Most accounts can cut Amazon ACOS by a third in ninety days without losing a single position. The trick is knowing which spend is buying rank and which is only renting clicks, and almost nobody separates the two before they start cutting.
Amazon PPC
You can cut Amazon ACOS by roughly a third in ninety days without losing a single position, provided you separate the spend that is buying rank from the spend that is only renting clicks. Almost nobody does that before they start cutting, which is why the usual result is a healthier looking ad account attached to a smaller business.
Every seller who has ever been told to bring ACOS down has had the same experience. They cut bids, ACOS improves for a fortnight, and then organic sales quietly fall off a cliff. Two months later total revenue is lower than when they started.
What is on this page
The connection the console never shows you
Amazon’s organic ranking rewards sales velocity on a search term. Paid clicks that convert contribute to that velocity. So when you cut the bid on a term you rank first for, you do not just lose the paid sales. You lose the velocity holding the organic position, and the organic sales follow it down with a lag of a few weeks.
Nothing in the advertising console shows you this, because the console reports on advertising. Your ACOS chart will look like a success story for a month. The damage appears in total sales, which is why total advertising cost of sale is the more honest measure while any of this work is happening, for the reasons we set out in ACOS vs TACOS.
It cuts the other way too, and this is the useful part. Some spend is not holding anything up. Cut that and nothing at all happens except your costs falling.
Split the spend into two piles first
Before you touch a bid, take every search term from the last ninety days and sort it into two groups.
Spend that is buying rank. Terms where you are climbing, or holding a position you could not yet hold organically. This is an investment with a return that compounds, and it should have an end date attached.
Spend that is renting clicks. Terms where you already rank first organically and are paying to appear above yourself, or terms that have taken money for months without producing a sale. There is no compounding here and no end date.
To sort them you need two numbers per term: your organic position, and the share of that term’s sales coming from ads. High organic position with a small paid share means the paid support is largely redundant. Page two organic position with paid carrying the term means cutting will cost you.
The four cuts that never cost you rank
One. Negate the search terms with clicks and no orders. Set the threshold from your actual conversion rate rather than a round number. If you convert at 12%, a term with thirty clicks and no orders is very unlikely to be a fluke. Negate it as exact rather than phrase, so you do not accidentally take out a variant that does work. The full method is in the search term report guide.
Two. Find the budgets capping before the day ends. A campaign that spends out at eleven in the morning is not efficient, it is rationed. It buys whatever the morning offered rather than the best of the day. Raising that budget can lower ACOS, which sounds backwards until you see which hours it was missing.
Three. Fix the placement modifiers. Many accounts pay a premium for product page placements that convert half as well as search. Pull the placement report and set the modifiers to match what the data says rather than what seemed sensible a year ago. The mechanics are in bid strategies and placements.
Four. Reduce paid support on terms you own organically. Not to zero, and not overnight. Take the bid down in steps of 15%, wait a fortnight between steps, and watch total sales rather than ad sales. If total holds, the paid spend was redundant and you have just found margin.
Cutting bids across the board is not a strategy. It is the absence of one, applied evenly. It removes your worst spend and your best spend in the same proportion, and the best spend is the part holding your rank up.
Want to know which third of your budget is safe to cut?
Send us the last ninety days. We will mark up your own report with what we would cut first, free.
The two cuts that always backfire
Blanket bid reductions. Taking 20% off every bid in the account feels decisive and treats your best performing keyword exactly like your worst. You lose impression share on the terms that were paying for themselves, and you keep the terms that were not, because they were cheap.
Pausing campaigns. It stops the spend immediately, which is the appeal. It also throws away the campaign history and the performance signals that took months to build, and restarting later means rebuilding both from a worse starting point. Reduce and negate instead. Leave the structure standing.
The chart is the reason this keeps happening. The ratio improves inside a fortnight while the organic damage takes six to twelve weeks to show, and by then it looks like an unrelated problem. Plenty of accounts do a second round of cutting in week six, on the strength of how well the first round appeared to go, and the two sets of damage then arrive together.
A ninety day plan you can follow
Weeks one and two: measure, cut nothing. Pull ninety days of search term data. Note your organic position on the terms carrying the most spend. Sort every term into the two piles. This is the whole job, and it is the step people skip.
Weeks three and four: remove the waste. Negate the zero order terms past your threshold. Fix the placement modifiers. Raise the budgets on campaigns capping early. None of this touches spend that is buying rank, so nothing should move except costs.
Weeks five to eight: reduce carefully. Start stepping bids down on terms you own organically, 15% at a time, a fortnight apart, watching total sales. Stop the moment total sales move.
Weeks nine to twelve: check the lag. Now the organic consequences of week three are visible. Compare organic sessions and organic sales against your baseline. If anything slipped, put the bid back on that specific term rather than reversing everything.
By the end you should be a third lower on ACOS with total sales flat or better. If total sales fell, the cut went too deep somewhere, and the record of what you changed and when will tell you exactly where.
They cut bids, ACOS improves for a fortnight, and then organic sales quietly fall off a cliff.
Rather have somebody run this properly than guess at it?
This is the first ninety days on every account we take over. Twenty minutes tells you if it fits.
What you should actually be aiming at
Your break even ACOS is your contribution margin. Above it you are paying for growth, below it you are taking profit. Which side you should be on depends on what the product is doing.
| Product stage | Sensible ACOS | What you are buying |
|---|---|---|
| New, no rank, no reviews | Well above break even | Velocity, deliberately, for a set period |
| Climbing, gaining position | Around break even | The last stretch of rank |
| Ranking, steady sales | Comfortably below | Defence and incremental sales |
| Mature, first organically | Lowest in the account | Very little, so pay accordingly |
Our own book runs a 17.84% lifetime ACOS across categories with very different margins, and that figure is an average of products sitting in all four rows. Treating it as a target for a single product would be a mistake, which is the same argument as there is no correct TACOS.
What to watch while you do it
Four numbers, weekly, on one page.
Total sales, not ad sales. This is the number that tells you whether a cut was free or expensive. Ad sales falling while total sales hold is exactly what a good cut looks like.
Organic sessions. Your early warning. It moves before organic sales do.
Impression share on your core terms. If it collapses after a bid step, you cut past the point where you were competitive.
Total advertising cost of sale. The honest summary of the whole exercise, because it counts the organic sales your paid spend was supporting.
Keep a note of every change with its date. In week ten, when something moves, that note is the difference between knowing what caused it and guessing.
Send us the reports and we will tell you where the safe third is
Free, no card, written answer in three working days, and yours to keep either way.
Frequently asked questions
A third inside ninety days is realistic for most accounts we take over, because that much of the budget is usually sitting on spend that produces nothing. Doing it in a fortnight is not, because the only lever fast enough is cutting bids across the board, and that takes your rank with it.
It can, and this is the mechanism nobody sees in the advertising console. Organic rank rewards sales velocity on a search term, and paid sales contribute to that velocity. Cut the bid on a term you rank for and you lose the paid sales plus the velocity holding the organic position, with the organic sales following a few weeks later.
Spend that buys rank is on terms where you are climbing or holding a position you could not yet hold organically. It compounds and it has an end date. Spend that rents clicks is on terms where you already rank first organically, or terms that have taken money for months without a sale. It compounds into nothing.
Almost never. Pausing throws away campaign history and performance signals along with the spend, and restarting later means rebuilding both. Reduce bids on the specific terms that deserve it, or negate what does not convert, and leave the campaign structure standing.
Check your organic position on that term, then look at what share of its total sales come from ads. If you rank in the top three organically and paid sales are a small share, the paid support is largely redundant. If you rank on page two and paid is carrying the term, cutting it will cost you.
Below your break even ACOS, which is your contribution margin, adjusted for what you are trying to do. A product buying rank should run above it deliberately for a defined period. A mature product should sit well below. There is no universal figure and anybody quoting one has not asked about your margins.
Advertising spend responds within days. The organic consequence, good or bad, takes three to six weeks to appear. That lag is why so many people conclude a cut worked, keep cutting, and then get hit by the delayed damage from the first round.
No. An unusually low ACOS on a growing product often means you are underinvesting while a competitor buys the rank you could have had. It is only good news when you already hold the positions that matter and are defending them cheaply.
Send the right reports. Get a straight answer back.
We read the account properly and write back with where the money is going, what it is buying, and the three things we would change first. You keep it either way.
No obligation · no card · a person replies, usually the same working day