The monthly agency report is thirty slides long. The dashboards are color-coded. The ACOS chart is trending in the right direction. And yet the brand’s overall Amazon margin keeps sliding in the wrong direction. Activity is everywhere. Outcomes are nowhere.
Most established brands aren’t asking whether their Amazon agency is working hard. They can see the work. What they can’t tell is whether it’s the right work. Whether the metrics they’re being shown were chosen because they measure progress, or because those are the metrics that are moving.
Amazon agency relationships often fail in a predictable pattern: enthusiastic onboarding, activity-rich reporting, gradual erosion of confidence, and an eventual realization (usually quarters too late) that nothing structural ever changed. The brands that catch this early have a framework for distinguishing tactical motion from strategic progress. But brands that don’t catch it churn through two or three agencies before the underlying pattern becomes obvious.
This piece gives you a framework for making that distinction. Six diagnostic questions to ask of any current or prospective Amazon agency. How to read a monthly report for what it omits, not just what it includes. The specific signals that distinguish a tactical execution vendor from a strategic diagnostic partner, and which one your business actually needs.
This is Part One of a two-part series on Amazon agency partnerships. Part One asks whether your agency is doing the right work. Part Two: “How to Make an Agency Work for You”, asks whether the conditions exist for any agency to do the right work at all. If you’ve worked through the six questions below and the answers concern you, Part Two is worth reading before you make any changes to the relationship.
Why Does My Amazon Agency Report Look Busy While My P&L Drifts?
Amazon agency reports are usually designed around activity, not outcomes, because activity is what the agency controls, plus it’s easy to chart. The metrics that would reveal whether the activity is moving the business, like contribution margin per ASIN, organic rank trajectory, and incremental ROAS, are harder to surface and often contradict the activity narrative.
Why Reports Get Designed Around Activity
Agency reports are built to demonstrate work, because activity/work is what the agency controls. Things like campaigns launched, keywords added, listings updated, A+ modules refreshed, bid adjustments deployed, etc. All of it is real but none of it is the same as moving the business in the right direction.
The structural issue is that activity metrics are abundant and easy to chart. Outcome metrics (contribution margin per ASIN, organic rank trajectory on commercially meaningful terms, customer acquisition quality, the gap between attributed ROAS and incremental ROAS) are harder to surface and often contradict the activity narrative. Agencies that aren’t equipped to do the harder work will report on easier metrics instead. In this scenario the reports look good but your business loses footing anyway.
The ACOS Trap
ACOS trending down is the most common banner metric, and one of the least informative. ACOS can improve because the agency is bidding more efficiently on the same wrong keywords. Or because branded search is being credited to paid spend it would have converted anyway. It can even improve because organic was already going to win that customer. None of those scenarios mean the business got better, they just mean the chart got better.
The deeper problem isn’t that ACOS is wrong. It’s that ACOS is being used as the headline when contribution margin per ASIN should be. A metric that can improve while the underlying business deteriorates is not a metric that belongs at the top of a monthly report.
What This Looks Like From the Inside
If you’ve inherited an agency relationship from a predecessor, or you’ve watched a previously promising engagement settle into routine, the pattern is recognizable. The reports stop surprising you. The monthly call covers the same ground. Optimizations happen at the edges. Nobody is doing anything wrong but nothing structural is changing.
If you’re trying to bring this issue to a CFO, a board, or a PE sponsor, the framing matters. “We don’t trust the agency anymore” doesn’t translate to a board deck. “Here are six diagnostic questions our agency cannot currently answer” does.
What Questions Should I Ask My Amazon Agency? Six Diagnostics That Surface the Truth Quickly
These six questions identify the difference between a strategic Amazon partner and a tactical execution vendor. An agency that engages substantively with all six is operating as a strategic partner. An agency that deflects on three or more is operating as an execution vendor. This may be fine, depending on what your business actually needs.
Each question is structured the same way: what to ask, what a strategic answer looks like, what a tactical or evasive answer looks like, and why the distinction matters.
1. “Can you show me contribution margin per ASIN, reconciled across advertising, FBA fees, returns, and storage?”
This is the single most diagnostic question available. A strategic partner can produce this view, even if assembling it takes effort. A tactical agency will offer ACOS and TACoS instead, or explain that Amazon doesn’t make this data easy.
The second answer is true. It’s also beside the point. Amazon doesn’t produce a unified P&L. Advertising lives in one system, FBA fees in another, chargebacks in invoices, inventory placement in settlement detail. Reconciling these into a true contribution-margin view per ASIN is detailed work. Doing it is what distinguishes a partner from a vendor. The reason most tactical agencies can’t do it is simply that they’re not scoped or staffed to.
2. “Which of our current keywords are profitable at the contribution-margin level, and which are losing money even at low ACOS?”
This tells you whether the agency is optimizing the right area. A strategic partner separates branded from non-branded, defends or cuts based on incrementality, and can name the specific keywords that look profitable on the surface but lose money once returns and acquisition quality are accounted for. A tactical agency reports on the keyword set as a whole and points to aggregate ACOS as evidence the program is working.
Low ACOS on the wrong keyword is still a loss. Knowing the difference requires the contribution-margin view from Question 1. So agencies that can’t answer the first question can almost never answer this second question either.
3. “What is our organic rank trajectory on our top twenty commercially meaningful keywords, and how is paid spend influencing it?”
The interaction between paid and organic is where most growth ceilings sit. A strategic partner tracks rank trajectory on a defined set of commercially meaningful terms (as opposed to aggregate impression share) and adjusts spend to support organic rank rather than cannibalize it. A tactical agency reports paid and organic in separate sections of the deck and doesn’t connect them.
The disconnect matters because paid spend on a term where you already rank organically is, for most brands, a transfer of margin from operations to advertising. Tracking the relationship is how you tell the difference between paid spend that builds the asset and paid spend that depletes it.
4. “Where is margin leaking outside of advertising… supply chain, packaging, AWD placement, chargebacks, IPI?”
This question tests whether the agency thinks beyond its scope of work. The best Amazon agencies (even those whose contract is nominally advertising) flag operational margin issues because they affect every other lever. A change in dimensional weight, a chargeback pattern, an IPI score drift, an AWD placement decision… each of these can move contribution margin further than a bid adjustment will.
A tactical agency points to its statement of work and explains that this is outside their scope of engagement. The answer is technically correct, and it tells you exactly what kind of relationship you have.
5. “What hypothesis are we currently testing, and what would change in our strategy if it were proven wrong?”
This question reveals whether there is a strategy at all, or just a recurring set of optimizations. A strategic partner can name a current hypothesis, about customer acquisition, content positioning, category dynamics, something specific, and describe the data that would invalidate it. A tactical agency describes its monthly workflow.
It’s the difference between a partner who is accountable for a position and a vendor who is accountable for activity. A position can be wrong. Activity can only be more or less efficient. Strategic progress requires partners who take positions.
6. “If we paused advertising entirely for thirty days on three representative ASINs, what would you predict happens to revenue, rank, and contribution… and why?”
This is the incrementality question, and the one tactical agencies resist most. A strategic partner has run, or can credibly model, the experiment. They can talk through what they’d expect for each ASIN, why, and what they’d learn from being wrong.
A tactical agency treats the question as theoretical, or actively resists it because their billing is tied to ad spend management. The resistance itself is the answer. An advertising program whose value cannot be tested is an advertising program nobody is sure is working.
How to Use the Six Questions
Bring them to a quarterly review, or to the diligence conversation with a prospective agency. The answers don’t need to be perfect. They need to exist. Watch for three patterns: substantive engagement with all six (a strategic partner), substantive engagement with two or three (a competent execution vendor that may or may not be the right fit), and deflection on most of them (an engagement that has settled into routine and is unlikely to produce strategic outcomes regardless of how long it runs).
How Do I Read an Amazon Agency Report Critically?
Start with what the report omits, not what it includes. The most diagnostic Amazon agency reports include contribution margin, incrementality, organic rank trajectory on commercially meaningful keywords, and the relationship between paid spend and organic conversion. Reports that omit these are almost always reporting on activity, not outcomes.
The Absent Metrics
A diagnostic read of any agency report should start with what isn’t there. The common omissions are predictable: contribution margin replaced by ACOS or TACoS; incrementality replaced by attributed ROAS; organic rank trajectory on commercially meaningful terms replaced by aggregate impression share; return rate by acquisition source not reported at all; and the relationship between paid spend and organic conversion treated as two separate stories rather than one.
If a report doesn’t include these, instead of asking whether the agency could include them, ask them why they don’t already.
The Narrative Test
A second test: read the report’s narrative section and ask whether anything in it would be falsifiable. “We saw strong performance across branded campaigns and continued to optimize bids on non-branded discovery terms” describes activity in a way that cannot be wrong because it’s so generic. So it’s not really a position.
Compare that to this: “Our hypothesis that non-branded discovery was acquiring lower-LTV customers held. We cut spend on those terms by 40% and contribution rose without revenue impact.” That describes a position the agency took, tested, and was accountable for. It could have been wrong. The fact that it could have been wrong is what makes it worth looking at.
The first pattern is what most monthly reports often sound like. The second is a strategic partnership in action.
The Question Cadence
Strategic partners ask their clients more and more difficult questions over time, not easier ones. If the agency relationship has become routine, with the same report cadence, same metrics, fewer probing conversations than in month three, that’s a signal the engagement has settled into execution mode, regardless of what the contract states. The most useful question to ask of your current agency relationship may not be about a metric at all. It may be: when did they last challenge you on something?
What’s the Difference Between a Tactical and Strategic Amazon Agency. And Which Do I Need?
A tactical Amazon agency provides execution capacity: PPC management, listing maintenance, content production, etc., for brands whose Amazon business is structurally healthy and whose need is more hands. A strategic Amazon agency provides diagnostic depth: forensic analysis of margin, advertising incrementality, and growth ceilings, for brands whose problem is not capacity but unanswered questions. Adding more tactical hours to a diagnostic problem produces more reports, not better answers.
When Tactical Execution Is the Right Fit
A tactical agency is the right partner for a brand whose Amazon business is structurally healthy, whose internal team has diagnostic capability, and whose need is execution capacity. PPC management, listing maintenance, and content production at scale are real, valuable services. A brand that knows what to do and needs someone to do it well is genuinely well-served by tactical execution. There is no inherent error in hiring a tactical agency. The error is hiring an agency to solve a problem it isn’t built for.
When Strategic Diagnosis Is the Right Fit
Strategic diagnosis is the right partner when the brand is not in that position: when margin is eroding for unclear reasons, when the internal team has hit a strategic ceiling, when previous engagements produced activity but not outcomes, or when an investor, board, or PE sponsor is asking questions the current data can’t answer. These are not problems tactical execution can solve, because the problem isn’t capacity, it’s diagnosis.
The Hybrid Trap
Many brands try to solve a strategic problem with a tactical agency by adding scope or hours. The result is usually more activity reports, not better answers. Diagnosis isn’t a function of effort applied to existing data. It is a function of asking different questions, of different data, in a different framework. Most tactical agencies just aren’t structured for diagnostic work, regardless of how many hours they bill.
The failure mode in most agency relationships isn’t hiring the wrong agency. It’s hiring one kind of agency to solve the other kind of problem.
What Strategic Partnership Actually Looks Like
For brands we work with, Troveres operates as a fractional VP of Amazon Operations. We are the strategic and diagnostic layer that sits above tactical execution, sometimes alongside the brand’s existing agency, sometimes in place of one. Our methodology is structured honestly: deep forensic research before any significant tactical intervention, and 12–18 month engagements built around diagnosis-first work rather than month-to-month optimization cycles.
That is what strategic partnership looks like when it is built to produce outcomes rather than activity. Whether a strategic partnership is really what your business needs is a separate question.
A Second-Opinion Review
The issue isn’t whether your Amazon agency is working hard. It’s whether the work being done is the work that is needed, measured against outcomes that would survive scrutiny from a CFO, a board, or your own honest assessment. The six diagnostic questions are designed to make that distinction visible quickly. To surface, in either direction, what kind of partner the business actually needs.
One important caveat: these questions diagnose the agency. They don’t diagnose the relationship. If you’ve worked through all six and the answers concern you, it’s worth asking a harder question before making any changes: whether the agency is the variable, or whether your organization has built conditions that would prevent any agency from succeeding.
That’s the subject of Part Two of this series, “How to Make an Agency Work for You,” which examines three partnership diagnostics that apply to the brand, not the agency: whether you’re validating claims against actual Amazon data, whether you’re operating as a true partner on inventory and financial planning, and whether your expectations are calibrated to what you’re actually paying. Many brands that have rotated through two or three agencies find the second set of questions more uncomfortable than the first.
A Second-Opinion Review is a low-commitment first step: a structured external read of the current agency relationship, the metrics being reported, and the strategic questions not being asked. It is not a pitch for replacement. It is a diagnostic of whether the current engagement is producing what the business needs, and what would need to change if it isn’t.
Other articles that may interest you:
Your Amazon Margin Problem Isn’t Your Ads — It’s Your Operations
What Your Amazon P&L Isn’t Telling You: A Forensic Breakdown
The Wrong-Customer Problem: How Bad Ad Targeting Actively Tanks Your Organic Rank
Frequently Asked Questions
How do I know if my Amazon agency is actually working?
The honest test is whether your agency can produce contribution margin per ASIN, name a current strategic hypothesis they’re testing, and predict what would happen if you paused advertising on representative ASINs for thirty days. An agency that can answer all three is operating as a strategic partner. An agency that deflects or treats those questions as theoretical is operating as a tactical execution vendor. This makes them valuable for capacity, but insufficient for diagnosis.
What questions should I ask my Amazon agency in a quarterly review?
Six questions surface the most: contribution margin per ASIN, keyword profitability at the margin level, organic rank trajectory on commercially meaningful keywords, where margin is leaking outside advertising, what hypothesis they’re currently testing, and what they’d predict if advertising paused. The answers don’t need to be perfect, they just need to exist. If three or more get deflected, the engagement is execution-only.
Is ACOS a reliable measure of Amazon agency performance?
ACOS is one of the least informative top-line metrics in Amazon advertising. It can improve because the agency is bidding more efficiently on the wrong keywords, because branded search is being credited to paid, or because organic was already going to convert that customer, none of which mean the business got better. Contribution margin per ASIN and incremental ROAS are the metrics that actually reveal whether advertising is working.
When should I fire my Amazon agency?
Consider replacement when monthly reports describe activity in ways that cannot be falsified, when the agency cannot produce contribution margin reconciled across advertising and operations, and when probing questions have decreased rather than increased over the course of the engagement. The decision is rarely about effort. It’s about whether the work being done is the right work for the problem your business actually has.
What’s the difference between a tactical and strategic Amazon agency?
A tactical Amazon agency provides execution capacity: running PPC, managing listings, and producing content for brands whose Amazon business is healthy and whose need is more people to do the work. A strategic Amazon agency provides diagnostic depth: forensic margin analysis, advertising incrementality testing, structural growth-engine identification, etc., for brands whose problem is not capacity but unanswered questions. Most failed agency relationships are mismatches between these two models.
Can I get a second opinion on my current Amazon agency without replacing them?
Yes. A second-opinion review is a structured external read of the current agency relationship, the metrics being reported, and the strategic questions not being asked. It is not a pitch for replacement, it is a diagnostic of whether the current engagement is producing what the business actually needs. Many brands use it to strengthen an existing relationship rather than end one.
Why do Amazon agencies struggle to produce contribution margin reports?
Amazon does not produce a unified P&L. Advertising lives in one system, FBA fees in another, chargebacks in invoices, inventory placement in settlement detail. Reconciling these into a true contribution-margin view per ASIN requires forensic work that most tactical agencies are not scoped or staffed for. The capability to do this work is one of the clearest signals of a strategic partner.