Amazon Sponsored Products PPC: How the Auction Sets Your Cost Per Click

What Sponsored Products PPC Actually Is

Sponsored Products is Amazon's cost-per-click ad format for individual listings. You pick which ASINs to promote, choose keywords or let Amazon's automatic targeting match your product to shopper searches, set a bid and a daily budget, and you're charged only when someone clicks — not for impressions.

Any professional seller, vendor, book vendor, KDP author, or agency account can run it. Ads can show at the top of search results, inside search results, on product detail pages, and increasingly off Amazon on select apps and sites. There's no monthly platform fee and no minimum spend to start — the cost is entirely the clicks you buy.

How Your Bid Becomes an Actual Charge

Your bid is a ceiling, not a price. When a shopper's search matches your targeting, Amazon runs an auction among every advertiser eligible for that placement. The winner is decided by bid combined with relevance and expected performance — not bid alone — and the actual CPC charged is typically lower than the bid, set just above what it took to beat the next competitor.

That gap between what you bid and what you pay is where most accounts leave money on the table or lose it. Here's what one keyword's numbers look like end to end:

Reading the Worked Example

Start with the actual CPC, not the bid — it's the number that determines whether the keyword can ever be profitable, and it moves independently of what you set. In the table above, the bid was $1.50 but the actual charge came in at $1.12, because the auction only needed to beat the next bidder by a cent.

ACOS of 60.8% is neither good nor bad on its own — it's only meaningful against your margin. If the product carries a 40% margin, that keyword is losing money on every sale it drives. If margin is 65%, it's still profitable. Sponsored Products reporting gives you every input in this table except your margin; that part you have to bring yourself.

When the Numbers Are Bad: What to Fix, in Order

High ACOS has a fixed troubleshooting order, and skipping steps is how accounts end up cutting a keyword that was never the problem.

• Check the search term report first. Automatic targeting and broad match both harvest irrelevant queries. Waste here inflates spend without touching sales, and it's the cheapest fix available.
• Look at placement performance before touching bids. Top-of-search and rest-of-search often carry very different ACOS on the same keyword; a blended number can hide a placement that's already profitable.
• Add negative keywords before cutting budget. Cutting budget slows the whole campaign, including the parts that work. Negatives remove only the waste.
• Give new campaigns two full weeks before judging them. Amazon's system needs volume to learn which queries convert; a campaign that looks expensive on day four often isn't the same campaign by day fourteen.
• Adjust bids last, and in small moves. A bid cut that looks reasonable on paper can take a keyword out of the auction entirely rather than just lowering its cost.

Mistakes We See Most, Including Ones We've Made

The single most common mistake is trusting an account-level average. A blended ACOS or blended CPC across every campaign hides which keywords are actually earning their spend and which are being carried by the rest. We've made this mistake ourselves at the DSP level: across 30 advertisers in our own book last month, the blended CPC came in at $1.42 — and the $0.41 CPC that gets quoted around this category is online-video only, not the whole book. Same trap, different channel: an average is a summary, not a diagnosis.

The second is treating a purchase as proof the ad worked. In that same book, 20.1% of purchases came from shoppers new to the brand — which means roughly four in five didn't. A keyword with a great ACOS that's mostly catching your own branded search looks identical in the report to one that's genuinely finding new buyers. Sponsored Products doesn't separate the two for you; you have to check whether the converting terms are your brand name or someone else's problem you're solving.

The third is leaving automatic targeting running indefinitely. Auto campaigns are excellent at finding terms you wouldn't have thought to bid on — but once a term proves itself, it belongs in a manual campaign where you control the bid precisely, not in an auto campaign where Amazon is still experimenting with it.

Where reMKTR Fits

Sponsored Products is a search-ads auction, and most sellers can run it themselves with the tools Amazon gives you for free. reMKTR doesn't manage Sponsored Products — we run Amazon DSP as a managed service, holding real DSP seats and reconciling in Amazon Marketing Cloud so display spend isn't double-counted against what search ads were already going to sell. If you're deciding whether Sponsored Products alone is enough, or whether display and DSP would add incremental sales on top of it, that reconciliation is the question we can actually answer with a holdout, not a dashboard.