Service
Amazon PPC management, measured in profit rather than ACoS.
Most Amazon PPC management is reported against a target ACoS because ACoS is easy to move and easy to present. It is also the number least connected to whether you made money. Elvatco builds the campaign structure, the bid rules and the reporting against contribution margin per unit instead.
The problem
ACoS is a ratio between two advertising numbers.
- Contribution before ads
- 44.3%
- Break-even ACoS
- 44.3%
- Ceiling at target net
- 29.3%
Try it with your own worst product. If the ceiling comes out near zero, no amount of campaign work fixes that one — the price or the cost has to move first.
ACoS divides ad spend by ad revenue. Nothing in that calculation knows your landed cost, your referral fee, your FBA fulfilment fee, your return rate or your storage bill. So a campaign can report a 15% ACoS and lose money on every unit it sells, while another reports 40% and is the most profitable line in the account.
This has to be calculated per ASIN, because fee structures, dimensional weight and COGS vary across a catalogue. One blanket ACoS target across a whole account is a guarantee that some products are being overspent and others starved.
The work
What actually gets run on the account.
Three structures, kept separate so the spend can be read. Click any of them.
Campaign architecture
Discovery, harvest and defence are kept in separate structures so that spend can be read. Auto and broad campaigns do the finding. Proven converters get promoted into their own exact-match campaigns where bids can be set against that term's own margin. Brand defence is budgeted separately from conquest.
Search-term harvesting
Converting terms get promoted on a fixed cadence and non-converting terms get negated at the right level, so discovery campaigns stay cheap instead of slowly turning into a tax. The negative lists are maintained, not generated once and forgotten.
Bid rules from unit margin
Bids are set from each ASIN's break-even and target ACoS, not from an account average. Products with different margin profiles get different ceilings, which is the entire point.
Placement strategy
Top-of-search is where rank gets bought and where the cost per click is highest. It gets its own margin maths and its own modifiers rather than being swept into one blended number.
Inventory coupling
Spend never gets scaled into a product that is about to stock out. A stockout is the most expensive advertising mistake available: you pay twice, once for the sales you lose and again to rebuild the rank you already owned.
Reporting you can take to the board
Contribution margin, TACoS as a read on ad dependency, and organic versus paid share by ASIN. If a report cannot be traced back to a line in the P&L it does not get sent.
Evidence
What this produced on a live account.
Scaling revenue and improving margin at the same time is the part that matters. Revenue alone is easy to buy: drop price, raise bids, accept a worse ACoS and the top line moves. Doing it while margin rises means the spend was allocated against unit economics rather than against a growth target.
The work was a full account restructure: PPC rebuild, catalogue cleanup, listing fixes and inventory systems, run directly rather than handed down.
The pattern
What an Amazon PPC agency usually gets wrong.
If you have hired an Amazon PPC agency before, some of this will sound familiar. The first month looks busy. Campaigns get rebuilt, negatives get added, and the ACoS comes down. You get a report with a lot of green on it.
Then month three arrives and sales are flat. The ACoS is still good, because the bids were cut to get it there, and cutting bids gives away impressions. You are paying a retainer to be smaller and more efficient. Nobody says that out loud, because the number everyone agreed to watch is still moving the right way.
The second pattern is the opposite. Spend climbs, sales climb with it, and the deck celebrates growth. Your bank balance disagrees, because the products doing the growing were the ones with the thinnest margins.
Both happen for the same reason. The agency is optimising a number that does not know what your product costs. Fix that and most of the usual problems stop happening on their own.
It is also why advertising is never run here as a standalone service in isolation. The same unit economics govern what you can afford to spend on Shopify, TikTok Shop and Walmart, and a product that cannot carry advertising on Amazon usually cannot carry it anywhere. Clients running several channels get them judged against one P&L rather than four separate reports.
Scope
Amazon PPC services, itemised.
No bundle names, no tiers. This is the actual list of what gets done on the account, so you can compare it against whatever you are being quoted elsewhere.
| Work | How often | Why it matters to you |
|---|---|---|
| Unit economics rebuilt per ASIN | At start, then quarterly | Every bid decision afterwards depends on this being right |
| Campaign structure rebuild | Once, then as the catalogue changes | You can finally read where the money goes |
| Search-term harvesting and negatives | Weekly | Discovery campaigns stay cheap instead of becoming a tax |
| Bid adjustments against margin ceilings | Weekly | Thin-margin products stop quietly draining budget |
| Placement and bid modifiers | Fortnightly | Top-of-search gets priced properly instead of blended away |
| Budget pacing against inventory cover | Weekly | You never pay to sell out early |
| Brand defence and competitor conquest | Ongoing | You stop bidding against yourself |
| Contribution-margin reporting | Monthly | A report you can take to the board without translating it |
Method
Your Amazon PPC strategy starts with your margins.
Most Amazon PPC strategy decks start with keywords. That is the wrong end. A keyword is only worth bidding on at a price, and the price depends entirely on what the product earns after Amazon has taken its share.
So the first week is spent on arithmetic rather than advertising. Every ASIN gets its own contribution figure, its own break-even, and its own ceiling. Three numbers per product. Once those exist, most bidding questions answer themselves.
It also tells you something uncomfortable fairly quickly. Usually a handful of products in the catalogue cannot be advertised profitably at all, at their current price and cost. Those need a price change, a cost change, or to be left to sell organically. Spending more on them will not fix them, and that conversation is better had in week one than in month six.
The products that remain get the budget. That is the whole strategy, and it is less exciting than a keyword deck, which is probably why so few agencies lead with it.
The rhythm
What Amazon PPC management services look like month to month.
You should know what you are buying before you buy it. Here is the shape of a normal engagement.
Weeks one and two
Unit economics per ASIN, a full read of the account, and the honest version of what is wrong. You get the break-even numbers whether or not you continue.
Weeks three to six
The structural rebuild. This is the part where reporting gets noisy, because campaigns are being rebuilt underneath it. Expect movement rather than a clean line, and expect to be told why.
Month two onwards
Weekly harvesting, weekly bid work, fortnightly placement review, and budget pacing tied to inventory. The account gets quieter to manage as the structure settles.
Every month
One report. Contribution margin, TACoS, organic versus paid share by ASIN, and what changed. If something went backwards, it says so and explains why.
Questions
Straight answers.
If you cut my ad spend in half and my profit went up, would you earn more or less?
The question worth asking any agency, including this one
Is a low ACoS a good result?
Not by itself. A low ACoS usually means bids were cut, which means impression share was given away. Whether that was correct depends entirely on where the break-even sits for that ASIN.
Will you manage PPC without touching the listings?
It can be done, but it caps the result. How well a page converts decides what a click is worth, and that comes down to the listing, the images and the price. Running ads into a listing nobody is allowed to fix puts a ceiling on the account. This is why PPC and account management usually run together.
How many accounts are you running at once?
Deliberately few. That limit is what makes the model work, and it does mean new accounts occasionally have to queue.
How long before I see anything?
Structural work shows up in weeks four to eight. Anyone promising a turnaround in two weeks is either cutting bids, which looks like a win and is not, or describing an account that only needed a small fix. You will know which yours is after the audit.
Do I keep control of my own ad account?
Yes. It stays yours, you keep full access throughout, and nothing is run through an agency-owned account. If we stop working together you keep the structure, the negative lists and the data, because they were always yours.
My ACoS looks fine already. Is there anything to gain?
Possibly not, and if so the audit will say that. A healthy ACoS on an account with low impression share usually means bids were cut too far and there is room to grow profitably. A healthy ACoS with strong share means it is working. The two look identical in a report and completely different in the account.
Do you work with small catalogues?
Yes, and small catalogues are often easier to move. The limit is not how many products you have, it is whether the margin supports the engagement. If it does not, that is better established at the audit than three months in.
What does it cost?
Pricing is published rather than held back for a discovery call. See Amazon agency pricing.
Free PPC audit
Find out what your real break-even is.
Send the account over and you will get the break-even ACoS for every ASIN, the products currently spending above it, and what that is costing you each month. You keep the numbers either way.
Audits are done personally, so there are only a few slots each month.
Request the audit