The reports arrive on time. The charts point upward. And still, the account doesn't feel like it's going anywhere. Most brands don't leave an agency because performance collapsed.
They stay too long because nothing was ever obviously broken, and by the time the problem is obvious, a year of compounding has already been spent.
A Good Agency Is Judged on Its Decisions, Not Its Dashboard
A good Amazon agency can explain the reasoning behind every decision it makes, reports on profit rather than volume, raises problems before the brand has to ask, and can show comparable results on brands of similar size and category.
When the reporting is opaque, the strategy hasn't changed in a year, and spending keeps rising while growth flattens, the agency is maintaining an account rather than growing one.
The distinction that matters is between activity reporting and decision reporting. A report listing what was done is not the same as a report explaining what was decided and why. Most agency reporting is the first kind. It's a record of motion, and motion is easy to produce.
Agencies rarely fail loudly. They fail by degrees, through a quarter with no real decisions in it, then another, while the account drifts on the momentum it built earlier. That's why this needs a deliberate audit rather than a gut feeling.
The Metrics an Agency Shows Us Reveal What It Optimizes For
Impressions, clicks, and gross ad sales are the easiest numbers on Amazon to make look good, because every one of them can be improved by simply spending more. An agency leading with them is showing us motion, not progress. Nothing in that set of numbers can tell us whether the business is better off than it was.
The numbers that can tell us are less flattering and harder to move. TACoS and the direction it's travelling. Contribution margin after ad spend and fees. Organic rank movement on the keywords that actually matter in the category. Share of voice against the competitors we are actually losing to. None of those improve simply because the budget went up, which is precisely what makes them worth reporting.
Here's the pattern we find most often when we audit an account that came from somewhere else. Revenue is up thirty percent year on year, and everyone is pleased. TACoS has climbed from eight percent to fourteen in the same window, which means the growth was purchased rather than earned, and the business now pays more for every dollar it makes than it did a year ago. The dashboard is green. The margin is worse. Nobody lied. The wrong thing was measured.
Strategy That Never Changes Isn't Strategy
A useful test takes about ten minutes. Pull the campaign structure and bid logic from twelve months ago and set it beside today's. If nothing meaningful has changed, nobody has been thinking. Amazon changed in that year. The category changed. Competitors entered, exited, and repriced. The account did not.
There's a related question about the stage. The right structure for a launch is the wrong structure at maturity, and an agency running the same architecture across every client regardless of where the brand sits is running a template. Templates scale beautifully for the agency. They rarely serve the account.
Proactivity is the other half of it. A strong agency raises the inventory risk, the suppressed listing, or the competitor's price move before we notice it ourselves. A weak one waits to be asked and then answers well. Answering isn't the job. Noticing is the job.
An Agency Comparison Should Compare Reasoning, Not Service Lists
Most agency comparison exercises line up service offerings and monthly fees, and both are close to identical across the market, which makes them nearly useless as a basis for choosing. Every agency on the shortlist does advertising, listings, creative, and account health. Every one of them says so on the homepage.
The comparisons that carry weight are harder to run and worth the effort. Ask each agency how it would approach the same account at launch, in growth, and at maturity, then listen for whether the three answers genuinely differ. Ask who does the daily work, because the people in the pitch are frequently not the people in the account. Ask how profit gets reported rather than revenue.
Case studies deserve the same scrutiny. A result without a starting position is a claim, not evidence. The three questions worth asking of any Amazon agency case study are what the account looked like beforehand, which specific decisions were made, and what would plausibly have happened without them. An agency that can't answer all three about its own published work is showing us marketing rather than proof.
This is also why searching for the best Amazon agency in the USA tends to disappoint. There isn't a single best one. There's the one whose strengths line up with a brand's stage, category, and marketplace complexity, and no ranking list can answer that question on a brand's behalf.
Some Problems Are Worth Fixing Rather Than Replacing
Switching agencies carries a cost that rarely gets priced in. Ramp time runs a quarter at a minimum. Account context, the accumulated knowledge of what's been tried and what failed, mostly doesn't transfer. And the replacement might be worse. Not every frustration is a firing offense.
So it's worth separating the fixable from the fatal. Unclear reporting, the wrong metrics, and a thin communication cadence are all fixable, and they usually respond to one direct conversation about what gets reported and how often. What isn't fixable is an absence of strategic reasoning, weak account health competence, or a full year of structural stagnation. Those are capability gaps, and capability gaps don't close because somebody asked politely.
Before deciding either way, we'd put three questions to the incumbent. What decision did you make on this account last month, and why. What will you do differently next quarter, and on what evidence. What is our contribution margin after ad spend and fees. The quality of those three answers usually settles it.
How We Read an Account That Came From Somewhere Else
When a brand comes to us from another agency, the first thing we look at isn't the performance chart. It's the decision history. What was changed, when, and on what basis. An account with clean numbers and no reasoning behind them is a fragile account, because nobody can say why it's working or what would break it.
That's how we run our own accounts too, and it's why how we approach an account starts with an honest assessment of where the brand actually is rather than with a proposal. The goal was never a better-looking report. It's a better decision about what actually converts, which is a different thing and a considerably harder one.
Sometimes that assessment tells us the current agency is doing a reasonable job and the real constraint sits somewhere else entirely, in pricing, in inventory, or in the product itself. We say so when that's what we find. It costs us the engagement, and it's still the right answer.
The question was never whether an agency is good in some absolute sense. It's whether it's making better decisions for this brand than anyone else would make. Remarkably few brands ever actually check, which is how so many agency relationships outlive their usefulness by a year or more.
If the reports look fine but the account feels stuck, that gap is worth investigating before the next renewal comes around. Let us read the last twelve months of decisions and say plainly what we find.
Frequently Asked Questions
How do I know if my Amazon agency is underperforming?
The clearest signals are structural rather than numerical. Reporting that lists activity instead of explaining decisions, a campaign architecture that hasn't meaningfully changed in a year, and rising spend alongside flattening growth all point the same way. A green dashboard sitting on top of a climbing TACoS is the most common version of this.
What metrics should my Amazon agency be reporting on?
TACoS and its direction, contribution margin after ad spend and fees, organic rank movement on the keywords that matter in the category, and revenue from customers new to the brand. Impressions, clicks, and gross ad sales can all be improved by spending more, which makes them poor evidence that the business is better off.
How often should an Amazon agency change its campaign strategy?
There's no fixed cadence, but a structure identical to how it looked twelve months ago is a warning sign. Amazon, the category, and the competitive set all move continuously. Strategy should change when the evidence changes, and an agency should be able to point to the evidence behind its last significant change.
What should I look for in an Amazon agency’s case studies?
Context rather than outcome. A result without a starting position is a claim, not evidence. Worth asking what the account looked like beforehand, which decisions were made, and what would plausibly have happened without them. Case studies on brands of similar size, category, and marketplace complexity carry far more weight than headline percentages.
Should I switch Amazon agencies or fix the relationship first?
It depends on whether the problem is a communication gap or a capability gap. Unclear reporting, the wrong metrics, and thin communication are usually fixable in one direct conversation. An absence of strategic reasoning, weak account health competence, or a year of structural stagnation are capability gaps, and those rarely close on request.
Which is the best Amazon agency in the USA?
There isn't a single answer, and any ranking offering one is selling something. The agency that suits a launch-stage seller in a low-competition category is rarely the one that suits an established brand scaling across several marketplaces. Fit against stage, category, and complexity is the thing worth optimising for.
How do I compare two Amazon agencies fairly?
Compare reasoning rather than service lists, because the lists are nearly identical across the market. Ask each how it would handle the same account at launch, growth, and maturity, and listen for whether the answers genuinely differ. Ask who performs the daily work. Ask how profit, not revenue, gets reported.
