How Amazon Keyword Bidding Works, With the Actual Math

What Happens When You Set a Bid

Amazon runs a real-time auction every time a shopper's search matches one of your targeted keywords. You set a max CPC bid — the most you're willing to pay for that click. Amazon then ranks eligible ads using both that bid and a relevance score built from predicted click-through and conversion rate. The ad that wins doesn't have to be the highest bidder; a highly relevant ad can beat a higher bid at a lower cost.

What you actually pay is usually less than your max bid. Amazon's auction works like a near second-price system, meaning you typically pay just enough to beat the next-best competing ad, not your full ceiling. That's the part sellers miss when a bid looks alarmingly high on paper: the max bid is a ceiling, not a price tag.

The Bid Formula, With Real Numbers

There's a standard formula experienced advertisers use to set a starting bid, and it's simple enough to run in your head:

Max CPC bid = Conversion rate × Average order value × Target ACOS

Say a keyword converts at 10%, your average order value on that ASIN is $40, and your target ACOS is 25%. The math: 0.10 × $40 × 0.25 = $1.00. That's your starting max bid. If the same keyword converted at 6% instead, the bid drops to $0.60 — same product, same price, just a worse-fitting search term.

This formula is a starting point, not a forecast. It assumes you already know the keyword's conversion rate, which for a brand-new keyword you don't. On anything under roughly 50-100 clicks, treat the output as a guess and expect to correct it once real data shows up.

Dynamic Bidding Strategies: Down Only, Up and Down, Fixed

Amazon gives you three ways to let that base bid move once it's set:

• Dynamic bids — down only: Amazon lowers your bid, never raises it, when a click looks less likely to convert. This is the conservative default and the right one for most established campaigns — it keeps your max bid an actual ceiling.
• Dynamic bids — up and down: Amazon can raise your bid (up to 100% for top-of-search, up to 50% elsewhere) when a click looks more likely to convert, and lower it otherwise. It can win more top-of-search volume, but it spends faster, and it's easy to lose track of what you're actually paying versus what you set.
• Fixed bids: No adjustment either way. Reasonable for the first week or two of a brand-new campaign when Amazon has no conversion data to adjust against — after that, it just means leaving efficiency on the table in both directions.

There's no universally correct choice. Down-only suits anything with real sales history. Up-and-down suits deliberate launches where you can absorb a higher ACOS for a few weeks in exchange for visibility.

Placement Modifiers Stack On Top — Here's the Math

Placement modifiers sit on top of whichever bidding strategy you've chosen, and they multiply rather than add. That's where a bid that looked disciplined on paper turns into a number you didn't mean to pay. Take a $1.00 base bid, a dynamic up-and-down strategy, and placement modifiers set separately for top-of-search and product pages:

Multiply the effective bid by your actual daily click volume at top-of-search and you can see how a $1.00 base becomes a real budget decision, not a rounding error. Set placement modifiers first, watch the placement report for at least two weeks, then adjust — not the other way around.

When the Bid Doesn't Work: Common Mistakes and What To Do

Two mistakes account for most of the wasted spend we see in bid setups.

The first is using a blended average order value or conversion rate across an entire ASIN family when the keyword only matches one variant. The formula still produces a bid — it's just the wrong bid. We've made this exact mistake ourselves on a multi-variant listing: one shared bid ceiling that was too high for the slow-converting variant and too low for the fast one.

The second is running dynamic up-and-down bidding with aggressive placement modifiers on a keyword with a thin sample size — chasing top-of-search on a term that's had twenty clicks. The data isn't stable enough yet to justify the multiplier.

When a bid change doesn't produce the result you expected — ACOS climbs, or sales don't move — the fix isn't to guess again. Check what specifically changed, give it enough clicks to be readable, and have a way back to the old number if it's worse. Every proposed change carries three things before it runs: the evidence behind it, a measurement plan, and a rollback trigger. That discipline holds whether you're doing it by hand in a spreadsheet or something else is doing it for you.

Where This Fits

If you're managing bids yourself, the formula and the modifier math above are the whole job — the rest is watching the search term report and adjusting weekly. Dr. PPC, run by Full Circle, does this as a managed service rather than a tool: it reads the account, proposes bid and modifier changes against your actual economics with the evidence and rollback plan attached, and runs at whatever autonomy level you set. It costs $300 a month plus 3% of ad spend, capped, with the first 30 days free, and includes Orbit — the analytics and tracking suite — at no extra charge. Full Circle has managed over $500M in revenue across 100+ brands. None of that changes the math in this article; it changes who's doing it.