"I'm spending money on ads but don't know if it's working." The account produces decisions every week. Who owns them is the only question that matters. The seller then asks whether an agency or an in-house team is better.
That is the wrong question. Anyone who answers it in general is selling one of the two.
The useful question is narrower. It is how many decisions this account produces in a week, who owns each one today, and which owner disappears during a holiday week.
Build what you can supervise and staff without a gap. Buy what you cannot. The rule is ownership, not money.
The numbers behind this guide
| Claim | Figure | Captured |
|---|---|---|
| Operators running our Amazon brands | 50 | 2026-09-04 |
| Brands those operators run by hand today | About 70, which is about 1.4 for each operator | 2026-08-28 |
| Managed pricing, every service included | $800 a month for one product to $2,400 for five | Standing term, SPEC §4 |
| Your own time once an account is onboarded | About 2 hours a month, rising to 4 to 6 hours a week during a launch | Standing term, SPEC §4 |
| Ad spend below which a tool run by you beats any hire | About $1,000 a month, with no hard minimum | Standing term, SPEC §4 |
| First measurable advertising improvement after handover | Typically inside 30 days | Standing term, SPEC §4 |
Each figure is one side of a comparison. The other side is your own payroll, your own hours, and your own calendar, and only you can fill those in.
Agency, in-house, freelancer, or software: what each one is quoting for
The four options are not competing for the same job, which is why comparing their prices produces nonsense. Software reports what happened and executes rules you already wrote.
A freelancer fixes one defined function well. An agency owns the whole account, including the parts nobody scoped, like cases, creative revisions, inventory timing, translations, and the decision to stop.
An in-house hire owns it too, at the cost of one salary per discipline and a single point of failure per person. Match the option to the symptom before you look at any price.
One broken function is a freelancer problem. Several functions with nobody accountable is an agency or a hire.
Good decisions with no visibility is a software problem. Most sellers who get this wrong bought the cheapest option for a problem it was never designed to solve, then blamed the option.
There is no volume discount hiding anywhere in this choice. Amazon advertising runs on an auction, so scale buys nothing off the price of a click.
What a good operator lowers is your cost of customer acquisition. That comes from better bidding terms, higher conversion, and stopped spend on products that should be discontinued. That is true on your payroll or on ours.
- Amazon agency vs in-house which is better
- Amazon account management freelancer vs agency vs software
- Cost comparison Amazon freelancer agency software
- Do agencies get better rates on Amazon ads
Building in-house: the roles, the tools, and the real cost
If you build, hire in the order that conversion is built. A brand manager first, because listing quality and images decide whether traffic buys anything. A creative producer second, and an advertising specialist third, since advertising only amplifies the conversion that already exists.
Until the account produces a full week of decisions for each role, none of the three is a full-time job yet. A full-time salary for a part-time workload is the most common way an in-house build goes wrong. It goes wrong slowly enough that nobody notices for a year.
Price the build from four lines: salaries, software, creative production, and your own management time. Only the last line looks free.
Compare the total against a published fee for the same scope. The fee side is fixed, while the payroll side also carries recruiting, ramp, and replacement.
The stack itself is the cheap part and the part sellers argue about most. You need Seller Central, a source of market and keyword data, and a way to place and adjust bids. You need somewhere creative and listing copy gets produced, and one report a human reads every week.
The subscriptions are the small line on that budget. The operator who reads them is the large one, and without that operator the stack produces dashboards nobody opens.
- Which roles to hire first for an in-house Amazon team
- Tool stack for running Amazon in-house effectively
- Alternatives to hiring a full-time Amazon manager
- Budget under 10k month agency or in-house
Software versus people, and where the line sits
Automation executes decisions. It does not make them.
A bid rule that lowers bids when the advertising cost of sale rises will run every night. If your targets are undefined, that rule runs the wrong guess every night at scale.
Dayparting and rule-based adjustment earn their setup once spend is large enough that a few percent of waste matters. Below about $1,000 a month the arithmetic rarely gets there.
Listing tools have the same shape. They flag a missing keyword, a thin bullet, or a low-resolution image in seconds.
They cannot tell you that conversion is low because the price sits above what your rating supports. They cannot see that the primary image loses to the row of competitor tiles around it.
Tools find problems. People decide what the problem means. The order matters, because a person who starts from the tool's list fixes the wrong thing first.
A human keeps the decisions that set direction. That is the stage each product is in, the budget ceiling for that stage, and the call to stop. Everything repetitive underneath those should go to software, because a person doing rule work is a person not doing judgment work.
- Compare done-for-you Amazon vs DIY software
- Alternatives to using Amazon automation software
- Amazon listing optimization tools vs human experts
- Dayparting and PPC automation for Amazon ads
The hybrid that works, and the one that fails
Most brands above a certain size end up hybrid, and the split that holds is by ownership rather than by task. You keep pricing, inventory, supplier relationships, and final approval, because those reward context and control.
An outside team owns advertising, creative, and listings, because those reward repetition, tooling, and volume of practice across many accounts. Bought as one scope, that half of the split is what Amazon brand management covers.
Hybrids fail when the same function is split across both sides. Two owners of advertising means no owner of advertising, and every underperformance becomes a boundary dispute with a weekly email thread attached.
Write down which side owns which job before anything starts. The document is boring and it is the whole arrangement.
Training your own team is a hybrid too, and it has a tuition bill. Amazon search and advertising are learnable, but the lessons are paid for in wasted ad spend and lost rank while people learn. Train when you have time and volume, and buy when the clock or the catalog is against you.
With a small team, growth comes from removing decisions rather than adding surface area. One listing template, one advertising structure per stage, one weekly review. Then buy capacity only for the work that does not compound, which is usually creative production and daily bid management.
- Is hybrid approach best for Amazon growth
- Build internal Amazon ops or partner with experts
- Amazon SEO and ads: outsource or train team
- How to scale Amazon growth globally with limited staff
When to switch, and how to choose whoever comes next
Seven signs say the current setup has stopped working. Four or more of them mean the question is no longer whether to change, only what to change to.
- Nobody owns advertising.
- The 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.
- Nobody has ever recommended killing a product.
Choose the replacement on diagnosis rather than on promises. Give each candidate read-only access, describe your worst symptom, and watch whether they name the cause or jump to a fix. The right operator names the mechanism, shows the data behind it, and tells you what they would not touch yet.
Then compare time to first fix, not time to kickoff. Software is usable the same afternoon and changes nothing by itself.
A freelancer starts fast and narrow. An agency starts slower and covers more, and in our own onboarding measurable advertising improvement typically lands inside 30 days.
For a full brand launch, buy the whole team. A launch needs research, sourcing, creative, and advertising running at once, and a freelancer covers one of those well. Hire the freelancer when the launch is one product into a category you already sell in and the rest of the machine exists.
- Signs your brand should switch from in-house to agency
- How to choose an Amazon account manager
- Compare onboarding time: freelancer, agency, software
- Who’s better for brand launch on Amazon: agency or freelancer
What most agencies will not tell you
Four things go unsaid on both sides of the build or buy pitch.
- An agency is a caseload before it is a capability. Every service on the list is delivered by a person who also carries other accounts. We run about 70 brands here with 50 operators. Ask any provider for that ratio before you ask what they do, because a double-digit caseload per manager is a queue.
- In-house is a single point of failure per person. The hire who holds all the knowledge is one resignation from taking the channel with them. Nobody in a recruiting conversation mentions the handover file, and it is the file that decides how bad the resignation is.
- Software vendors sell the day everything is configured. Nobody mentions the month when the catalog has changed, a new marketplace has launched, and the person who wrote the rules has left. Ownership, not features, is where stacks die.
- We turn down accounts. A brand whose market or margin cannot repay a fee is better served by waiting. A provider that has never declined a seller has one filter, and it is a payment method.
Run the ownership test on us the same way you would run it on a hire. If we cannot name the person on your account, their other brands, and the day they hand everything back, do not buy.
Do this week
One thing to do this week, at no cost. Open a blank page and list every decision your Amazon account produced in the last seven days. Bids, a listing change, a reorder, a case, a price move, and a creative brief all count.
Next to each one, write the name of the person who made it. Then write the name of the person who would have made it if the first was away.
Blank names are the answer to build or buy. Whatever you choose next has to fill them.
Bring that list to us and a written audit with prioritized fixes comes back inside 48 hours at no charge from Flapen.
Keep learning
Every question in this cluster
The site lists every answer in this cluster under this heading. They are grouped by the benchmark each one turns on, from caseload and in-house delivery to stop criteria, capital, and the buyer-side view. Pick the family that matches the decision on your desk this week.
