How to Evaluate Amazon Agencies: 20 Questions to Ask Before Hiring

Knowing how to evaluate Amazon agencies comes down to asking questions the sales deck wasn’t built to answer. Every agency pitch covers the same ground — growth stories, logos, a team slide — and none of it predicts what your account will look like in month six. What does predict it: who actually touches your account, how the agency measures its own work, what happens when Amazon suspends a listing, and what the contract says about leaving. This guide gives you the 20-question interview framework we’d use to evaluate any agency, including ourselves, grouped into five topics, plus a scorecard for comparing finalists side by side. It pairs with the agency red flags list — that page is about elimination; this one is about selection.

Run the Evaluation as a Process, Not a Gut Check

Before the questions, three ground rules.

Interview two or three finalists, not six. A real evaluation takes several hours per agency. Use the broader hiring an Amazon agency guide and a pricing sanity check to get to a shortlist first, then go deep.

Ask every finalist the same questions in the same order. The scorecard at the bottom of this page only works if the answers are comparable. Free-flowing conversations favor the best salesperson, not the best operator.

Grant read-only access and demand a real audit. Add each finalist as a view-only user in Seller Central and ask for three account-specific findings with numbers. The quality of that audit is the best preview of the quality of their monthly work. If you want all of this in a printable format, the Amazon agency evaluation checklist packages the full question set.

Now the questions.

Team Structure: Who Actually Works My Account

1. Who will run my account day to day — by name — and can I meet them before signing? The strategist on the sales call is often not the person who’ll touch your campaigns. A good answer names the account manager and puts them on a call before the contract. A bad answer is “we’ll assign the best fit at onboarding,” which means you’re buying a queue.

2. How many accounts does that person manage? This is the single best predictor of attention. Somewhere in the range of 6–10 accounts per manager allows real work; 20–30 means triage and templated reports. Ask for the number, not the philosophy.

3. What’s your account manager turnover, and what happens to my account when mine leaves? Agency account managers churn — that’s the industry. What matters is the handoff: documented account playbooks, overlap periods, and a second person who already knows your account. If the agency claims nobody ever leaves, discount everything else they say.

4. Who does the execution work — bids, flat files, cases — and where do they sit? Offshore execution is normal and keeps pricing sane. What you’re testing is honesty about the org chart: who makes strategic decisions versus who runs the playbook, and whether the senior person from the pitch will ever look at your account again after onboarding.

Reporting and Attribution: How They Grade Themselves

5. Do you optimize to ACoS or TACoS, and why? The answer reveals whether the agency thinks about your business or just its own ad campaigns. Optimizing purely to ACoS lets an agency look brilliant while shifting sales that would have happened organically into paid. The right answer centers total advertising cost against total revenue — the full logic is in our breakdown of ACoS versus TACoS.

6. How do you handle incrementality — how do you know your ads drove sales that wouldn’t have happened anyway? This is the honesty question. Amazon’s attribution gives ads credit for plenty of sales that were coming regardless, especially branded search. A strong answer discusses branded versus non-branded segmentation, TACoS trend lines, and tools like Amazon Marketing Cloud for overlap analysis. A weak answer is a dashboard screenshot. The best answer includes the words “we can’t measure it perfectly.”

7. Show me an actual monthly report from a current client, anonymized. Don’t ask what reporting includes — ask to see it. You’re looking for numbers tied to actions taken and actions planned: wasted spend identified and negated, specific search terms promoted to exact match, a listing test launched with a hypothesis. Charts without decisions attached are decoration, and a report that could describe any account describes no account.

8. What KPIs will you commit to reviewing every month, and what triggers a strategy change? Good agencies operate on thresholds: if TACoS drifts above target for 60 days, if conversion rate drops X% after a listing change, a defined response follows. If there are no thresholds, there’s no system — just vibes and monthly calls.

Compliance Experience: When Amazon Breaks Something

9. How many account suspensions and ASIN reinstatements have you handled in the last 12 months? Suspensions, listing takedowns, and inauthentic complaints are operating conditions on Amazon. An agency that claims its clients never have compliance problems either has very few clients or isn’t telling you the truth. You want a number and a couple of war stories with specifics.

10. Walk me through your plan-of-action process. A real POA process has structure: root-cause analysis, documentation gathering (invoices, supply chain records, test reports), a written corrective-action narrative, and an escalation path when the first appeal bounces. If the answer is “we open a case with seller support,” your account will sit suppressed for weeks when it matters. This is the capability behind our compliance and reinstatement service, and any full-service agency should be able to describe an equivalent.

11. Do you audit listings for compliance risk proactively, or only react to enforcement? Restricted claims in supplement copy, missing hazmat documentation, pesticide flags on cleaning products — most enforcement is predictable from the listing content itself. Proactive agencies review claims and documentation before Amazon does.

12. Have you worked in my category, and what category-specific enforcement have you seen? Compliance is category-specific. Topicals get inauthentic complaints; toys need children’s product certificates and test reports; supplements draw claim reviews and random document requests; anything with batteries or aerosols hits hazmat review. An agency that’s operated in your category can name the traps before you fall into them — and will ask you documentation questions during the sales process, which is itself a good sign.

Commercial Terms: The Contract You’ll Live With

13. What’s the contract length, and what’s the exit? The fair structure: an initial 90-day commitment — real work genuinely takes a quarter to show — then 30-day notice. Push back hard on 12-month lock-ins with no performance out-clause and on auto-renewal language. How the agency reacts to this question is itself data.

14. Who owns the ad account, the campaign history, and the data when we part ways? Everything should live in accounts you own, with the agency as a removable user. Campaign history — harvested search terms, negative keyword lists, audience data — is the compounding asset of PPC management. If it lives in the agency’s account or tool, leaving them costs you months of accumulated optimization. Get ownership in the contract.

15. Who owns the IP of creative you produce — A+ Content, storefront design, photography, video? Some agencies license creative rather than transfer it, which means your own product images become leverage at renewal time. The answer you want: all creative produced under the retainer is work for hire, transferred to you, with source files delivered on exit.

16. What exactly is in scope for the retainer, and what triggers additional fees? Get the deliverables schedule in writing with cadence attached — search term harvesting weekly, listing tests quarterly, whatever the actual system is. Then ask what common needs fall outside scope: reinstatements, new product launches, DSP, international marketplaces. Surprise fees live in that gap. For market context on what retainers should cost, see our breakdown of Amazon agency pricing.

17. Is any compensation tied to my ad spend? Percentage-of-spend pricing creates a structural incentive to grow spend regardless of efficiency. It’s workable when paired with TACoS accountability, but you need to know the incentive exists. Flat retainers and performance components tied to profitable growth align better.

References and Proof: Verify Before You Sign

18. Show me two case studies with numbers from brands near my size and category. Real case studies specify starting condition, intervention, timeframe, and measured outcome — the format looks like “TACoS from 28% to 11% in 90 days,” not “significant growth for a leading brand.” Then pressure-test one: ask what the mechanism was. Teams that did the work explain it fluently.

19. Give me two current-client references — and one client who left. Current references will be cherry-picked; ask them specific questions anyway: response time on urgent issues, whether the team changed after signing, what the agency is weakest at. The departed-client request is the revealing one. An agency that offers a gracious answer about a departure — or actually provides the contact — is confident in its record.

20. What kind of client do you turn away? Agencies with a real operating model know who they’re wrong for — categories they don’t know, brands too small for the retainer math, sellers who want daily approval on every bid. “We can help anyone” means a sales-driven shop that will take your money while learning your category.

The Finalist Scorecard

Score each finalist 1–5 per row while the calls are fresh. Weight the rows however you like, but don’t skip any — a 5 on reporting doesn’t offset a 1 on data ownership.

Two patterns to watch. A finalist that scores high everywhere except commercial terms is a good operator with a bad contract — negotiate, because the terms are usually more flexible than the pitch implies. A finalist that scores high on the sales-facing rows (case studies, audit polish) but low on the operational rows (team structure, thresholds, compliance) is a marketing organization, not an operations one. That gap is exactly what the red flags list is designed to catch.

What a Good Evaluation Feels Like on the Other Side

Here’s the tell that you’ve found the right agency: the questions don’t make them defensive. Strong operators enjoy this interview, because the answers are their differentiation — the named team, the documented deliverables, the TACoS thresholds, the reinstatement war stories, the contract that lets you leave.

If you’re evaluating full-service Amazon management, run us through all 20. Ask for the deliverables schedule, the exit terms, and the case studies with numbers, and score us on the same sheet as everyone else. We built this framework because we’re comfortable with where we land on it.