Broad, Phrase, and Exact Match on Amazon Ads: What Each One Actually Controls

What Broad, Phrase, and Exact Match Actually Control

Match type controls how closely a shopper's search term has to resemble the keyword you bid on before Amazon lets your ad compete for that search. It is not a quality setting and it does not make your ad better — it makes your targeting looser or tighter.

Broad match is the loosest. Amazon can show your ad for search terms that share the theme of your keyword — synonyms, related words, different word order, sometimes both. Bid on "stainless steel water bottle" broad and you might show up for "insulated metal flask" or "gym water bottle 32 oz."

Phrase match requires your keyword's words to appear in that order somewhere inside the search term. Amazon can still add words before or after it. "Stainless steel water bottle" phrase will match "best stainless steel water bottle for hiking" but not "steel stainless bottle" or "water bottle, stainless steel, gift."

Exact match requires the search term to match your keyword almost word for word. Amazon allows close variants — plurals, minor misspellings, a stray function word — but nothing wider than that. It's the tightest of the three, and the one where you're paying, almost exclusively, for the keyword you actually chose.

A Worked Example: One Keyword, Three Match Types

Run the same keyword and the same bid across all three match types for a week and the shape of the result is predictable, even if your exact figures won't match this. Say you bid $0.75 on "stainless steel water bottle" under each match type separately:

• Broad: the widest net — most impressions, most clicks, but a chunk of those clicks come from loosely related searches that were never going to convert. Highest wasted spend, highest discovery value.
• Phrase: fewer impressions than broad, but the search terms it pulls in sit closer to what you actually sell. Click volume drops, conversion rate usually climbs.
• Exact: the fewest impressions of the three, but almost every click lands on the term you picked. If that term converts well, this is where ACOS is lowest. If it doesn't, you have nowhere to hide — every dollar spent was your choice.

The mistake is treating this as a ranking — broad isn't worse than exact. Broad is doing a discovery job. Exact is doing a harvesting job. The numbers above are illustrative, not a benchmark to hit; your own search term report is the only place to get real ones.

When to Use Each — the Funnel, Not the Favorite

Most accounts that get this right run all three match types at once, structured as a funnel rather than picking a single best one:

• Broad for new keywords, new products, or categories with no search term history yet. Its job is to surface search terms you didn't think to bid on.
• Phrase for keywords you've already validated but still want to catch nearby variations of. Its job is to scale spend on a proven theme without opening targeting all the way up.
• Exact for keywords you already know convert, where you want maximum control over bid and ad rank without paying again to relearn what broad and phrase already taught you.

The connective tissue is the search term report and the negative keyword list. When broad or phrase surfaces a term that converts, add it as its own exact match keyword and negate it out of the looser campaigns — so you stop paying twice to learn the same lesson.

Common Mistakes — Including Ones We've Made

The most common mistake is running broad match with no negative keyword list and checking back a month later. Broad match doesn't get more efficient over time on its own — it accumulates irrelevant spend until someone actually reads the search term report.

The second most common mistake is the opposite: going exact-only too early, before enough search terms have been mined to know which exact keywords are worth bidding on. We've done this ourselves — tightened an account to exact match to chase a target ACOS, and starved it of the new search terms that would have kept sales growing. ACOS looked great for a few weeks. Total sales did not.

A third, quieter mistake: assuming exact match means zero waste. Amazon's close-variant matching means an exact match keyword can still pull in a plural, a minor misspelling, or an added word — it's tighter, not sealed.

What to Do When Changing Match Type Doesn't Fix It

If you tightened from broad to phrase or exact and ACOS didn't move, the match type wasn't the problem. Check the listing's conversion rate first — a match type change can't fix a product page that isn't converting the traffic it already gets. Check the bid next — an exact match keyword with a bid too low to win auctions just shows fewer impressions, not better ones.

If you loosened from exact to broad to find volume, and CPC spiked instead, add negatives before you add more budget. The fix for a broad match that's gone wide isn't switching straight back to exact and losing the discovery entirely — it's a negative keyword list built off the search term report you now have.

A match type change is really a hypothesis: this keyword, this tightness, will move performance in this direction. Treat it that way. At Dr. PPC, every proposed change — including a match type change — carries three things before it runs: the evidence behind it, a measurement plan, and a rollback trigger. If the measurement plan shows the change didn't work, the rollback trigger reverses it instead of leaving it live out of inertia.