How Amazon's auction system works

Every time a customer searches on Amazon, an auction takes place in milliseconds. All advertisers bidding on relevant keywords take part. Amazon evaluates two factors here: your bid and the relevance of your ad to the search query.

Amazon uses a second-price auction. This means: the winner doesn't pay their own bid, but one cent more than the second-highest bid. If you bid $1.50 and the next-best bidder bids $1.20, you pay $1.21 per click. So the bid is a maximum, not a fixed price.

Definition

In a second-price auction, as the winner you only pay one cent more than the second-highest bid. Your set bid is the maximum amount you're willing to pay, not the actual CPC (cost per click).

Besides the bid, relevance plays a decisive role. Amazon wants to show customers relevant results because that leads to purchases, and Amazon earns from them. That's why an ad with a lower bid but higher relevance can beat an ad with a higher bid. Relevance is determined among other things by the click-through rate (CTR), the conversion rate and the listing quality.

In practice this means: a higher bid alone doesn't guarantee a top position. But if two ads are equally relevant, the higher bid wins. The art lies in setting the bid so that you win auctions without paying more than necessary.

The three bidding strategies in detail

At the campaign level, Amazon offers three bidding strategies. Each strategy changes the way Amazon handles the bid you set.

Dynamic bids: down only

"Down only" is the default setting for new campaigns and the most conservative strategy. Amazon automatically lowers your bid when the probability of a conversion is low. But your bid is never raised above the value you set.

When does Amazon lower the bid? For example, when the search query does match your keyword, but historical data shows that clicks from this position or for this search query rarely convert. Amazon then reduces the bid by up to 100%, which means that in some cases you don't take part in the auction at all.

Recommended for: Most campaigns in regular operation. You keep cost control, and Amazon automatically filters out unpromising auctions. Particularly suitable for ACoS-focused sellers who want to save budget.

Dynamic bids: up and down

"Up and down" gives Amazon the greatest control over your bid. Amazon can raise the bid by up to 100% when there's a high conversion probability, and lower it by up to 100% when the probability is low.

An example: your set bid is $1.00. For a search query with high conversion probability, Amazon could bid up to $2.00. For a search query with low probability, it could lower it to $0.30 or even $0.00.

Recommended for: Campaigns with enough historical data (at least 50 conversions in the last 30 days), aggressive launch phases and campaigns where revenue growth is more important than ACoS control. Caution: this strategy can significantly raise the CPC.

Fixed bids

With "fixed bids", Amazon uses your bid exactly as you set it. It's neither raised nor lowered. You bid the same amount in every auction, regardless of the predicted conversion probability.

Recommended for: Brand-awareness campaigns where impressions are more important than conversions. Also useful for tests where you want to keep the CPC factor constant in order to analyze other variables in isolation. For performance campaigns, fixed bids are usually not the best choice.

Strategy Raise bid Lower bid Ideal for
Down only No Up to -100% Regular operation, ACoS control
Up and down Up to +100% Up to -100% Launches, aggressive scaling
Fixed bids No No Awareness, controlled tests

Placement modifiers: control bids by ad position

In addition to the bidding strategy, you can set placement modifiers at the campaign level. These increase your bid for certain ad positions by a percentage markup.

Top of search (first page)

The top-of-search position is the first ad slot in the search results, directly above the organic results. This slot usually has the highest click-through rate and the best conversion rate. A placement modifier of 50% on "top of search" means: if your base bid is $1.00, Amazon bids up to $1.50 for the top position.

Product pages

This modifier applies to ads that appear on product detail pages of competitors or related products. Customers on product pages have a different purchase intent than searchers. They compare specifically and click on alternatives. Whether a modifier is worthwhile here depends heavily on the product and the pricing strategy.

Practical tip

In the placement report, analyze which position delivers the best conversion rate for your campaigns. Often top of search is the clear winner. Only set the modifier where the data confirms a profitable ROAS. Start with +25% and increase gradually.

Placement modifiers and the bidding strategy work together. With "dynamic bids: up and down", a top-of-search modifier of 50% and a base bid of $1.00, the effective bid can theoretically rise up to $3.00 ($1.00 base, +100% through up and down = $2.00, +50% top of search = $3.00). That sounds like a lot, but if the top position has a conversion rate of 15% and your product costs $30, it can definitely be worth it.

Optimize Amazon PPC automatically

Sellantica creates, manages and optimizes your Amazon PPC campaigns. Try it free for 30 days.

Start for free

Calculating the optimal bid: the formula

Instead of guessing bids, you can derive them mathematically. The basis for this is your target ACoS and the historical conversion rate of the keyword.

The formula is:

Maximum CPC = selling price x target ACoS x conversion rate

An example: your product costs $25.00. Your target ACoS is 25%. The keyword's conversion rate is 10% (every tenth click leads to a purchase).

Maximum CPC = $25.00 x 0.25 x 0.10 = $0.625

So you should bid a maximum of $0.63 per click to keep your target ACoS of 25%. If you bid more, the ACoS rises above your target. If you bid less, you may win too few auctions.

The extended formula with break-even

If you want to know the absolute maximum amount you can bid without making a loss, use the break-even ACoS:

Break-even CPC = selling price x profit margin x conversion rate

At a selling price of $25.00, a profit margin of 30% and a conversion rate of 10%:

Break-even CPC = $25.00 x 0.30 x 0.10 = $0.75

Anything over $0.75 per click would mean a loss for this keyword. The break-even CPC is your absolute maximum, which you should never exceed.

Definition

The break-even ACoS equals your profit margin before advertising costs. If your margin is 30% and your ACoS is also 30%, you make exactly zero profit and zero loss. Anything above it is a loss-making business, anything below it is profitable.

When to choose which bidding strategy

The right strategy depends on the context. Here are the most important scenarios and the matching recommendation:

Scenario 1: Launching a new product

At launch you have no historical data. Amazon can't reliably estimate the conversion probability. Nevertheless, you need visibility.

Recommendation: Start with "down only" and a slightly higher base bid (10-20% above the suggested bid). Set a top-of-search modifier of +25-50%. That way you secure visibility without "up and down" raising the budget uncontrollably before data is available.

Scenario 2: Scaling a profitable campaign

You have a campaign with an ACoS well below your target and want more revenue at consistent profitability.

Recommendation: Switch to "up and down" and raise the base bid gradually by 10-15%. Set a top-of-search modifier of +50-100%. Amazon's algorithm has enough data to adjust the bids intelligently. Monitor the ACoS daily in the first week after switching.

Scenario 3: ACoS is too high

Your ACoS is above the target value and you want to make the campaign more efficient without pausing it.

Recommendation: Switch to "down only" if not already active. Lower the base bid by 15-20%. Remove any placement modifiers. Check whether the keywords are individually profitable. Pause keywords with an ACoS that's more than 50% above the target value.

Scenario 4: Brand-protection campaign

You bid on your own brand name to prevent competitors from appearing there.

Recommendation: "Down only" with a moderate bid. Brand keywords usually have a high conversion rate (20-40%), so even low bids are profitable. A top-of-search modifier of +50% ensures you keep the first position for your own brand.

Manual vs. automated bidding

The fundamental question is: should you adjust bids manually or leave them to an algorithm? The answer isn't either/or, but a question of scale.

Manual bidding

With manual bid adjustment, you regularly analyze keyword performance and adjust bids individually. The advantage: you keep full control and can factor in contextual information an algorithm doesn't know (e.g. upcoming promotions, seasonal trends, inventory levels).

The disadvantage becomes clear quickly as your portfolio grows. With 50 campaigns of 30 keywords each, you have 1,500 bid decisions to make. Weekly. That's not only time-consuming but also leads to many keywords being reviewed too rarely.

Rule-based bidding

Rule-based bidding works with defined if-then rules. Example: "If a keyword's ACoS in the last 7 days is above 35% and it had at least 10 clicks, lower the bid by 15%." These rules are executed automatically.

The advantage: scalability. The rules apply to all keywords simultaneously. The disadvantage: rules are rigid. They don't account for interactions between keywords and can make wrong decisions in unusual situations (e.g. Prime Day).

AI-powered bidding

AI-powered systems like Sellantica go a step further. They analyze historical data, recognize patterns and adjust bids proactively. Unlike rigid rules, they take the overall context into account: time of day, day of the week, seasonality, keyword clusters and the interaction between keywords.

The advantage: you get the precision of manual adjustments with the scalability of automation. The disadvantage: you have to trust the system and give it enough data to make good decisions.

Practical tip

Start with manual bids until you have enough data (at least 4 weeks, better 8 weeks). Then switch to rule-based or AI-powered bidding. But always keep the option to intervene manually. Automation is a tool, not a substitute for strategic thinking.

The bid-adjustment workflow

Regardless of whether you work manually or automated, you need a systematic workflow for bid adjustments. Here's a proven process:

Step 1: Collect data (be patient)

Before you adjust a bid, the keyword needs enough data. The rule of thumb: at least 15 clicks, better 20-30. With fewer clicks, the statistical significance is too low. A keyword with 5 clicks and 0 orders isn't necessarily bad. It simply doesn't have enough data yet.

Step 2: Evaluate performance

Compare the keyword's actual ACoS with your target ACoS. There are three possible outcomes:

  • ACoS below target: The keyword is profitable. You can raise the bid to gain more impressions and clicks.
  • ACoS in the target range: The bid is optimal. No change needed.
  • ACoS above target: The keyword is too expensive. Lower the bid or check whether the conversion rate can be improved (listing optimization).

Step 3: Adjust the bid

Always change bids in small steps. An adjustment of 10-20% per week is a good benchmark. Larger jumps can cause you to suddenly get no more impressions or to raise the ACoS uncontrollably.

To calculate the new bid:

New bid = current bid x (target ACoS / current ACoS)

Example: current bid $0.80, current ACoS 40%, target ACoS 25%.

New bid = 0.80 x (25 / 40) = 0.80 x 0.625 = $0.50

Since a reduction from $0.80 to $0.50 corresponds to a decrease of 37.5%, you should implement it in two steps: first lower it to $0.65, observe for a week, then to $0.50.

Step 4: Check the result

After each bid change, wait at least 5-7 days before intervening again. Amazon's data has a delay of 24-48 hours (attribution window). Changes that are too frequent cause you to optimize based on incomplete data.

Common mistakes in bid optimization

Even experienced sellers regularly fall into the same traps with bids:

Mistake 1: Adjusting too much too fast

Anyone who wants to push a keyword's ACoS from 40% to 25% overnight lowers the bid too aggressively and loses all impressions. The keyword gets no more clicks, and you have no more data to learn from. Patience and gradual adjustments are the key.

Mistake 2: Never adjusting bids

The opposite is just as harmful. Anyone who sets their bid once and then forgets it gives away money. The market changes constantly: new competitors, seasonal fluctuations, changes to your listing. Bids need to be reviewed regularly.

Mistake 3: Up and down without a data basis

Setting "dynamic bids: up and down" on a new campaign without historical data gives Amazon permission to double your bid without the algorithm knowing whether the clicks convert. Always start with "down only" and only switch once enough conversions are available.

Mistake 4: Placement modifiers without analysis

Setting a top-of-search modifier of +100% because "top of search surely converts better" is expensive and often wrong. First check in the placement report whether the top position actually has the best conversion rate for your campaign. For some products, "rest of search" converts better.

Practical tip

Document every bid change with the date, old bid, new bid and the reason for the change. That way, after 2-3 weeks you can trace which adjustments worked and which didn't. Tools like Sellantica log this automatically.

Summary: bid systematically instead of guessing

Bid optimization isn't an art, but a craft. Those who understand the basics and proceed systematically achieve better results than someone working on gut feeling. The most important principles:

  • "Down only" is the safe default for most campaigns. Start with it.
  • "Up and down" you switch on when enough data is available and you want to scale.
  • "Fixed bids" you use only for awareness campaigns or controlled tests.
  • Placement modifiers you set based on data from the placement report, not on assumptions.
  • You calculate the optimal bid with the formula: CPC = selling price x target ACoS x conversion rate.
  • Bids are adjusted in small steps (10-20%), never in big jumps.
  • After each change, wait at least 5-7 days before intervening again.
  • Automation saves time but doesn't replace the strategic understanding behind the numbers.

If you follow these principles and work on your bids regularly, you'll see a clear improvement in your PPC performance over time. Every small bid optimization adds up, and the difference between a well-optimized and a neglected account can amount to thousands of dollars a month.