Amazon PPC Bid Optimization Rules: A Weekly Rulebook You Can Copy
The best Amazon PPC bid optimization rules are simple: lower bids on keywords running above your target ACoS, raise bids on keywords running well below it, cut or negate targets that spend past a click threshold with no orders, and leave alone anything without enough data. The key is setting clear thresholds, applying them on a fixed schedule, and limiting the size of each change so you never swing an account too hard.
Below is a rulebook you can copy into a spreadsheet or a rules engine. Every number is a starting point. Tune it to your margins and conversion rates.
Before you start: three numbers you need
- Target ACoS per product. Start from break-even ACoS and decide how much profit you want from ad sales. The break-even ACoS calculator helps.
- Clicks per order (CPO). This is 1 divided by your conversion rate. At a 10% conversion rate, CPO is 10. Many rules below are expressed as multiples of CPO, so they scale with your product.
- A lookback window. Use 14 to 30 days, ending 2 to 3 days ago so Sponsored Products attribution has mostly settled.
The rulebook
| Rule | Condition | Action |
|---|---|---|
| 1. Bleeder | Clicks >= 2 x CPO and 0 orders | Lower bid 30-50%, or negate the search term if irrelevant |
| 2. Over target | Orders >= 2 and ACoS > target x 1.2 | New bid = CPC x (target / ACoS), max 25% cut per week |
| 3. Slightly over | Orders >= 2 and ACoS between target and target x 1.2 | Lower bid 5-10% |
| 4. Under target | Orders >= 3 and ACoS < target x 0.7 | Raise bid 10-20%, max 25% per week |
| 5. No impressions | Impressions < 100 in 14 days | Raise bid 15%, up to your calculated max bid |
| 6. Impressions, no clicks | Impressions > 1,000 and CTR far below campaign average | Check relevance and main image before raising bids |
| 7. Not enough data | Clicks < 1 x CPO | Do nothing yet |
Why the thresholds are tied to clicks per order
A fixed rule like 'negate after 10 clicks' is too harsh for a product that converts at 5% (20 clicks per order) and too lenient for one that converts at 25% (4 clicks per order). Expressing thresholds as multiples of CPO keeps the rule fair across your catalog.
Why changes are capped
Amazon's auction responds to bids nonlinearly. A 40% bid cut can drop a keyword from page one to nowhere, and a 40% raise can double spend. Capping weekly changes at about 25% lets you walk toward the right bid while watching what happens.
Worked example: applying the rules to five keywords
Product: $30 average order value, 10% conversion rate (CPO = 10), target ACoS 30%. Lookback: 30 days ending 3 days ago.
| Keyword | Clicks | Orders | CPC | ACoS | Rule | New bid |
|---|---|---|---|---|---|---|
| bamboo cutting board set | 120 | 14 | $0.95 | 27% | None (on target) | Keep $1.00 |
| large cutting board | 80 | 4 | $1.10 | 73% | 2. Over target | $0.83 (25% cap) |
| cutting board with juice groove | 60 | 9 | $0.70 | 16% | 4. Under target | $0.84 (+20%) |
| wood serving tray | 25 | 0 | $0.80 | n/a | 1. Bleeder | $0.48, review terms |
| butcher block board | 6 | 0 | $0.90 | n/a | 7. Not enough data | No change |
For 'large cutting board', the formula gives $1.10 x (30 / 73) = $0.45, a 59% cut. The weekly cap limits it to 25%, so the bid moves to about $0.83 this week and gets re-evaluated next week.
Rules for search terms, not just keywords
Bid rules adjust keywords and targets. Search term rules decide what to add and what to block. Run them alongside your bid rules:
- Harvest: search term with 2 or more orders and ACoS at or below target in an auto, broad, or phrase campaign. Add it as exact match in a manual campaign and as negative exact in the source. More detail in how to harvest keywords from auto campaigns.
- Negate: search term with clicks >= 2 x CPO and 0 orders, or any clearly irrelevant term. Add as negative exact (or negative phrase for a whole irrelevant theme).
- Watch: search term with 1 order and high ACoS. Leave it running until it reaches the harvest or negate threshold.
How often to run the rules
- Weekly: rules 1 to 4 on your top-spending campaigns.
- Every two weeks: all rules on the rest of the account.
- Monthly: placement adjustments and bidding strategy review. See Amazon placement bid adjustments.
- Before and after peak events: temporarily loosen raise limits and tighten review cadence.
Running everything daily usually does more harm than good for smaller accounts. Yesterday's sales are not fully attributed yet, so daily rules react to noise. If your account is large enough to see dozens of orders per keyword per week, more frequent changes can make sense.
Guardrails every rulebook needs
- A minimum bid floor, such as $0.15 to $0.25, so a keyword is never cut to the point of zero impressions without you deciding to pause it.
- A maximum bid ceiling per product, based on your calculated max bid, so raises never run away.
- A cooldown: do not change the same keyword more than once in 7 days.
- Exclusions for branded keywords and launch campaigns, which usually have their own targets.
- A change log with the old bid, new bid, date, and reason, so you can undo a bad batch.
Adapting the rulebook to your account
The thresholds above assume a typical private label product with a conversion rate between 8% and 15%. Adjust them when your situation is different. For low-priced products (under about $15), margins are thin and CPCs eat a large share of each sale, so tighten rule 2 to trigger at target x 1.1 and raise the bleeder threshold only slightly. For high-priced products, orders are rarer, so lengthen the lookback to 60 days and lower the order minimums in rules 2 and 4 to one or two orders. For launch campaigns, use a higher temporary target ACoS and skip rule 2 entirely for the first few weeks, since the goal is sales velocity rather than efficiency.
Whatever you change, write the rulebook down. A written rulebook makes bid management consistent between team members, makes it easy to hand off to a freelancer or a tool, and gives you something concrete to review when results shift.
AdSpark's rules engine runs rules like these on your schedule, alongside a continuous bid optimizer, and logs every change so you can reverse it.
Start your 14-day free trialWhen rules are not enough
Rules are transparent and predictable, but they only know what you tell them. They struggle with seasonality, sudden competitor moves, and keywords with thin data. That is the main argument for algorithmic or AI-driven bidding, which can weigh more signals at once. Many sellers combine both: AI or algorithmic bidding for the bulk of keywords and hard rules as guardrails. We compare the approaches in rule-based vs AI bidding on Amazon, and you can see how this fits a broader routine in our weekly Amazon PPC optimization routine.
FAQ
How many clicks before I lower a bid on Amazon?
A useful rule is about two times your clicks per order with no sales. If your conversion rate is 10%, that is roughly 20 clicks. For keywords that do convert, wait for at least two orders before judging ACoS.
How much should I change a bid at once?
Keep single changes to about 10 to 25%. Larger swings can cause big jumps in impressions and spend that are hard to read.
Should I use Amazon's built-in rules?
Amazon offers some rule features, such as budget rules and, for eligible campaigns, rule-based bidding with a ROAS guardrail. They are worth testing, but they do not replace a full keyword-level rulebook.
Do the same rules work for Sponsored Brands?
The logic works, but Sponsored Brands uses a 14-day attribution window and often has lower conversion rates on headline ads, so use longer lookbacks and adjust thresholds.