How I Cut Amazon ACOS From 63% to 40% and Doubled Sales Year Over Year 

Quick answer: I cut ACOS from 63% to 40% on one Amazon account in about two months by auditing wasted PPC spend, setting up keyword harvesting rules, adding ACOS-based bid rules, and dayparting my ads to stop paying for dead clicks in the middle of the night. On a second account, the same advertising approach plus refreshed images and listings doubled sales year over year without adding a single new product. Every tool I used is inside Helium 10, and I walk through the whole thing below. 

I run real Amazon accounts with real money and real risk. Not demo accounts, not someone else’s case study. So when I tell you these numbers, they came out of my own P&L. In this post I am pulling back the curtain on two of those accounts: one I am bringing back from the brink of death, and one I already doubled year over year. The thread connecting both is advertising, because advertising is the biggest lever you have. 

How do you bring an Amazon brand back from low performance? 

One of these accounts is a brand a couple built around kids’ learning products and Bible verse mapping. They got big on TikTok, hit around two to three million dollars across TikTok and Amazon, and then the bottom fell out. Their TikTok Shop got shut down for a while, shipping errors wrecked their metrics, and since TikTok was driving their Amazon sales, both platforms went to almost nothing. 

From January to May of last year they did 1.5 million dollars across both platforms, with a million of that on TikTok alone. This year, same five months, TikTok fell to about 30,000 dollars. The account did not just dip. It fell off a cliff. 

My goal was simple to say and hard to do: get Amazon working on its own so this brand is not 90% dependent on TikTok ever again. That starts with cleaning up the listings and the advertising, then rebuilding the keyword positions they lost while they were invisible. 

Before I touched ads, I made sure the listings deserved traffic. New main images and A+ content from AMZ OneStep, then I ran the main image options through Helium 10 audience polling before choosing. The winner was not the image I personally liked. Real buyers picked the one that showed the inside of the workbook and actual handwriting, because they had never seen the product before and needed to understand how it works. As the brand owner, you already know your product, so your favorite image is often the wrong one for the person seeing it for the first time. 

Reclaim Your Time

Elevate Your Brand Performance

Helium 10’s Diamond Plan automates the tasks eating your day so you can focus on decisions that actually move the needle across Amazon, Walmart, and TikTok Shop.

Sign Up Today

How did you cut ACOS from 63% to 40%? 

When I got into the account, advertising was out of control. 63% ACOS, 32% TACOS, 20,000 dollars of spend in two months for 37,000 dollars of sales. Here is the exact order I worked in. 

1. Audit the wasted spend first. I went to the search term level and asked one question: where did I get a lot of clicks and zero orders? There were 43 search terms bleeding money, including one with 69 clicks and no sales. If 69 clicks have not produced a sale, the odds the 70th does are low. I negative-matched all of them out of the auto and broad campaigns to stop the bleeding immediately. Helium 10 now has an ads audit that surfaces this automatically, but the point is that if you never check, you will always find 63% ACOS waiting for you. Amazon is not going to refuse your money. 

2. Split bloated campaigns and set up keyword harvesting. They had campaigns with 400 targets in them, which is not workable unless you have a huge daily budget. I split them into tight, focused campaigns, then built harvesting rules so all the campaigns for one product talk to each other and do not duplicate targets. A keyword that proves itself in an auto or broad campaign gets funneled into the right exact or product-targeting campaign automatically once it hits two orders under 100% ACOS. There is even a setting that stops it from suggesting a search term that already exists with the same match type, so you never bid against yourself. 

3. Add bid rules that manage ACOS while you sleep. I put every ACOS-based campaign into one rule group. Above 200% ACOS, pause it, that one is beyond saving. Between 100% and 200%, bring it down toward 60%. Between 75% and 100%, bring it down toward 45%. All the way down to my target of about 20%. I also have a rule that raises bids when ACOS drops too low, because sitting at 10% ACOS usually means I am leaving sales on the table. And one rule almost nobody talks about: if a target got paused for no sales but a delayed order comes in through the attribution window, the rule re-enables it and gives it another shot. 

4. Daypart to kill dead-of-night spend. They were spending around 300 dollars a night between midnight and 4 a.m. on people browsing in bed who never buy. You cannot do this inside Seller Central on its own, but with dayparting I set bids to drop 40% in that window. I did not turn ads off completely, because I still want brand visibility. Spend in that 2 a.m. window went from 300 dollars to about 45. 

5. Rebuild the keywords that used to convert. Using search query performance and brand analytics, I found keywords that drove sales before the collapse but had since dropped out of the top ten organically and in sponsored. Those went into ranking campaigns designed to hold top three or four in sponsored results, with a rule that raises the bid automatically when my tracked rank slips below position five. If a keyword converted well for me before, I just need to be seen again. 

The result: in two months, March to April, ACOS dropped from 63% to 40%, TACOS from 32% to 18%. I spent 10,000 dollars and got 24,000 dollars of PPC sales, where before they paid more than double that for less. No single change did it. The audit, the harvesting, the bid rules, the dayparting, and the ranking campaigns all worked together. Amazon is now carrying this account on its own, and the profit margin is still 30%. 

Can you double Amazon sales without launching new products? 

Yes, and I did. The second account is an InMotion Hemp cream and roll-on brand. Same five-month window, year over year, gross sales went from 50,000 dollars to 106,000 dollars, and we have not even added the new products we have planned. 

Two things drove it. First, refreshed images, copy, and A+ content, again validated with audience polling. That lifted the conversion rate from 20.39% to 21.86%. That sounds tiny, but on this account it was worth over 7,000 dollars in extra sales by itself. Bump a weaker listing 4% and you are looking at a 25,000 dollar swing. 

Second, the same advertising work from the first account. More keyword coverage from search query performance and brand analytics meant page views nearly doubled from 7,600 to 13,000. More visibility plus a better conversion rate is a two-prong approach, and it is why the account doubled. His ACOS came down from 51% to 47%, which is fine here because strong subscribe and save means these buyers come back on replenishment. 

Which Helium 10 plan do you need to do this yourself? 

Everything above uses tools any Helium 10 member can run. There was no special access. The rules-based advertising, dayparting, and historical keyword data that did the heavy lifting on these turnarounds live in the Diamond Plan. If you have been fighting a high ACOS or watching a brand slide, this is the plan that gives you the controls to fix it. 

Key takeaways 

• ACOS this high almost always means wasted spend you have never audited. Start by finding clicks with zero orders and negative-matching them. 

• Keyword harvesting and ACOS-based bid rules keep your ads clean without daily manual work. 

• Dayparting stops you from paying premium prices for clicks that never convert. 

• Better images and listings compound with better advertising. A 1% conversion lift funded five figures of extra sales. 

You can bring a near-dead brand back and double a healthy one using the same Helium 10 solutions.