You’re staring at an Amazon P&L that doesn’t make sense.
Revenue is flat or falling. Ad costs keep climbing. Your margin is eroding in places you can’t quite name. And none of it tracks with the work that’s supposedly being done on your behalf: the keyword refreshes, the new images, the dashboards, the monthly calls where someone walks you through a deck and assures you things are “on track.”
You’ve done everything right: hired more people, spent more money. You’ve tried two agencies, maybe three. And the harder you push, the worse it seems to get.
Here is the thing no one has told you yet: you almost certainly do not have a performance problem. You have an undiagnosed structural problem. And until someone actually finds it, no amount of optimization is going to fix it.
Most Amazon agencies don’t diagnose. They optimize. Those are not the same thing, and the difference is often why “nothing works.”
This post is about that difference. What the standard agency playbook misses, the three hidden categories where margin actually disappears on Amazon, and what a real forensic diagnosis looks like when someone is willing to do the work.
The standard agency playbook, and why it fails
If you’ve been through an Amazon agency engagement or two, the rhythm is probably familiar. Onboarding call. Keyword audit. A round of image refreshes and A+ updates. Some PPC restructuring. A monthly report with a lot of graphs. A quarterly business review where everyone agrees to “keep testing.”
None of this is wrong. Keywords and images do matter. Bids do need to be managed. The problem is that these actions are tactical. But tactics, applied to a business with a structural issue, are the equivalent of tuning the engine on a car with a cracked axle. The engine runs smoother for a minute. The car still pulls to the right.
The activity trap
Agencies are incentivized to show activity. Reports, dashboards, keyword counts, campaigns launched, creative refreshed. Activity is easy to invoice. It feels productive. It fills the monthly call with things to talk about.
But activity can also mask the absence of real diagnosis. A brand can cycle through two or three agencies, each one running essentially the same playbook with slightly different branding, and never once have someone step back and ask the question that actually matters: why is this business losing money?
That question does not get answered by a keyword audit. It only gets answered by a forensic look at the entire operation, but few agencies are structured to do that work.
Why this is especially painful for proven brand builders
If you’ve built a successful consumer products brand, you’re used to understanding your business. You know your margins, and your customer. You can walk through your wholesale numbers, your DTC funnel, your retail sell-through, and explain what’s happening and why.
Amazon takes that away. The data is opaque. The reports are fragmented. The levers that matter in other channels don’t behave the same way here. And when the agency you’ve hired to translate that opacity into clarity instead hands you more dashboards, you don’t just lose money. It starts to feel like you’ve lost control of a channel that should be within your competence.
You don’t need more activity. You need answers.
The three hidden categories of margin erosion
In our experience, margin erosion on Amazon almost always lives in one or more of three places. These are the areas that standard agency reviews do not touch, usually because the agency’s scope was never built to look at them in the first place.
1. Operational and supply chain leakage
This is the quiet killer. Master carton sizing that’s inefficient for AWD and FBA is bleeding storage and placement fees every week. Inbound shipping structures that made sense at launch are now wildly out of step with current volume. Inventory is mis-allocated, overstocking slow SKUs and starving the winners, which tanks a brand’s rank as soon as demand peaks.
These costs don’t show up in any agency report because operations aren’t in the agency’s scope. Nobody is asking whether a carton change would save six figures a year. Nobody is looking at whether your inbound freight terms still match your volume. The money is walking out the door before the first ad impression ever loads, and no one is watching the door.
2. Advertising misalignment
Ad spend is one symptom that many founders fixate on, but it’s rarely the root cause. Even when it is part of the problem, the issue is almost never “the bids are wrong.” The issue is often who the ads are reaching.
A lot of Amazon ad budgets are busy acquiring the wrong customers… shoppers who click but don’t convert, or who buy once and never return. Campaign structures optimize for impressions or clicks rather than for the acquisition of a customer who’s actually profitable over their lifetime with the brand.
It compounds. Bad ad data trains Amazon’s algorithm to serve the listing to the wrong audience. That audience converts poorly, which signals to the algorithm that the listing itself is weak, which suppresses organic ranking. Now you’re paying more for worse placement and blaming the ads for a problem they’re really only one part of.
A concrete example from our own work: a brand came to us spending $1.9M a year on Amazon ads. Their previous agency’s dashboards looked fine. When we looked at who those ads were actually acquiring, the answer was customers who were never going to be profitable for the brand. We cut ad spend. Profit went up $1.3M. Revenue didn’t change. The ads had been performing perfectly, for the wrong audience.
3. Content-to-search-intent mismatch
The third hidden category is the one almost nobody checks: whether your listings are actually answering the question the shopper is asking when they search.
A listing can rank for a keyword and still lose the sale, because ranking is not the same thing as fit. Shoppers search with an intent: a job to be done, a problem to solve, a gift to buy for a specific occasion. If your title, images, bullets, and A+ content are written for your brand story rather than that shopper’s intent, the click-through is low, the conversion is lower, and Amazon’s algorithm reads both as signals that your listing isn’t relevant. Organic rank drops. Ad costs rise to compensate. The cycle tightens.
Beautiful A+ content and a polished Storefront do not fix this. They often make it worse, because they don’t align with the shopper’s intent… they don’t answer the right questions and counter objections the way they need to.
What a real diagnosis looks like
If you’ve only experienced the standard playbook, it’s worth saying plainly what a different kind of engagement actually involves. Not as a pitch, but as a benchmark you can hold any agency partner against.
A forensic diagnostic phase is not a week of onboarding and a kickoff deck. It’s three to six months of deep research before any significant implementation. It exists because you cannot build a durable strategy on top of data you haven’t fully understood yet, and because several of the questions that matter most require a full seasonal cycle to answer honestly.
What actually gets examined
- A full P&L teardown at the SKU level. Not just top-line revenue and blended ACoS, but margin by unit, by variation, and by channel, with a clear view of where profit is made and where it’s lost.
- A supply chain and logistics audit. Carton sizing, inbound terms, AWD utilization, FBA placement, inventory allocation. Where is margin disappearing before a unit ever reaches a shopper?
- An advertising ecosystem analysis. Not just bid efficiency, but who is actually buying, how often they come back, and whether your ad spend is acquiring customers who are good for the brand.
- A content-to-search-intent audit. Are your listings aligned with what shoppers are actually searching for, and does your content convert the right shopper instead of performing for the brand?
- A competitive and algorithmic positioning review. Where you sit in the category, why, and which levers will actually move that position.
Why it takes time
Some of the most important data only makes sense in context. Seasonal keywords cannot be tested outside of their season. A hypothesis about customer quality needs a repeat-purchase window before you can validate it. Supply chain changes take weeks to implement and months to measure. Skipping any of that is how brands end up back where they started — with new reports and the same problem.
We know the objection here, because we hear it on almost every first call: “I need to see results faster than 12 months.” It is a completely rational thing to say. It is also the exact reasoning that leads to agency engagements that don’t work. Shortcuts are what got the brand to this point. A real diagnosis is what gets it out.
What it looks like in practice
Consider a brand in the $15M–$50M range, on Amazon for several years, revenue plateaued, multiple agencies tried, each departing with a different explanation.
The symptom: “Our ads aren’t working. Margin is falling every quarter and we can’t see why.”
The diagnosis: The ads were not the core issue. At the SKU level, master carton configuration was inflating AWD and placement fees enough to compress unit margin by a meaningful amount on the best-selling variations. A separate but related issue: the ad mix was acquiring a customer cohort that rarely repeated, and the content was written for the brand’s internal narrative rather than the search intent of the shoppers Amazon was actually serving it to. All three problems were compounding each other.
The outcome: The early months looked, on paper, like nothing was happening. Research, data collection, hypothesis testing. From roughly the seventh month forward, structural changes began to compound — a restructured ad strategy aligned to profitable acquisition, a content overhaul aligned to search intent, and a carton/logistics rework that returned margin on every unit sold. None of that would have shown up in a 90-day “fix.” All of it held, because it was built on an actual understanding of the business.
What to look for in your next Amazon partner
If you take nothing else from this, take the checklist. These are the questions we wish every founder asked before signing with us or anyone else.
- Do they diagnose before they optimize? If the engagement starts with “we’ll optimize your campaigns in week one,” that’s a red flag, not a selling point.
- Do they look at your operations, or just your ads? If supply chain, logistics, and catalog structure aren’t in scope, the diagnosis is incomplete by design.
- Do they ask about your customer, or just your keywords? Understanding who buys and why is the foundation of everything on Amazon. Keyword lists are downstream of that answer, not a substitute for it.
- Are they willing to tell you things you don’t want to hear? The best partners challenge your assumptions. If every finding conveniently confirms what you already believed, you’re being managed, not diagnosed.
- Do they commit to a timeline that allows real diagnosis? Be skeptical of 90-day guarantees. Pattern-matching works on simple problems. The one you have isn’t simple. If it were, the last agency would’ve caught it.
If nothing has worked, the problem was never correctly identified
The hardest thing to accept, after several agency engagements, is that none of the fixes were wrong. They just weren’t aimed at the real problem. And until someone does the unglamorous work of finding that problem, the next engagement is going to feel exactly like the last one.
That is the work we do. Not a sales call, not a pitch, not a promise that we’ll turn things around in a quarter. A genuine evaluation of whether there is a structural issue in your Amazon business worth investigating. And, if there is, what it actually is.
If your Amazon business is underperforming and you’ve tried everything short of understanding why, apply for a diagnostic consultation. We’ll start by figuring out whether there’s a problem worth solving — and if there is, what it actually is.