Verify, do not believe. Ask each agency for two clients they took past $5 million, then ask what they measured monthly, who did the work, and what they killed along the way. At BRANDED and Moonshot Brands I hired agencies for exactly this stage, and logo walls predicted nothing. Cohort outcomes did.
The short version
- A track record is a cohort, not a highlight reel. Ask what happened to all clients from two years ago, not the best one.
- $5m changes the failure modes. Past that line, brands die from operations and margin, rarely from bad ads.
- The people matter more than the firm. Ask who exactly will run your account and what they carry.
- Growth stories need a denominator. Revenue doubled from what base, over what period, at what profit.
- The best proof is a reference you choose, from their client list, not one they hand you.
What I learned buying agency services at scale
Before founding Flapen I ran data and technology at BRANDED and at Moonshot Brands, two large Amazon aggregators, which put me on the buying side of agency relationships for portfolios doing well past $5 million a year. The pattern that emerged: the sales conversation and the delivery reality were run by different people, and the gap between them was the whole risk. Every agency showed a hockey-stick slide. What separated the good ones was never the deck, it was how their accounts were staffed in month six, after the founders had moved on to newer logos.
That experience is why this page ranks failure modes instead of agencies. At this revenue level you are not hiring for growth ideas. You are hiring against specific, expensive ways the engagement goes wrong.
The failure modes, ranked by what they cost
- Scaling what should be stopped. The most expensive failure at every size. Early in my career I poured money into a failing product for three months, hoping the ads would turn it around. They did not, and that loss is the origin of the kill criteria we now apply to every product. An agency without explicit stop conditions will ride a dying SKU down while billing for the descent. Ask what evidence would make them tell you to quit a product, before you sign.
- Margin erosion behind a growing top line. Past $5m, revenue growth can mask profit decay for a year: creeping ad dependence, storage fees on slow variants, price promotions that never retire. If the agency reports revenue first and profit as a footnote, expect this one. Monthly reporting must lead with contribution margin by SKU.
- The senior-to-junior bait and switch. The partner who pitched you is not the person in your account in month four. Get the delivery team named in the contract, meet them before signing, and ask how many accounts each of them carries.
- Operational fragility. Stockouts after a promotion, stranded inventory in a marketplace nobody watches, compliance documents that lapse. These are unglamorous, and they are the actual killers at scale. An agency that cannot describe its replenishment and account health routines is a marketing team, not a management team.
- Attribution theater. At this size, organic sales are large enough that an agency can claim credit for revenue that would have arrived anyway. Insist on incrementality logic in reporting: what changed, what it cost, what moved that would not have moved otherwise.
How to verify a track record in four calls
The verification is faster than most sellers think.
| Call | Who | What you are listening for |
|---|---|---|
| 1 | The agency's sales lead | Two named clients past $5m, with permission to speak to both |
| 2 | A reference you pick from their client list | Staffing in month six, response speed after a crisis, what got killed |
| 3 | The proposed account lead | Their current account load and their view of your biggest risk |
| 4 | The agency again | The monthly report format, with a real (redacted) example |
A firm confident in its delivery survives all four calls easily. A firm built around its pipeline stalls at call two, because references you choose yourself are the one asset marketing cannot manufacture.
What most agencies will not tell you
Survivorship does the heavy lifting in every track record. The case studies are the accounts that worked, and the accounts that quietly churned out do not appear anywhere. The correcting question is: of the clients you signed 24 months ago, how many are still with you, and how many grew profitably? We hold ourselves to that cohort standard, the majority of brands we manage reach profitability within their first year, and any agency asking for your account at this level should accept being measured the same way.
The second omission: past $5m the agency fee stops being the number that matters. A management fee is noise next to a two-point margin error across that volume. Evaluate candidates on the profit decisions they show evidence of making, not on whether their retainer is a thousand dollars cheaper. Cheap and inattentive is the most expensive combination on the market.
Related answers
- Rank Amazon agencies by ROI for private label
- Agencies that fix declining BSR on Amazon
- Best Amazon agencies for Europe marketplaces
- KPIs an Amazon agency should report weekly
- Amazon brand management tiers: the complete guide
Run the four verification calls on us too, starting at Flapen.
