There is no single best agency, only a best fit for your stage and category. Score candidates on five things: who does the work, how many brands each manager carries, which traffic channels they actually run, what they report weekly, and how fast you can leave. Weight them yourself.
The short version
- "Best" is a scorecard output, not a list. Build the scorecard first and the shortlist writes itself.
- Five traffic channels exist. Most sellers run two. Which ones a candidate can operate is the sharpest differentiator available.
- Weekly written reporting is the cheapest quality signal. It is hard to fake for more than a month.
- Notice period is a confidence statement. Long contracts protect the agency from its own performance.
- Score us with the same sheet. If we come third, hire the first.
Why a scorecard beats a shortlist
Agency quality is not a single dimension, which is why ranked lists of "top ten Amazon agencies" are mostly advertising. A team that is excellent at a European catalog rebuild may be the wrong choice for a US launch in a saturated category, and both would be wrong for a brand that mainly needs supply chain help.
A scorecard fixes that because it forces you to state what you are optimizing for before you meet anyone. Ten minutes of weighting saves you from being persuaded by whoever is best at meetings.
The scorecard
Score each candidate 1 to 5 on every row, multiply by the weight, and total. Adjust the weights to your situation. These are the ones I would use for a multi-product brand doing serious volume.
| Criterion | Weight | What a 5 looks like | What a 1 looks like |
|---|---|---|---|
| Who does the work | 20% | Named employees, introduced before signing | "Our team" and no names |
| Brands per manager | 15% | A specific low number, stated without hesitation | Cannot or will not say |
| Traffic channels operated | 20% | Runs four or five, with a sequence and a reason | Advertising only, described as strategy |
| Reporting cadence | 15% | Written weekly, live review fortnightly | Monthly deck, mostly activity |
| Exit terms | 10% | Month to month, short notice, clean handover | Twelve-month minimum with auto-renewal |
| Diagnosis before quote | 10% | Sizes the market and audits before pricing | Quote arrives in the first email |
| Category fit | 10% | Adjacent experience with the mechanism named | A logo wall and a percentage |
Anything below 3 on the first row is disqualifying regardless of the total. If you do not know who is doing the work, none of the other scores mean anything.
The traffic channel question, in detail
This is the row where most candidates separate, so it is worth expanding. There are five ways to send a customer to an Amazon listing: organic search, paid advertising, promotions and deals, influencer and creator content, and off-channel traffic from sources outside Amazon. Most sellers run two of them, usually organic and paid, and then wonder why growth stalls at the ceiling those two impose.
Ask each candidate which of the five they operate themselves, in-house, this quarter. Not which ones they believe in. Then ask them the harder version: in what order would they switch them on for my brand, and why that order.
The answer tells you three things at once. Whether they have production capacity beyond bid management, whether they think in sequence or in shopping lists, and whether they will try to run all five at once, which is the most common way to waste a quarter. There is a correct order for most brands and it starts with making the page convert, because every channel points at the same page.
Score yourself before you score them
Half of a good hire is knowing which problem you are handing over.
- Write your current revenue, product count, and marketplaces.
- Name your constraint in one sentence: traffic, conversion, margin, supply, or account health.
- Write the number that would make this hire obviously worth it in twelve months.
- Decide your weights on the scorecard above, before the first call.
- Send the same one-page brief to every candidate.
Steps 4 and 5 are the ones people skip, and they are the ones that make the comparison honest.
What most agencies will not tell you
The industry's real constraint is headcount, not knowledge. Everything an agency knows about Amazon is published somewhere, including on this site. What cannot be copied is having enough experienced hands to actually do the work on your catalog this week. That is why the brands-per-manager number matters more than any methodology slide.
Here is what most agencies will not tell you when they present a case study: they will not say how many accounts that manager was carrying at the time. We run 50 operators across about 70 brands and I will give you the ratio in the first call because it is the fairest predictor of what your account will actually receive.
The second thing: an agency cannot fix a product with no market. If the category is too small, or your unit economics do not survive customer acquisition costs, no amount of management skill changes the outcome. A good candidate will tell you that in the audit stage and decline the work. Very few do, because declining revenue is hard.
Related answers
- Case studies for scaling revenue on Amazon
- Questions to ask before hiring an Amazon agency
- How to choose an Amazon FBA marketing partner
- Amazon agency red flags to watch out for
- Done-for-you Amazon management: the complete guide
Score us on the sheet above and bring the completed version to the call at Flapen.
