Free Tool — No Signup

Bid Optimizer

“Decrease by 10% if ACOS is over target” is a guess wearing a rule’s clothing. At 500% ACOS it takes months to arrive, and running it twice in a week compounds an error you never chose. This calculates the exact bid instead — from revenue per click, in one move — and tells you which of its own formulas are safe to run again tomorrow.

The whole tool, in one line Target CPC = RPC × Target ACOS
Four categories, four formulasCalculated from CPC, never from bid Frequency-safety warningsData hierarchy for low-click keywords Grace range — leaves working bids aloneUpload-ready bulk file out

What This Tool Does

Sorts every keyword and product target into one of four categories, then applies the formula that category needs.

📉
High ACOS → RPC formulaCalculates the exact bid to hit target in one move, not ten percent at a time
💸
Spend, no sales → CPA projectionBid steps down on its own as non-converting clicks pile up — no arbitrary $20 cutoff
👀
Low visibility → measured step-upAnd it separates “never served” from “served but nobody clicked”, which is not a bid problem
📈
Low ACOS → scale to the ceilingSteps up toward the maximum you can afford, and stops there
🔒
Frequency safetyTwo of the four formulas compound if you re-run them. The tool knows which and locks them
🏷
Ranked by dollars, not percentages96% ACOS on $450 outranks 400% ACOS on $12, every time

📋 How To Use

The standard Sponsored Products bulk file — the same export the other tools take.

Open Bulk OperationsCampaign Manager → Sponsored Ads → Bulk operations
Choose your date range deliberately30–90 days for routine work, 7–14 for a targeted fix. Start it after your last big change
Include performance dataWithout clicks and sales there is no RPC, and no tool can invent one
Drop the file belowRead locally — nothing is uploaded anywhere
Set your target ACOS and modeThen export the bulk file and upload it back
💰

Drop Your Sponsored Products Bulk File Here

Bulk operations export with performance data included — .xlsx or .csv

🔒 Never leaves your browser✅ No row limit📝 No signup
One thing worth getting right before you export. Your date range is a trade between confidence and relevance — longer is more reliable, shorter reflects current conditions. Begin it after your most recent significant change: a deal, a stockout, a price move, or your last bid optimization. You are not optimising for what happened; you are optimising for what is about to happen.

Reading your bulk file

Parsing rows…

📄
The Real Problem

“Decrease by 10%” is a guess with a rule’s reputation.

It is the industry default, and almost every part of it is arbitrary.

1

Why ten percent?

Nobody can tell you. It is not derived from your margin, your conversion rate or your revenue per click. It is a number that sounds cautious, applied to accounts that have nothing in common.

2

At 500% ACOS it takes months

Cutting ten percent at a time from five times your target is a long walk. The maths can tell you the right bid immediately, and every week spent walking there is money spent at the wrong bid.

3

It compounds without telling you

Percentage rules step from the current bid, so running one twice in a week applies it twice. Ten percent a day halves a bid inside a week on data that never changed.

4

It is usually applied to the wrong number

Most people apply the percentage to the bid. The bid is a status — whatever it happens to be set to. The CPC is what actually happened. Calculate from the data, not the setting.

What Makes This Different

One formula, applied honestly to four situations

Your cost per click should be a calculated percentage of your revenue per click. Everything else is that idea in context.

It calculates, it does not nudge

Target CPC = RPC × Target ACOS. A keyword making $100 on 10 clicks has an RPC of $10, so at a 30% target its bid is $3.00 — reached in one move rather than eleven weekly cuts. And because the formula recalculates from the file instead of adjusting whatever the bid currently is, the same data always produces the same answer.

It knows which of its own formulas are safe to repeat

Categories 1 and 2 recalculate from data: run them hourly if you like, the output only moves when the data does. Categories 3 and 4 step up from the current bid, so running them twice applies them twice. The tool asks when you last optimized and locks the compounding half until a week has passed. No other free tool draws this distinction, and it is the one most likely to cost you money.

Non-converting keywords get maths, not a $20 rule

Archiving everything with $20 spend and no sales treats a $200 product like a $15 one. Instead: Target CPC = Target ACOS × (AOV ÷ (clicks + 1÷CVR)). The bid steps down on its own as non-converting clicks accumulate — $0.36 at fifteen clicks, $0.30 at twenty — so the target keeps trading until it has spent about one target CPA, then prices itself out.

A hierarchy instead of a fabricated number

One click and one sale on a $20 product implies a $6.00 bid — a bet that a 100% conversion rate holds at thirty times the current CPC. So low-data targets inherit from the ad group, then the campaign, then the account, and every row is labelled with the level it used. Borrowed numbers are marked as borrowed.

A grace range, because churn has a cost

Anything within 10% of your target is left alone and counted. Constantly adjusting keywords that are already working buys volatility, and you are not only paid to optimize — you are paid to know when not to. Most tools generate a recommendation for every row, because a longer list looks like more value.

Ranked by dollars, and it shows the caps

A 96% ACOS on $450 of spend outranks 400% ACOS on $12, always. And when the correct bid is further away than one cycle’s cap allows, the tool shows both numbers and how many cycles it will take — rather than quietly capping and letting you believe you are done.

Questions

Frequently Asked Questions

Which file do I need?
The Sponsored Products bulk file: Campaign Manager → Bulk operations, with performance data included. The tool reads Keyword and Product Targeting rows — their bids, clicks, spend, sales and orders — plus ad group and campaign rows for the data hierarchy. A settings-only export has bids but no performance, and there is no revenue per click to calculate from.
Why calculate from CPC instead of from my bid?
Because the bid is a status and the CPC is data. Your bid is whatever it happens to be set to right now; your CPC is what you actually paid over the date range you chose. The distinction matters most exactly when it is easiest to get wrong: if someone dropped every bid to $0.05 yesterday, your 30-day ACOS still reflects the old, higher CPCs. Cutting again because "ACOS is still high" would collapse the account. Every formula here starts from the CPC.
What is the difference between macro and micro mode?
Macro is routine maintenance: the whole account, a 30–90 day window, caps of ±15%, run weekly. Small changes across many keywords, low risk. Micro is targeted intervention: specific campaigns, a 7–14 day window, caps of ±50%, run as needed. Higher risk per keyword but applied to a small pocket of spend. Most people only ever do one of the two; the framework needs both.
Why does it sometimes refuse to give me a bid?
Four reasons, and each is deliberate. The target is inside the grace range, so it is working and changing it buys volatility. It has impressions but no clicks, which is a relevance problem — a higher bid buys more impressions of something shoppers are already declining. It is already at the affordable ceiling, so stepping up is choosing to lose money. Or there is no conversion data anywhere in the hierarchy, in which case there is no honest number to give.
Why are some recommendations locked?
Because two of the four formulas compound. Categories 1 and 2 recalculate from your data, so the same file always yields the same bid — run them as often as you like. Categories 3 and 4 step up from your current bid, so running them daily multiplies: 10% a day doubles a bid within a week on data that never justified it. Tell the tool when you last optimized and it locks the compounding half for seven days. Set the date to something older and they unlock.
My CPC is higher than my bid. How?
Placement modifiers. Your effective CPC is base bid × placement multiplier, so a +50% top-of-search modifier on a $1.00 bid can produce a $1.50 click, and dynamic up-and-down bidding can push it further. This is why bids and placements have to be managed in the same session — the tool flags the gap when it sees it, and the Placement Analyzer handles the other half from the same file.
How often should I actually run this?
Once or twice a week. Not daily. On a 30-day lookback one extra day barely moves the averages, so you make the same decision you made yesterday; step-ups compound; and if ACOS spikes a fortnight later across thousands of changed keywords, you cannot trace which change caused it. There is also an art to doing nothing: if sales are growing, ACOS is on target and trends are positive, extending budgets and leaving bids alone can be the better call.
Can I upload the export straight back to Amazon?
Yes. It contains only the rows whose bid changed, with Operation set to Update, the Bid column carrying the new value, and every other field copied verbatim from your own file. IDs are written as text so 15-digit values survive Excel. Review it before submitting, as with any bulk upload — then send it through Bulk operations.
Should I use Amazon's suggested bid instead?
Only as a sanity check. Amazon's suggestion reflects auction market price — it knows nothing about your margins, your target ACOS or your product economics. A $0.75 suggested bid might be $2.00 more than you can profitably afford, or $3.00 less than you need to win the placements that convert. Your own revenue per click is the only number that accounts for your business.
Is my data uploaded anywhere?
No. The file is read by JavaScript in your browser and discarded when you close the tab. The only thing stored is the "last optimized" date, kept in this browser's local storage so the frequency lock survives a refresh. There is no server and no account.
Go Deeper

Related Guides & Tools

Want every bid in your account done properly?

The formulas handle eighty percent of the work. The rest — date range selection, when to hold, when to push — is judgment, and that is what you are actually hiring.