Most Sponsored Products accounts I audit have the same problem: bids were set once, by feel, and never tied back to what the product actually earns. The campaigns "work" in the sense that they generate sales, but nobody can say whether those sales make money.
This guide covers how I set and manage bids so that ad spend is pointed at profit.
Start with your break-even ACoS
ACoS (advertising cost of sale) is ad spend divided by ad sales. It only means something once you compare it with your margin.
Break-even ACoS = your profit margin before ad spend.
If a product sells for $40 and, after product cost, Amazon referral fees, FBA fees and shipping in, you keep $12, your margin before ads is 30%. At a 30% ACoS you break even on ad-driven sales. Below 30% you make money on every ad sale; above it you're paying to sell.
| Selling price | Profit before ads | Break-even ACoS | Target ACoS (profit) |
|---|---|---|---|
| $25 | $6.25 | 25% | 15–20% |
| $40 | $12.00 | 30% | 20–25% |
| $60 | $21.00 | 35% | 25–30% |
Work this out per product, not per account. A blended account ACoS hides the products that are losing money.
Use TACoS to judge the whole picture
ACoS only looks at ad sales. TACoS (total advertising cost of sale) divides ad spend by all sales, organic included. It tells you whether advertising is building the business or just renting it.
- ACoS stable, TACoS falling: ads are lifting organic rank. This is what you want.
- ACoS stable, TACoS rising: you're becoming more dependent on paid traffic.
When I'm launching or ranking a product, I'll accept an ACoS above break-even for a period, as long as TACoS is heading down. That's a deliberate investment, not an accident.
Turning a target ACoS into a bid
You can work backwards from your target to a sensible bid for any keyword:
Bid = average order value × conversion rate × target ACoS
For a $40 product converting at 12% with a 25% target ACoS: $40 × 0.12 × 0.25 = $1.20. If the keyword's suggested bid is $2.50, it won't be profitable at today's conversion rate. Either the listing needs to convert better or the keyword isn't worth chasing yet.
Recalculate as data comes in. Two weeks of real conversion data on a keyword beats any estimate.
Dynamic or fixed bidding?
Amazon gives you three bidding strategies at campaign level:
- Dynamic - down only: Amazon lowers your bid when a click looks less likely to convert. This is my default for most campaigns. It protects spend without inflating it.
- Dynamic - up and down: Amazon can raise bids by up to 100% for top-of-search (and up to 50% elsewhere) when a conversion looks likely. Useful on proven, high-converting exact-match keywords. Risky everywhere else.
- Fixed bids: Amazon uses your bid as set. Good for tests where you want clean data, and for tightly controlled ranking pushes.
Placement modifiers (top of search, product pages, rest of search) are often more powerful than the base bid. Check your placement report: if top of search converts at twice the rate of product pages, bid more there and less elsewhere.
Auto and manual campaigns work as a system
Auto and manual targeting aren't alternatives. They do different jobs.
Automatic campaigns are for discovery. Amazon matches your ads to searches and products based on your listing. They'll find search terms you never thought of, and they'll also find plenty you don't want.
Manual campaigns are for control. Keyword targeting (exact, phrase, broad) and product targeting (ASINs and categories) let you set a bid for each target based on its own performance.
The workflow that ties them together:
- Run auto and broad campaigns to collect search-term data.
- Each week, review the search-term report.
- Move converting terms into exact-match campaigns with their own bids.
- Add those terms as negative exact in the discovery campaigns, so you don't bid against yourself.
- Add terms that spend without converting as negatives.
That weekly loop is the core of good PPC management. The account gets more efficient over time instead of drifting.
Rules of thumb that save money
- Give a keyword enough clicks before judging it. If your conversion rate is 10%, a keyword with 8 clicks and no sales hasn't proven anything yet. Around 20–30 clicks without a sale is a reasonable point to cut or reduce.
- Don't duplicate targets across campaigns. The same exact keyword in three campaigns means you're competing with yourself and the data is split three ways.
- Separate branded and non-branded terms. Branded searches convert well and make ACoS look better than it is.
- Fix the listing before raising bids. If a keyword gets clicks but doesn't convert, more traffic won't fix it.
The bottom line
Bid from your margins, measure with TACoS as well as ACoS, and run auto and manual campaigns as one loop. Do that every week and Sponsored Products becomes a profit lever rather than a cost you tolerate.
This is exactly the weekly routine behind our Amazon PPC management: AI tools work through the search-term reports, and I review every change before it goes live.