What is Bid Optimization?

Bid optimization is the work of setting each keyword and target bid from your own economics, then adjusting it as new evidence arrives. The goal is not the lowest ACoS. It is the most profit, or the most growth you have chosen to pay for, at a cost per click your margin can carry.

This entry defines the term and its formula. For the full process, from setting a target to pricing placements and deciding how often to change bids, read the Amazon bid optimization guide.

Definition

Your bid is the most you are willing to pay for a click. Amazon then adjusts it inside the auction. The campaign bidding strategy can move it: Amazon says dynamic bids up and down can raise or lower a bid by up to 100% based on performance (Amazon Ads guide, checked 2026-10-03). Placement adjustments then apply on top of that strategy (Amazon Ads Support, checked 2026-10-03).

Bid optimization decides the starting number those layers multiply. Amazon's suggested bid is a reference for what the auction costs. It knows nothing about your margin.

The formula

ACoS is an output. Its only inputs are what you pay per click and what a click earns:

Revenue per click (RPC) = Sales ÷ Clicks
ACoS                    = CPC ÷ RPC
Target CPC              = RPC × Target ACoS
Target ACoS             = Break-even ACoS × Profit retention

Break-even ACoS is your margin before advertising as a share of price. Profit retention sets how much of that margin advertising may spend, and so how much of it you keep. A target set this way moves with your costs instead of being a number you copied from somewhere.

A target with no orders yet needs a different form, because RPC is zero. Project it instead:

Target CPC = Target ACoS × AOV ÷ (Clicks so far + Clicks per order)

Each click without a sale lowers the result, so the bid winds down gradually rather than falling off a cliff.

Worked example

Suppose a product sells for $40, Amazon's fees are $12 and the landed cost is $10. All figures here are example inputs.

• Contribution margin = 40 − 12 − 10 = $18
• Break-even ACoS = 18 ÷ 40 = 45%
• Target ACoS at 0.75 retention = 45% × 0.75 = 33.75%

Thirty days of results for three keywords:

• Keyword A costs $1.50 a click and earns $3.20 a click, so its target CPC is $3.20 × 0.3375 = $1.08. It is above break-even, and its bid comes down toward that number.
• Keyword B costs $0.80 and earns $4.00, so it could carry $4.00 × 0.3375 = $1.35. There is room to raise it.
• Keyword C has no orders. If the product typically needs about 12 clicks per order, the projection is 0.3375 × $40 ÷ (40 + 12) = $0.26.

The max CPC calculator works out the same ceiling for a single keyword.

Using it in practice

Work from the CPC you actually paid rather than the current bid, leave out the newest days while their sales are still being attributed, and move bids in measured steps rather than by fixed percentage rules. Placement adjustments and the bidding strategy multiply the bid you type, so check the effective bid, not just the number you set. The Amazon bid optimization guide covers search terms, placements, time of day, stock and how often to change bids.

How it relates to adjacent metrics

Bids set CPC. Conversion rate and average order value set RPC. ACoS is the ratio of the two, ROAS is its inverse, and TACoS shows what the result means against total sales. Break-even ACoS sets the ceiling every target is measured against. Timing is a separate layer, covered under dayparting.

How Shurq helps

Shurq computes break-even ACoS for each product from its own P&L, which needs the product's cost of goods entered. Targets are set as break-even times a profit retention setting, with Profit-First, Balanced and Growth presets or a custom target. An inventory guard holds ad groups whose products are out of stock, blocks raises when stock runs short and lets cuts slow sales so the stock lasts to the restock, and placement adjustments are set from each placement's own results.

Three bid managers share that math. The Bid Optimizer previews every proposed change so you apply only what you agree with. Automate manages bids toward one target, with a suggestions-only mode where you approve changes before they apply. Rules run a playbook you define on your own schedule. Dayparting works with all three, raising and lowering bids by hour against the base bid, and applied changes are logged in Shurq's change and optimization history.