The bid is not everything — what actually decides Amazon ad auctions.

"I raised my bid and nothing happened" is one of the most common questions in any Amazon PPC support thread. The answer is almost always the same: the bid is not the only thing the auction prices. Amazon runs a quality-weighted auction, and quality factors quietly multiply on top of whatever you bid.

What Amazon's ad auction actually optimises

Amazon's goal in the auction is not to pick the highest bidder. It is to maximise expected revenue per impression, where expected revenue is roughly:

bid × predicted CTR × predicted conversion rate × predicted order value

Every factor on the right-hand side is multiplicative. An ASIN with double the predicted conversion rate at half the bid will out-rank a higher-bidding competitor — because Amazon's expected revenue is higher.

The quality signals that move the multiplier

Conversion rate on the keyword

Conversion rate is the single largest non-bid factor. Improving CVR from 8 % to 12 % is worth more in ad rank than a 30 % bid increase, on average, and it lowers your CPC at the same time because Amazon is happier serving your ad.

Click-through rate on the placement

CTR matters most for top-of-search. A weak main image or thin title can suppress your CTR enough that Amazon stops serving you the slot even at a competitive bid.

Historical performance

Amazon's models lean on the ASIN's history. A campaign on a brand-new ASIN starts cold and is initially over-priced until the model has data. This is why launch CPCs are always higher than steady-state CPCs on the same keyword.

Buy Box state

Lose the Buy Box, lose the ad. We covered this in Episode 03; it bears repeating here because it is the single non-bid factor that takes you from "delivering" to "zero impressions" the fastest.

Listing completeness

Bullet count, A+ content, image count, video presence, keyword coverage in title — all feed Amazon's relevance model. Each one is small individually; together they explain why two near-identical products at the same bid deliver very different impressions.

What this means in practice

• Before raising a bid, ask "is the bid the constraint?" Check impression share and CTR. If CTR is low, the bid is not the problem.
• New campaigns deserve a 7–14 day patience window before you judge them. The model is still learning.
• Listing improvements compound with PPC. The same €1.20 bid wins a different volume of impressions on a well-optimised listing than on a thin one.