← Back to case studies
42% less ad spend, no drop in sales, 63% more profit
The client: a 9-figure omnichannel supplement brand in the US, selling 8 figures a year on Amazon, with ad spend averaging over $1M a month before the cut.
with $450k+ a month cut from a $1M+ ad budget
of steady sales after the cut, with average monthly sales up 2%
ad spend on the hero product, with its organic sales flat
Monthly averages, seven months before the cut against four months after. Sales include organic.
Other results over the same period:
TACOS down from 33.5% to 19.0%
ACOS down from 96% to 60%
Organic orders up 7%
A new product line grew sales 2.7x in five months while its TACOS halved
Summary
The brand runs heavy direct-to-consumer marketing alongside Amazon. We reset its Amazon advertising around one question: how much of the ad spend was creating sales, and how much was paying for sales the brand would win anyway?
•
We cut ad spend by 42%, in steps, over five weeks
•
Total sales held and organic orders rose 7%
•
Organic sales on the hero product held flat while its ad spend fell 59%
•
Keyword-level organic rankings held steady across the product range, tracked throughout
•
Monthly profit rose 63%
•
Every step was measured on the whole account, not just ad metrics, and reported weekly
The starting point: strong demand, expensive ads
The brand creates much of its own demand. Its direct-to-consumer advertising sends a steady stream of shoppers to Amazon already searching for the brand by name. Amazon ads were set up to push for growth on top of that.
In the six months before the cut:
•
Ad sales roughly matched ad spend, with ACOS between 94% and 100%
•
Ads cost 32-36% of total sales (TACOS)
A strategy review with the founders set Amazon’s job: convert the demand the brand already creates, at a profit, and keep growth spend only where it shows a return.
The thesis: a large share of non-branded spend was buying sales the brand would win anyway
Two things pointed to an account that could spend far less without losing sales:
•
Branded demand. Shoppers searching the brand name have already chosen the brand. Branded ads defend demand that direct-to-consumer advertising has already paid for
•
Organic strength on the hero product. The hero product ranked well organically on its core category searches. Part of the paid traffic there was reaching shoppers who would have found it anyway
The founders wanted to cut spend. We defined the strategy for how to do it without losing sales:
1.
Cut non-branded ad spend in stages, not in one step
2.
Cut deepest on the broadest searches, such as ingredient searches, where shoppers were furthest from choosing a product
3.
Cut less on the specific searches that support the hero product’s organic rank
4.
Measure every week whether sales, organic orders and share held
The main risk was clear from the start: organic rank can slip weeks after a cut, once the lost paid sales feed through. So we kept measuring for four months after the cut began.
How we made the cut: five weekly steps, deepest where clicks paid back least
Weekly ad spend came down 52% over five weekly steps. Two phases set the pace:
•
Phase 1: TACOS target lowered from 30% to 25%
•
Phase 2, three weeks later: deeper non-branded cuts, with ACOS goals of 120% for non-branded keywords and 110% for competitor targets
•
Automated bid rules moved each search toward its goal. Spend came out search by search, based on efficiency, with no blanket budget cuts
Before the cut, the main non-branded searches ran at 170-250% ACOS, far above break-even. Every one of them had room to come down, but not by the same amount:
Search type
Ad spend change
ACOS before
ACOS after
Broad ingredient search
-84%
213%
116%
Largest category search
-73%
210%
124%
Second category search
-69%
172%
99%
Specific product search A
-62%
168%
130%
Specific product search B
-51%
250%
131%
Five weeks before the cut against five weeks after it settled.
•
The broad ingredient search took the deepest cut
•
The specific searches, closest to the hero product, kept the most spend
•
Branded keyword spend came down least of any segment we cut, to keep converting the demand the brand already pays to create
How we measured it: seven numbers, reported every week
Ad metrics alone could not answer the question. A spend cut always makes ACOS look better. The test was whether the business kept its sales, so we measured the whole account.
Metric
The question it answers
Total sales, including organic
Did revenue hold?
Organic orders
Did organic demand hold once paid traffic came down?
Organic keyword rankings
Did the products keep their organic positions on the searches that matter?
Purchase share on core non-branded searches
Are we holding our place on the searches that drive the hero product?
Branded purchase share
Are we still converting the demand the brand pays to create?
Category demand on the same searches
Did the market move, or did we?
Profit after ads, TACOS and ACOS by segment
Is the channel more profitable, and is each segment inside its goal?
How we kept the read honest:
•
Like-for-like search baskets. Amazon’s search report caps at 1,000 searches a week, so the mix changes week to week. Every comparison used only the searches present in both periods
•
A market control. Every report set our share against category demand. A gain in a growing market and a gain in a shrinking one mean different things
•
Clean branded and non-branded split. Shoppers misspell the brand name constantly, and dozens of misspellings were being counted as non-branded. Moving them to branded stopped them flattering the non-branded numbers
•
A weekly report to the client team, with a running log so every trend carried forward
The result: sales held on 42% less ad spend
•
The cut was complete within six weeks. Sales stayed in the same range for the four months that followed
•
Total orders rose 2% and organic orders rose 7%
The hero product kept its organic sales
Five weeks before the cut against five weeks after it:
-59%
Ad spend
0%
Organic sales
-31%
Ad-attributed sales
-10%
Total sales
+46%
Profit after ads
•
Organic sales did not move. The sales given up were ad-attributed, and each $1 of them had cost $1.66 in ad spend
•
The hero product kept its best-seller rank in its main subcategory
•
Organic sales held even though shopper demand on the hero product’s core searches fell 9% over the same weeks
Purchase share held on the hero product’s core searches
•
We tracked our share of purchases across 51 core searches for the hero product every week after the cut
•
Share held within a one-point band for nine straight weeks
•
On the most specific search for the hero product, share held close to its level before the cut
•
Branded purchase share stayed above 80% throughout
A new product line: sales up 2.7x in five months
The cut freed budget and attention for growth where it paid. A new product line launched shortly before the cut. Its first full month against five months later:
New product line
First full month
Five months later
Sales
Base
2.7x
Ads as a share of sales (TACOS)
60%
30%
Organic share of sales
19%
53%
Profit after ads
Loss-making
Profitable
•
Reached around the top 10 organically on its two main non-branded searches
•
Branded searches for the product doubled with the brand’s own marketing. Our job on Amazon was to convert that demand and build organic rank behind it
•
It kept higher growth ACOS goals after the hero product went back to efficiency goals, because it was still on target and gaining ground
After the cut: a leaner account, set up to test growth properly
With the waste out, every new dollar of spend had to prove itself. Each test had its success measure written down before it started.
•
New campaigns on the efficient base. Around 50 new campaigns launched in the eight weeks after the cut, covering competitor searches and gaps in keyword coverage, all run to the new ACOS goals
•
A ring-fenced growth test. A separate set of campaigns ran at higher ACOS goals to test whether more spend would buy extra non-branded sales, judged on four metrics set in advance. On the hero product the extra spend did not produce enough extra sales, so those campaigns went back to efficiency goals. The new product line kept the higher goals
•
A branded banner bid test. The Sponsored Brands banner on branded searches cost nearly three times as much per click as branded Sponsored Products. After the bid cut, branded ad spend fell 23%, branded ad sales held (+2%) and branded ACOS fell from 39.6% to 29.9%
The client’s view
When we reviewed the quarter with the client team, their Growth Marketing Manager summed up the result:
“We’ve scaled back significantly and it’s great that we haven’t seen a huge impact on our top line. When we look at contribution margin at channel level, we’re in a much better position than a few months back.”
Spending heavily on Amazon ads?
We’ll show you which of your spend is creating sales, and which is paying for sales you’d win anyway.
![](data:image/svg+xml,<svg display="block" id="13676550" role="presentation" viewBox="0 0 16 16" xmlns="http://www.w3.org/2000/svg"><path d="M 2.36 1.333 C 2.4 1.927 2.5 2.507 2.66 3.06 L 1.86 3.86 C 1.587 3.06 1.413 2.213 1.353 1.333 Z M 8.933 9.347 C 9.5 9.507 10.08 9.607 10.667 9.647 L 10.667 10.64 C 9.787 10.58 8.94 10.407 8.133 10.14 Z M 3 0 L 0.667 0 C 0.3 0 0 0.3 0 0.667 C 0 6.927 5.073 12 11.333 12 C 11.7 12 12 11.7 12 11.333 L 12 9.007 C 12 8.64 11.7 8.34 11.333 8.34 C 10.507 8.34 9.7 8.207 8.953 7.96 C 8.887 7.933 8.813 7.927 8.747 7.927 C 8.573 7.927 8.407 7.993 8.273 8.12 L 6.807 9.587 C 4.92 8.62 3.373 7.08 2.413 5.193 L 3.88 3.727 C 4.067 3.54 4.12 3.28 4.047 3.047 C 3.8 2.3 3.667 1.5 3.667 0.667 C 3.667 0.3 3.367 0 3 0 Z" fill="rgb(255, 255, 255)" height="12px" id="DbyFBxHNE" transform="translate(1.462 2.46)" width="12px"/></svg>)
jake(at)levo.agency
Join our newsletter, where we open up the toolbox on all things Amazon.
@2026 Levo Agency