Amazon Bid Strategy: How to Set and Adjust PPC Bids

What 'bid strategy' actually means

On Amazon, 'bid strategy' is a campaign-level setting with three options, and it changes what happens to the bid you've already set — it doesn't replace it. Fixed bids use your exact CPC every time. Dynamic bids — down only let Amazon lower your bid in real time when a click looks less likely to convert, and never raise it. Dynamic bids — up and down let Amazon raise your bid by up to 100% at the top-of-search placement and up to 50% elsewhere, on top of lowering it, when the auction looks favorable.

This used to be called 'Bid+' before Amazon folded it into the current three-option structure. The setting sits above your actual bid, not instead of it — so the real work is still getting the base bid right before you decide how much room to give Amazon to move it.

How to calculate a starting bid, with real numbers

Most sellers set a bid by feel and adjust when ACOS looks wrong. A better starting point is a formula: CPC bid = target ACOS × conversion rate × average selling price. It's not exact — nothing about an auction is — but it gives you a defensible number instead of a guess.

Say the product sells for $40, converts at 10% on that keyword, and the target ACOS is 30%. Bid = 0.30 × 0.10 × $40 = $1.20. That's the fixed-bid number you'd enter for a keyword with that profile. Change any input — a price drop, a conversion rate that's actually 7% not 10%, a tighter 20% target — and the correct bid moves with it. The formula is only as good as the conversion rate you feed it, which is why it needs revisiting, not setting once and forgetting.

What dynamic bidding does to that number

Set that $1.20 bid to dynamic bids up and down, and at top-of-search Amazon can push it as high as $2.40 for a single auction it judges likely to convert. If that click converts at the same 10% rate, the real ACOS on that click is $2.40 ÷ ($40 × 0.10) = 60%, not the 30% you designed for. Multiply that across enough top-of-search clicks and your blended ACOS drifts up without any single change explaining why.

That's not a flaw in dynamic bidding — placements that convert better can absolutely justify a higher bid. The mistake is turning it on everywhere and not checking the placement report afterward to see whether the extra spend actually bought better conversion, or just bought more expensive clicks at the same rate.

The mistakes that actually cause the damage

• Turning on dynamic bids up and down across every campaign without checking placement reports, effectively doubling exposure on top-of-search and not noticing until the invoice.
• Recalculating the bid formula off a 7-day conversion rate on a keyword with ten clicks a week — the rate is noise, and the bid swings on noise.
• Cutting a bid the moment ACOS spikes without checking whether the real cause was a price change, a stock-out, or a competitor's listing update. We've made this one ourselves: dropped a bid to fix ACOS when the actual problem was a broken buy box, and the bid cut just cost impressions on top of the existing problem.
• Never revisiting a bid after a price change. The formula depends on price. A $5 price drop changes the correct bid by a meaningful margin and almost nobody goes back to recheck it.

When the bid is wrong — what to check before you touch it again

If a bid change didn't do what you expected, don't immediately reverse it and move on — find out why first. Check whether the conversion rate you used to calculate it was based on enough clicks to trust. Check the placement report to see if dynamic bidding pushed spend somewhere it didn't pay off. Check whether something else moved at the same time — price, stock, a competitor's coupon — because a bid change gets blamed for problems it didn't cause.

The discipline that actually prevents this is treating every bid change like a hypothesis, not a fix: know what evidence justified the change, know how you'll measure whether it worked, and know the number that triggers you to undo it. Every proposed change carries three things before it runs: the evidence behind it, a measurement plan, and a rollback trigger. That's true whether a person is making the change by hand or software is doing it — the discipline is what's missing from most PPC bid strategies, not the bidding tool.