketplaces. A broad keyword that behaves sensibly on amazon.de can sprawl across loosely-related queries on amazon.it or amazon.es, because the semantic match is weaker and the competitive data thinner. The same match type is a different amount of risk in each language.

So the right match-type mix is not a global template you clone across DE, FR, IT, and ES. It is set per marketplace, in each language, against each market's own data. This is the same trap that gives cross-marketplace copy its +28% ACoS penalty when it is shipped without per-region tuning: a setting tuned for one language quietly misbehaves in the next. Matching German keywords against German shopper intent — rather than translating them through English first — is the difference between broad match helping and broad match bleeding.

What actually decides profit, and it is still not the match type

You can run the cleanest exact-phrase-broad structure in your category and still lose money, because none of the three match types sets a bid. The match type decides which shoppers are eligible to cost you money; the bid decides whether each click is worth winning. And the only bid that protects profit is one priced against the margin of the specific ASIN behind the ad.

Break-even ROAS = 1 ÷ margin. A product carrying 35% margin after COGS and fees breaks even at about a 2.9x ROAS — a 35% ACoS. A click is worth bidding for only if its expected revenue clears that line with the profit you actually want on top. That math is identical whether the keyword is exact, phrase, or broad. The match type just decided whether the shopper was in the auction at all.

We make the same point about ROAS versus ACoS: the framing you obsess over is a distraction from the number both ignore, which is your margin. Match types are the structural cousin of that distraction. Pick the width deliberately, then spend your real attention on the bid.

How to use match types in 2026

1. Run all three with different jobs. Exact for your proven money terms, bid hard against break-even. Phrase for controlled expansion around those terms. Broad as a low-bid discovery sensor — never as your main spend.
2. Treat broad as discovery, not delivery. Give it a conservative bid, watch the search-term report, and harvest converting queries into phrase or exact where you can price them deliberately.
3. Use negative match types as a real tool. Negative exact to snipe a single wasteful query; negative phrase to shut down a whole irrelevant seam. A broad or phrase keyword without negatives is an unsupervised budget.
4. Stop the duplication. When a term graduates from broad to exact, negate it in the broader campaign so you are not paying to outbid yourself — the same self-competition leak that auto-and-manual structures create when nobody closes the loop.
5. Tune the mix per marketplace, never blended. The match-type balance that works on amazon.de is a guess on amazon.it; set it in each language and currency, the way CPCs themselves differ by region.

The one-line version

Match types choose how wide you cast and how much control you keep — exact for the terms you know, broad to find the ones you don't, phrase in between, and negatives to keep all of it honest. They set the funnel; they never set the bid. Your profit lives in what every eligible click is worth against the margin of the ASIN behind it, in the language and currency of the marketplace it runs in. The match type just decides who gets to be in the auction.

See the relevance score and margin context attached to a single bid, or read why the negatives nobody maintains are the cheapest profit lever you have.