Outsource the disciplines you cannot supervise and train the ones you can. Amazon search and advertising are learnable, but the tuition is paid in wasted ad spend and lost rank while your team learns. Train when you have time and volume. Outsource when the clock or the catalog is against you.
The short version
- The deciding number is how many months of underperformance you can fund. Training is cheap in salary and expensive in elapsed time.
- You cannot supervise a skill you do not have. An untrained manager hiring a trained specialist is a governance problem, not a staffing one.
- Rank lost during the learning period is not automatically recoverable. Competitors take the position and keep it.
- Hold either route to the same outcome standard. Ours is a majority of brands profitable inside their first year.
- The two routes are not exclusive. Most brands that get this right outsource execution while one internal person learns to read the work.
The failure modes, ranked by what they cost
I have watched both routes fail, and they fail in a predictable order. This table is sorted by cost, most expensive first.
| Failure mode | Where it happens | What it costs |
|---|---|---|
| Learning on a launch | Training | The launch window. Rank and review velocity are hardest to buy back |
| Nobody internally can grade the work | Outsourcing | Months of paying for activity that is not producing outcomes |
| The trained person leaves | Training | The entire investment, plus the account sits unattended |
| Optimizing the wrong lever | Both | Ad budget spent to fix a conversion problem that was never about traffic |
| Fragmented ownership | Both | Search and advertising drift apart, and neither party owns the result |
| Tool sprawl | Training | Subscriptions that duplicate each other and a stack nobody fully uses |
The first row is the one that matters most. Search and advertising are cheapest to learn on a stable, mature product where mistakes cost efficiency. They are most expensive to learn on a launch, where mistakes cost position, and position compounds.
Why the first failure is so costly
Amazon search rewards conversion history. A product that ranks and converts keeps ranking. A product that spends its launch window learning does not just lose that period, it enters a worse competitive position permanently, because the competitor who launched properly is now the one accumulating the reviews and the rank.
That is the honest argument for outsourcing during a launch specifically, even if you intend to bring the work in-house later. A full brand launch runs around seven months. That is a long time to be paying tuition on your own catalog.
How to decide, route by route
Train the team when: you have more than a handful of products, the catalog is stable, your revenue can absorb a slow quarter, and there is someone internally who wants this as a career rather than as an extra duty. In that situation the compounding works for you, because they get better every month on an account that tolerates learning.
Outsource when: you are launching, you are entering a new marketplace, your catalog is small enough that one person's downtime stops everything, or nobody internally can currently tell good work from busy work.
Do both when: you want the capability eventually. Outsource execution now, and appoint one internal owner whose job is to read the weekly update, ask hard questions on the fortnightly review, and gradually take over the parts they understand. Around two hours a month is enough to hold an outsourced team to account once onboarding is done, rising to four to six hours a week during a launch.
The standard both routes must meet
Whichever route you take, hold it to an outcome rather than an activity. At Flapen the standard I use internally is that a majority of the brands we take on are profitable within their first year. That is the number I would ask any agency for, and it is the number I would set for an internal team too.
Activity metrics will look healthy on both routes long before results arrive. Campaigns launched, keywords added, listings updated: none of those are outcomes. Ask for profitability, acquisition cost, and rank on the terms that convert.
What agencies and course sellers will not tell you
Course sellers will not tell you how much of Amazon advertising skill is account-specific rather than general. The mechanics are teachable in a week. The judgment, knowing which of six plausible explanations for a bad month is the real one, comes from having watched a lot of accounts, and you cannot get that from a curriculum.
Agencies will not tell you that a well-run internal team beats an average agency comfortably. If you have volume, patience, and a person who wants the job, building is a strong answer. The question is not which route is better in the abstract. It is which one you can actually staff and supervise.
Related answers
- Which roles to hire first for an in-house Amazon team
- Tool stack for running Amazon in-house effectively
- Best Amazon keyword research solutions
- Amazon agency vs in-house team pros and cons
- Build vs buy for your Amazon channel: the complete guide
If you want the outcome standard applied to your account first, start with the free audit at Flapen.
