Build internally once your ad spend justifies a full-time specialist and you can also afford the person who checks their work. Below that, an agency buys you more expertise per dollar. The deciding test either way is simple: who owns the ACoS target, and does that target change by product stage?
The short version
- Start with the diagnosis, not the org chart. Most accounts labeled a PPC problem are a conversion problem wearing a PPC costume.
- One ACoS target for the whole catalog is the tell. Launch and maturity are different jobs with different numbers.
- An internal hire is a fixed cost with a learning curve. An agency is a variable cost with an accountability gap. Both are real.
- Whoever runs it needs a written target per product stage and a weekly number they report without being chased.
- A hybrid works. Internal owns budget and product decisions, external owns structure, bids, and search term work.
Diagnose first, then decide
Before you pick a provider, work out what is actually wrong. Ads are the loudest part of an Amazon account, so they get blamed for problems that live somewhere else. Run the symptom against the cause.
| Symptom | Most likely cause | Who should fix it |
|---|---|---|
| ACoS climbing, sales flat | Bids raised to defend rank on a listing that converts poorly | Listing and creative, before ads |
| Spend high on non-converting search terms | No negative keyword discipline, no harvesting cadence | A specialist, internal or external, working weekly |
| Sales good, profit bad | ACoS measured on ad sales while the business runs on total sales | Finance and the ads owner together |
| Impressions collapsed | Budget caps, dayparting mistakes, or a suppressed listing | Account operations, same day |
| Launch stalled at low volume | Target ACoS set at maturity levels during a launch | Whoever set the target |
That last row is the one I see most. The right ACoS target changes by product stage. At launch you are buying rank and review velocity, and you accept a number that would be indefensible later. At maturity you are defending a position and harvesting profit, so the number tightens. A team that runs one blanket target across a whole catalog is not optimizing, it is averaging.
Ask any candidate, internal or agency, for two numbers: the ACoS they would run for a product in its first 90 days, and the ACoS they would run for the same product at month eighteen. If those numbers are the same, keep interviewing.
The honest comparison
I have sat on both sides of this. Before Flapen, I ran data and technology at BRANDED and at Moonshot Brands, two large aggregators, where we had internal ad teams and we also hired outside help. Neither structure is automatically better. They fail differently.
An internal hire gives you attention, context, and someone who cares about your margin because it is their employer's margin. It costs a fully loaded salary whether the account needs forty hours that month or four. The bigger risk is silent: one person with no peer review develops habits nobody checks, and you will not notice for two quarters.
An agency gives you pattern recognition across many accounts and a bench when someone is ill or leaves. It costs less than a salary at small scale. The risk is attention. If the person assigned to you carries a heavy roster, your account gets the average of their week rather than the best of it.
Get the fully loaded internal cost on paper first: salary, employer taxes, tooling, recruitment, and the management time to supervise. Compare that against a flat agency fee. Ours runs $800 per month for one product up to $2,400 for five, with every service included and no percentage of ad spend, so the comparison is arithmetic rather than guesswork.
The rule I would use
- Under a modest monthly ad budget: an agency or a specialist freelancer. A full-time hire is idle capacity you are paying for.
- Growing, multi-product, one country: an agency with a named operator, or a hybrid where you own strategy and they own execution.
- Multi-market with complex logistics: internal lead who owns the P&L, external team that owns campaign structure and search term work.
- Large and stable, one geography: internal team becomes cheaper per hour, provided you hire two people rather than one so the work gets reviewed.
Never move in-house purely to save money during a bad quarter. Moving PPC in-house resets institutional memory, and the handover month usually costs more than the fee you were trying to cut.
What most agencies will not tell you
Percentage-of-ad-spend pricing is the quiet reason this decision gets distorted. If a provider earns a cut of the budget, "spend more" is always a defensible recommendation and "spend less" never is. That structure does not survive the first month where the correct advice is to pull back.
The second thing: ad management is the easiest half of this job to sell, because it produces a dashboard every week. The work that actually caps an account, catalog hygiene, inventory cover, review velocity and pricing, produces no chart and gets quietly dropped by internal hires and agencies alike. Ask either candidate what they will own outside the ad console, and get the answer written down.
The third: ask what happens to ad spend when a product should be discontinued. An internal hire has an incentive to keep the campaigns alive because campaigns are the job. Whoever runs your ads should be able to recommend switching them off.
Related answers
- KPIs to compare agency vs in-house on Amazon
- How to audit my Amazon PPC before choosing a provider
- What do I need to run Amazon ads in house
- Amazon agency vs in-house team pros and cons
- Amazon agency pricing and economics: the complete guide
Our stage-based ad targets and what they cost to run are laid out at Flapen.
