Most CPG brands hire the wrong Amazon agency on the first try. The pattern is consistent: founder gets pitched, picks the agency with the best deck, signs a 12-month retainer billing on percentage of ad spend, watches ad spend triple and net revenue stay flat, fires the agency at month 10. The fix is structural, not tactical. Here's the framework.
The single biggest filter: how does the agency get paid?
Three pricing models exist. (1) Percentage of ad spend (industry default, 8 to 15%). (2) Flat retainer (rare, $5K to $20K/mo). (3) Percentage of gross revenue (rare but growing, 2 to 5% plus base retainer). The first model rewards budget growth. The second rewards nothing. The third rewards revenue growth. Pick the one whose incentive matches what you actually want.
The percentage-of-ad-spend model is the industry default for one reason: it's easier to forecast agency revenue than to share in client revenue. It's not designed for client outcomes. It's designed for agency cash flow predictability.
That doesn't make every percentage-of-ad-spend agency wrong. It makes the structure misaligned. Some good operators work inside that structure and produce real results in spite of it. But you should at least know what you're signing.
Percentage-of-gross-revenue pricing flips the incentive. The agency makes more only when your top line grows. The agency makes less when ad spend inflates without revenue lift. We charge this way at Eleviam because we believe it's the only structure that survives a five-year client relationship without a forced renegotiation.
What questions should you ask in the first sales call?
Seven questions that separate operator-style agencies from sales-deck agencies. Ask all seven. The agency that answers six of seven well is usually a real fit. The agency that deflects on three or more is showing you the version of itself it markets, not the version that will operate your account.
- How do you charge? Listen for ad spend versus gross revenue. If ad spend, ask why their incentive is aligned with yours.
- What's your minimum brand size? Agencies that take any brand at any revenue are optimizing for fee revenue, not client outcomes. Real agencies turn down small brands.
- Can I talk to two of your current clients in my category? If they hesitate or only offer "anonymous case studies," the case studies are probably weaker than they sound.
- What's the first thing you'd fix in my account? A real operator can answer this within a few minutes of seeing your storefront. A sales rep deflects to "after we onboard."
- How do you handle MAP enforcement and unauthorized resellers? If they only offer "monitoring tools," they don't actually fix the channel. They just flag the problem.
- What's the 30/60/90 day milestone framework? Real engagements have written milestones. Vague timelines are a signal of vague execution.
- What kinds of brands have you fired or declined? Agencies with a real disqualification framework (revenue, margin, category) will give you specific examples. Agencies that take everyone won't.
How do you read agency case studies critically?
Look for four signals. Named brands (or specific enough descriptions to identify them). Absolute numbers (revenue, units, contribution margin) not just percentages. Acknowledgment of what didn't work alongside what did. A clear before-state, the intervention, and the after-state. Case studies missing any of these are usually marketing fiction.
"We grew this brand 300%" is meaningless without the starting revenue. 300% growth from $5K/month to $20K/month is a different story than 300% growth from $200K/month to $800K/month. The first is a startup case. The second is an operator case. Most marketing decks blur the difference.
"We achieved a 2.5x ROAS" is meaningless without TACoS context. ROAS measures attributed sales against ad spend. TACoS measures total ad spend against total revenue. A campaign with 4x ROAS and 28% TACoS is buying clicks at scale without growing the organic base. A campaign with 2x ROAS and 11% TACoS is investing in branded growth where the organic flywheel does the heavy lifting. The case study that quotes ROAS without TACoS is hiding the second number for a reason.
What red flags should make you walk away?
Five patterns to walk on. (1) Multi-year contract pushed in the first call. (2) Reluctance to provide client references in your category. (3) ROAS-only metrics with no TACoS or contribution margin discussion. (4) No 30/60/90 milestones in the proposal. (5) Refusal to operate transparently against agreed targets at the 60-day evaluation.
Any single red flag isn't conclusive. Two or more is reason to walk. The agencies that ship long-term client value are the ones comfortable with milestone-based renewal and honest unwind clauses. Agencies that need to lock you in for 12 to 24 months upfront are protecting against their own retention problems, which usually means their results aren't strong enough to retain clients voluntarily.
How should you structure the first 90 days?
Two-week diagnostic, written operating plan with explicit targets, then 60 to 75 days of execution against those targets, ending with a 90-day evaluation that triggers either renewal or transparent unwind. The structure protects both sides. You get accountability. The agency gets a defined window to prove value.
The agency should deliver a written Account Business Plan within 14 days of the engagement starting. The ABP should specify Buy Box trajectory, advertising baseline, contribution margin progression, and the 30/60/90 milestones. If the agency can't or won't write this down, that's its own red flag.
At the 60-day mark, both parties review against the milestones. Brands that try to skip the 60-day check usually find out at month 8 that the engagement was off track from month 3. Brands that hold the 60-day check honestly either get a real pivot or an early unwind, both of which are better than month-8 surprise.
When should a CPG brand consider 3P distribution instead of an agency retainer?
When capital is the actual bottleneck, not operations. If your unit economics are strong but you can't fund the inventory or ad spend the demand data is begging for, a 3P partnership transfers inventory and ad-spend risk to the distributor in exchange for resale-margin economics. If you have the capital and only need operational execution, an agency retainer is usually cheaper and more flexible.
We're biased here because Eleviam offers both. Honestly: most brands don't need 3P. Most brands need an operating partner who can run the marketplace cadence while the founder focuses on demand generation, brand strategy, and retail/D2C. 3P fits the 20 to 30% of brands where the unit economics are already strong but cash flow is the constraint blocking the next inflection. We turn down 3P engagements where the brand is better served by an agency engagement, which is most of the time.
Frequently asked questions
+What's the single most important question to ask any Amazon agency?
How do they get paid. Agencies billing on percentage of ad spend earn more when your budget grows, regardless of whether your brand grows. Agencies billing on percentage of gross revenue (or with skin in the game via 3P inventory) earn more only when your top line grows. The compensation structure predicts every recommendation you'll get.
+How long should an Amazon agency engagement be?
The right structure is a 90-day pilot with explicit milestones, then renewal based on whether the milestones were met. Multi-year retainer commitments without performance gates are a warning sign. Month-to-month with no commitment is a different warning sign (the agency won't invest in deep work). Pilots with milestones are the middle path that protects both sides.
+Should we hire an Amazon agency or build an internal team?
Internal teams are usually cheaper at scale (above ~$2M annual Amazon revenue) but slower to ramp and harder to staff with category-specific expertise. Agencies are faster to ramp, deeper on category and tooling, and don't carry W-2 overhead. The hybrid model (internal demand-and-brand team + external agency for marketplace operations) is what most CPG brands at $1M to $10M end up converging on.
+How do we evaluate an agency's case studies?
Real case studies name the brand (or describe the brand specifically enough that you could identify them with a Google search). Real case studies include both the starting state and the ending state with absolute numbers, not just percentages. Real case studies acknowledge what didn't work alongside what did. Anonymous percentage-only case studies are usually marketing fiction.
+What red flags should we watch for in agency proposals?
Five recurring patterns: (1) percentage of ad spend pricing without revenue gates, (2) no clear 30/60/90 day milestones in the proposal, (3) reluctance to provide direct client references, (4) focus on vanity metrics like ROAS without contribution margin or TACoS, (5) high-pressure close on a multi-year contract during the first call. Any one of these is reason to slow down. Two or more is reason to walk.
+How much should an Amazon agency cost for a CPG brand?
For brands at $75K to $300K monthly Amazon revenue, expect a $4K to $10K monthly retainer plus 2 to 5% of gross revenue (or 8 to 15% of ad spend, depending on commercial structure). Total annual cost typically lands at 6 to 10% of gross Amazon revenue. Above that, the agency is taking too much. Below 5%, the agency is probably under-resourcing your account or absorbing margin loss in ways that won't be sustainable.
+What's the biggest mistake CPG brands make when hiring an Amazon agency?
Hiring on slick sales-deck pattern matching instead of category fluency. The agencies with the polished deck and aggressive sales process are not always the agencies that operate well. Ask to see how they think about your specific category's reseller dynamics, FBA fee profile, and seasonal patterns. If the answer is generic, the execution will be generic.