The behavioral signs arrive before the numbers do. You find problems before they report them, questions get answered with reassurance instead of data, the named person keeps changing, and nothing has ever been recommended for removal. By the time revenue moves you have already lost a quarter.
The short version
- You find problems first. The clearest sign of all.
- Reassurance instead of numbers when you ask a direct question.
- The named person changes every quarter.
- No proactive bad news, ever. Real accounts generate bad news.
- Reports get longer and less specific over time.
The behavioral signs
I run Flapen with 50 operators managing about 70 brands. These show up months before any metric does, and unlike metrics they need no analysis to notice.
| Sign | What it looks like | What it means |
|---|---|---|
| You find it first | You spot a suppressed listing, they did not | Nobody is watching daily |
| Reassurance answers | "It is trending in the right direction" | They do not have the number |
| Named person churn | New account manager each quarter | Your context is rebuilt repeatedly |
| No bad news | Every update is positive | Filtering, not reporting |
| Reports lengthen | More slides, fewer numbers | Activity substituting for outcomes |
| Slow on suppression | Days to fix a suppressed listing | Caseload too high |
| Vague next steps | "Continue optimizing" | No plan |
You find it first
The single most reliable sign. If you notice a suppressed listing, a Buy Box loss, a stockout risk, or a competitor undercutting you before your partner mentions it, the monitoring is not happening.
This is about attention, not competence. Ask how many brands your named manager carries. We run about 1.4 brands per operator. Above eight, daily monitoring across every account is arithmetically impossible, and nobody will say so.
Reassurance instead of data
Ask a specific question and notice the shape of the answer. "How has our cost of customer acquisition moved over six months" should produce a number. If it produces a sentence about positive momentum, the number was not to hand, and a partner who does not have that number to hand is not managing to it.
No bad news
Every real Amazon account generates bad news. A product underperforms, a competitor launches, a supplier slips, a listing gets flagged. A partner reporting only good news is filtering.
The version of this that costs most: no product has ever been recommended for removal. Any partner on a recurring fee has a structural incentive to keep everything alive. Ask what they have recommended you stop doing in the last two quarters.
Separating underperformance from a hard market
Worth doing before you act, because they look similar from the outside.
Some things are outside a partner's control: category demand decline, a well-funded new competitor, a supply problem. What is inside their control is whether the diagnosis was early, specific, and accompanied by a proposal.
Two accounts can both show falling revenue. In one, the partner identified a demand shift in month two, showed you the data, and proposed a response. In the other, the explanation arrived in month seven after you asked. Only the second is underperformance.
What to do before deciding
- Pull four numbers yourself: cost of customer acquisition trend, organic share of revenue, conversion rate, return rate.
- Lay three monthly plans side by side and count repeated items.
- Ask three direct questions and grade the answers on specificity.
- Say what you need in the next 60 days to stay.
- Get a second opinion. Most agencies audit free.
Step four is the one people skip. A direct conversation with concrete asks often produces a reassignment or a scope change. It also tells you quickly whether the problem is fixable, because the response to being told plainly is very informative.
What most agencies will not tell you
The most common cause of underperformance is not skill, it is caseload. A capable account manager carrying twelve brands produces exactly these symptoms: slow to notice, reassurance instead of analysis, plans that repeat because nothing shipped.
That is also why it is often fixable. Ask for a reassignment or ask what the caseload is and whether it can change. A good agency will address it. An agency that treats the question as an accusation has told you the number is bad.
The second thing: partners know which clients are drifting long before those clients speak up, because it shows in meeting attendance and the tone of questions. If nobody has proactively asked whether you are getting what you expected, that silence is data too.
Related answers
- Red flags your Amazon agency is underperforming
- How to audit current Amazon agency performance
- When to switch Amazon agency
- KPIs an Amazon agency should report weekly
- Hiring an Amazon agency: the complete guide
Ask us our caseload number before anything else. It is published at Flapen.
