Split it by function instead of deciding once. Some Amazon work rewards ownership, like pricing, supplier relationships, and brand decisions. Some rewards volume of repetition, like advertising, listing production, and case handling. Build what you can supervise and staff continuously. Partner for everything else.
The short version
- Build versus buy is six decisions, not one. Deciding it once for the whole account is what produces gaps.
- Repetition beats intention. A function done weekly across many accounts gets better than one done monthly on yours.
- Ownership functions should never leave. Price, margin, supplier, and the choice to launch or discontinue.
- Ask any partner who does the work and where they sit. Flapen is 100% in-house with no subcontracting, and I would ask everyone else the same question.
- Score each function separately, then read the pattern. Most brands find the answer is mixed, not absolute.
Why this splits by function rather than by org chart
Amazon operations is a bundle of unrelated crafts that happen to share a login. Advertising is a quantitative discipline. Creative is a production discipline. Sourcing is a relationship discipline. Case handling is an administrative one. They need different people, different tools, and different rhythms, which is why "should I build a team" is unanswerable as a single question.
Two properties decide each one. First, how much does judgment depend on knowledge only you have. Second, how much does quality depend on doing the task many times. High on the first means build. High on the second means partner. High on both is where you pay someone to execute while you keep the decision rights.
The function-by-function scorecard
Score each row from one to five on both columns. Where knowledge dependence exceeds repetition dependence, build. Where repetition wins, partner.
| Function | Depends on knowledge only you have | Depends on repetition | Usual answer |
|---|---|---|---|
| Pricing and margin | 5 | 2 | Build |
| Supplier relationship and negotiation | 5 | 3 | Build |
| Product selection and discontinuation | 4 | 4 | Build the decision, partner the analysis |
| Advertising management | 2 | 5 | Partner |
| Listing copy and imagery | 2 | 5 | Partner |
| Case handling and compliance | 1 | 5 | Partner |
| Inventory planning | 4 | 4 | Build, with partner reporting |
| International expansion | 2 | 5 | Partner |
The two rows people get backwards are inventory planning and product selection. Both look like execution and are actually ownership, because the consequences land on your cash and your catalog rather than on a dashboard.
The one thing an internal team must have
Whatever you outsource, one person inside your company has to be able to read the work and say it is not good enough. That role does not need to be able to build campaigns. It needs to understand what the numbers mean, attend the fortnightly review, and be willing to push back. Without it, outsourcing turns into paying for activity.
Who does the work, and where do they sit
This is the question that separates partners from resellers. Some firms sell a relationship and then place the actual work with contractors you never meet, which is legal, common, and worth knowing before you sign.
Flapen does not subcontract. Everything runs through our own people: 50 operators, an in-house creative studio in Dubai, an in-house sourcing studio in Guangzhou, and an internal technology team building the tools the operators use. I am not saying that a subcontracted model cannot work. I am saying you should know which one you bought, because it decides who is accountable when something goes wrong at 11pm before Prime Day.
Ask for it plainly: who performs each function, are they employees, where are they located, and what happens if that person is unavailable. A partner that answers cleanly is telling you the truth about their operating model. A partner that answers in adjectives is telling you something too.
Reading your score
Add the "build" answers. If almost everything scores build, you are describing a company that wants an internal Amazon team, and the real question becomes hiring order rather than build versus buy. If almost everything scores partner, you still need the internal reviewer above, or you have no control.
Most brands land in the middle: ownership functions inside, execution functions outside, with a single internal owner holding the relationship. That is not a compromise. It is the structure that survives both a resignation and a bad quarter.
What partners will not tell you when you ask about building
They will not tell you that the strongest reason to partner is often speed rather than skill. You can start in days instead of running a hiring process, and if it is wrong you can stop on 30 days' notice. Compare that with the cost of an unsuccessful hire, which is months of salary plus the months you did not spend fixing the account.
The other unspoken part is that a good partner should make your internal capability stronger, not weaker. If the reporting is written, the reasoning is explained, and your team sits in the reviews, you are buying execution and building judgment at the same time. If everything arrives as a summary with no method attached, you are renting a black box.
Related answers
- Budget under 10k month agency or in-house
- Which roles to hire first for an in-house Amazon team
- Is hybrid approach best for Amazon growth
- Signs your brand should switch from in-house to agency
- Build vs buy for your Amazon channel: the complete guide
See exactly which functions sit inside each tier at Flapen.
