Seven signs: nobody owns advertising, listings have not changed in six months, you cannot say what changed last week, one person holds all the knowledge, expansion keeps slipping, ad spend rises while margin falls, and nobody has ever recommended killing a product. Score them. Four or more and the current setup has stopped working.
The short version
- Score the signs before you take a sales call. A vendor cannot diagnose you and should not be asked to.
- The seventh sign is the heaviest. A team that never recommends stopping is not evaluating anything.
- Two or three signs usually means one hire, not a change of model.
- Switching costs you a quarter. Budget for the ramp instead of pretending it is instant.
- Some of these are fixable in a fortnight with a written cadence and one owner per job.
Score yourself out of twenty
Give each sign 0 if it does not describe you, 1 if it partly does, and 2 if it clearly does. The weights are not equal, because these do not cost the same.
| Sign | Weight | What a clear 2 looks like |
|---|---|---|
| No owner for advertising | 3 | Bids last touched more than a month ago, no negative keywords added this quarter |
| Listings static for six months | 2 | Title, bullets, and images unchanged while competitors have all moved |
| No answer to what changed last week | 3 | No change log anywhere, and reporting is a revenue figure |
| One person holds everything | 3 | Their leave or resignation would stop the account |
| Expansion keeps slipping | 1 | A second marketplace has been next quarter for three quarters |
| Spend up, margin down | 3 | Revenue growing, contribution margin after fees and ads falling |
| Nobody has proposed stopping anything | 5 | Every product ever launched is still live, including the ones losing money |
Total the weighted score. Under 10 and you have a staffing gap, not a model problem, and one experienced hire plus a written weekly routine usually fixes it. Between 10 and 20, you are choosing between a serious internal hire and outside help. Above 20, the internal setup has stopped functioning as management and has become maintenance.
Why the last sign carries the most weight
Early on, before Flapen existed, I poured money into a failing product for three months, convinced that better advertising would turn it around. It did not. The product had a rating problem and a return rate problem, and no amount of spend fixes a product people do not want to keep. That quarter is where our kill criteria came from.
Now the rule is written down before launch: rating trend, return rate, conversion rate, and acquisition cost trajectory, each measured over a defined window, with an agreed decision at the end of it. Scale, fix, or kill.
In-house teams almost never kill a product, and it is not incompetence. The person who chose the product still works there. The person who commissioned the packaging sits two desks away. Everyone is invested, and the decision to stop feels like an accusation. An outside team has no such loyalty and can say the unpopular thing without it costing anybody a relationship. That is a real and underrated part of what you buy.
Which cuts both ways. If an outside team has never told you to stop anything either, they are managing the relationship rather than the business, and you should say so directly.
What switching actually costs
Budget two to three months of ramp. A new team has to read the account, understand your supply constraints, and rebuild campaign structure without losing the ranking history that structure carries. Anyone promising results in week one is either inheriting an easy account or changing things they have not yet understood.
Protect yourself with terms rather than promises. We work month to month on 30 days of notice, the first invoice covers the first and last month, and on exit the client keeps the Seller Central account, the campaigns, the creative, and a written handover. Onboarding runs as an audit, then a named brand manager, then a blocker list, then execution, and a measurable ACoS improvement inside 30 days is the usual first visible result.
What most agencies will not tell you
At least two of the seven signs are usually fixed without changing your model. A written change log and a weekly thirty minute review of search terms, rating, and inventory closes the visibility gap on its own. If that is your whole problem, hiring anyone is an expensive way to buy a calendar reminder.
The second thing: switching resets institutional knowledge. Your internal person knows why the blue variation was discontinued and which supplier missed a deadline in 2024. That context is valuable and it does not transfer in a handover document. Weigh it honestly rather than treating the change as pure upside.
Related answers
- Amazon agency vs in-house which is better
- Build internal Amazon ops or partner with experts
- Is hybrid approach best for Amazon growth
- KPIs an Amazon agency should report weekly
- Build vs buy for your Amazon channel: the complete guide
If you scored above ten and want a second opinion in writing before you decide anything, the audit is free at Flapen.
