How Amazon Sponsored Products Targeting Actually Works

What Targeting Actually Controls

Targeting is the setting that decides which searches and product pages your Sponsored Products ad is allowed to appear against. It's separate from bidding — targeting decides eligibility, bidding decides how hard you compete for that eligible placement. Get targeting wrong and no bid fixes it, because you're paying to be relevant to the wrong shopper in the first place.

Amazon gives you three targeting types inside Sponsored Products: automatic, manual, and negative. Manual splits again into keyword targeting and product targeting. None of these are mutually exclusive at the account level — most mature accounts run all of them at once, just not inside the same campaign, because Amazon locks the targeting type once a campaign goes live.

The Three Targeting Types, Side by Side

Here's what each one actually does, who's driving it, and where it earns its keep.

How Automatic Targeting's Four Strategies Split Your Budget

Automatic targeting runs on four strategies, and you can bid on each one separately instead of accepting one blended bid. Say HexClad is running automatic targeting on a cast iron skillet listing:

• Close match — shoppers searching close to the listing itself, e.g. "cast iron skillet."
• Loose match — broader, related searches, e.g. "cookware set."
• Substitutes — product pages for other skillets, cast iron or not.
• Complements — product pages for things bought alongside a skillet, like trivets or oven mitts.

Close match usually converts best because intent is highest. Complements usually gets the most impressions for the least conversion, because a shopper looking at oven mitts isn't necessarily buying a skillet. The most common early mistake is leaving all four on one default bid: complements and loose match soak up budget with a fine-looking CTR and a weak conversion rate, and the campaign looks healthy in aggregate while quietly wasting spend one strategy at a time. Split the bids, watch complements on its own, and cut it if it hasn't earned its place after a few weeks of data.

Manual Targeting: Keywords vs Products, and the Match Types That Change Everything

Manual keyword targeting is you choosing the exact search terms, with three match types controlling how loosely a shopper's search has to resemble your keyword before it triggers your ad:

• Broad match — matches searches containing your keyword's terms in any order, plus close variations.
• Phrase match — matches searches containing your exact phrase, with words allowed before or after.
• Exact match — matches only the search term itself, or very close variants.

Manual product targeting skips keywords entirely and lets you pick specific ASINs or whole categories, refined by price band, star rating, and Prime eligibility. Choose keyword targeting when you already know how shoppers search for what you sell. Choose product targeting when you want to appear on a specific competitor's page, defend your own listing from being targeted, or sell an accessory against the product it's used with.

Negative Targeting and the Auto-to-Manual Workflow

Negative targeting doesn't generate impressions — it removes them. Add a search term or ASIN as negative and your ad stops being eligible to show against it, in either an automatic or a manual campaign.

The workflow that actually works: run automatic targeting to find out what shoppers search, pull the search term report once enough data accumulates, move the terms that convert into a manual exact-match campaign where you control the bid precisely, then add those same terms as negative exact inside the automatic campaign. Skip that last step and you're bidding against yourself for the same shopper in two campaigns at once — we've done this on our own accounts, watched CPCs creep up for no visible reason, and traced it back to an automatic campaign that never got the negative added after harvesting.

When Your Targeting Data Says Something's Wrong

Targeting problems show up as one of three symptoms, and each has a different fix:

• ACOS is climbing on automatic targeting. Check which of the four strategies is spending — usually loose match or complements. Lower that specific bid or negative out the worst offenders before touching the campaign as a whole.
• You harvested a converting term into manual and it stopped converting. This usually isn't a targeting problem. Check the listing itself — price, stock, star rating — before assuming the targeting setup is broken.
• You added negatives and spend didn't drop. Confirm you added the negative at the right match type and in the right campaign; a negative phrase match added when the waste was coming in on broad match won't block it.

And the one setting people forget: once a campaign is live, you cannot switch it between automatic and manual. If you picked wrong, the fix is a new campaign, not a settings change.

A Note on Benchmarks

One habit worth breaking: treating a single average as the whole picture. It happens inside Sponsored Products when someone quotes "average ACOS" for a category without saying which match type or strategy it came from, and it happens across channels too. In our own DSP book, one CPM and CPC figure gets quoted constantly — but that same book ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost-per-click, while the $0.41 CPC people quote in this category is online-video only, not the whole book. A number without its source line isn't a number you can act on.

That distinction matters more once Sponsored Products isn't running alone. Once Sponsored Display or DSP enters the mix, a purchase can get credited to both a search ad and a display ad on a last-click basis, and last-click can't tell you which one actually caused the sale. reMKTR runs Amazon DSP as a managed service and reconciles that overlap in Amazon Marketing Cloud, using holdouts and matched controls rather than last-click, specifically so a brand can tell whether display added anything beyond what search targeting was already going to close. It's part of Full Circle, which has managed over $500M in Amazon ad spend across more than 100 brands.