Amazon's bidding strategy is a multiplier, not a setting — price the base bid first.

Your Amazon bidding strategy is not a setting you pick once — it is a multiplier stacked on your base bid, and choosing "up and down" on a bid that ignores your margin just tells Amazon to pay more, faster, on the auctions it likes. Every Sponsored Products campaign carries one of three bidding strategies, and most sellers leave it on the default without realising it is silently rewriting the price of every click. The strategy does not make a click more valuable. It changes what you pay for it — and if the base bid underneath was never priced against break-even, the strategy is just deciding how quickly you overspend.

What the three strategies actually do

Amazon gives Sponsored Products three campaign bidding strategies, and they are not three flavours of the same thing. Two of them adjust your bid in real time based on Amazon's own guess at conversion probability; one does not.

• Dynamic bids — down only. Amazon lowers your bid in real time when a click looks less likely to convert, and never raises it. Your bid is a ceiling. This is the default on new campaigns.
• Dynamic bids — up and down. Amazon raises your bid when a click looks more likely to convert and lowers it when it looks less likely. Per Amazon's own documentation, the uplift can reach up to 100% for top-of-search placements and up to 50% elsewhere. Your bid is a midpoint, not a ceiling.
• Fixed bids. Amazon uses your exact bid for every auction, with no real-time adjustment up or down. What you set is what you enter the auction with.

Read that middle one again. "Up and down" hands Amazon permission to double your top-of-search bid in the moment, on its own model of whether the click will convert — a model that is optimised for Amazon's revenue as much as yours. That is not a reason to fear it. It is a reason to make sure the number it is doubling was priced against your profit in the first place.

The multiplier no one shows you: strategy times placement

Here is the part the campaign UI keeps quiet. Your bidding strategy is not the only multiplier acting on the base bid — it compounds with your placement bid adjustments, which we unpacked in the Top-of-Search placement bid. Both are percentage uplifts, and on a top-of-search auction they stack on top of each other.

Final top-of-search bid = base bid × (1 + placement adjustment) × (1 + dynamic uplift). A €1.00 base bid with a +100% Top-of-Search adjustment becomes €2.00, and "up and down" can raise that a further 100% — so a €1.00 bid can clear at up to €4.00 on a single top-of-search auction. The number you typed is a floor for what you might actually pay, not the price.

Four times the bid you set is not a bug — it is the documented behaviour of two levers most accounts have switched on without connecting them. If that €1.00 was your break-even bid, the auction you just won at €4.00 lost money on the click, and the dashboard will still call it a conversion. The strategy did exactly what you told it to. You just did not know what you told it.

Why "down only" is not automatically the safe choice

The obvious reaction is to run everything on down-only and never let Amazon raise a bid. That caps the damage, but it is not free either. Down-only means you never bid up for the auctions most likely to convert — the exact clicks worth paying more for. On a high-margin product with room above break-even, refusing to bid up is leaving profitable volume on the table so a competitor on up-and-down can take it.

So the honest framing is not "up-and-down is aggressive, down-only is safe." It is that both are multipliers on the same base bid, and neither is a substitute for pricing that base bid correctly. Down-only protects a margin-blind bid from getting worse. Up-and-down lets a margin-aware bid capture the conversions it can afford. The strategy is the second decision. The base bid is the first.

Price the base bid, then choose the strategy

The base bid should come from the same place every profit-first number does: your break-even. A bid is affordable when the expected cost per sale it implies still clears your margin after Amazon's fees, COGS, and returns. If you have not set your ceiling yet, start with what a good ACoS actually is — the break-even ACoS in that piece is the same ceiling the bid has to respect.

1. Compute the break-even bid per ASIN: break-even ACoS × price × conversion rate. That is the most you can pay per click before the average advertised sale stops being profitable.
2. Set the base bid at or below that number, leaving whatever profit margin you want out of advertised sales — tighter for mature heroes, looser for a launch buying rank.
3. Choose the strategy against that base. Down-only where you want a hard ceiling on a thin-margin SKU; up-and-down only where the margin has room for Amazon to raise the bid and still clear profit.
4. Account for the stack. If you also run a placement adjustment, remember it multiplies with the strategy — model the worst-case combined bid, not the base, and make sure even that clears break-even on top-of-search.
5. Re-check as CPCs and conversion rates drift. A base bid that was break-even in Q1 is stale by peak season, and the multiplier on top of it drifts with it.

Where an autonomous system changes the equation

The reason strategy-versus-base-bid trips up so many accounts is that it is a per-auction decision made at a scale no human can watch. This is the gap Mirox is built to close. Its Tactician agent prices bid-level moves against the margin math directly, so the number entering the auction is already a break-even-aware bid rather than a guess Amazon is then free to double. And because every bid carries an exportable trace — the Bayesian conversion estimate, the alternatives ruled out, and which of the seven safety gates were checked — you can see the ceiling on a single bid instead of discovering it in a month-end report. The ROAS circuit breaker and per-marketplace spend caps sit underneath as the hard limit on how far any "up" can run.

The one-line version

Bidding strategy does not set your price — it multiplies it. Down-only caps a bid, up-and-down can double it, fixed leaves it alone, and none of them fixes a base bid that ignores your margin. Price the base bid against break-even first, model the multiplier stack second, and the strategy stops being a coin flip and becomes a guardrail.

See the break-even ceiling on a single bid, or read how the same multiplier logic plays out on Top-of-Search placement bids.