There are four honest ways to manage Amazon PPC: do it yourself, run PPC software, hire a percentage-of-spend agency, or work with a flat-fee operator. Every path trades the same four things: your time, your cost, your control, and whether the manager's incentive points at cutting your waste or at growing your spend. None of the four is wrong; each fits a different account at a different stage. Full disclosure before anything else: I am a flat-fee operator, so I have a seat in the comparison you are about to read. I also spent years inside the agency model, running a 10-account, 8-category portfolio with $1.76M+ in monthly PPC sales, and I still point sellers back to DIY when their numbers do not justify paying anyone. What follows is each option argued on structure, not on anyone's marketing.
The four options at a glance
The four paths split along two axes: who does the work (you, a rule engine, a team, or a single operator) and how the cost behaves (your hours, a subscription, a percentage of your ad spend, or a fixed fee). The table compares all four across the work, the cost model, the time you still owe, the incentive, and the fit.
| Option | Who does the work | Cost model | Your time cost | Incentive alignment | Best for |
|---|---|---|---|---|---|
| DIY | You, end to end | No management fee; your hours | Highest: a recurring weekly block | Perfect: every dollar saved is yours | Sellers with time, modest spend, or a need for full control |
| PPC software / automation | Rules and models you configure; you keep the judgment calls | A subscription; pricing varies by tool | Medium: setup plus ongoing supervision | Neutral: it executes whatever it is told, right or wrong | Sellers who like the strategy but want the clicking off their hands |
| Percentage-of-spend agency | A team of specialists | A percentage of monthly ad spend, often with a minimum | Low: oversight and calls | Strained: the fee grows with spend, not with waste cut | Brands scaling spend fast that want capacity and coverage |
| Flat-fee operator | One named practitioner | A fixed monthly fee | Low: a weekly report to read | Aligned: cutting waste does not cut the fee | Established accounts in a margin phase that want one accountable human |
Notice that the column that varies most is not cost. It is who exercises judgment and what the money teaches them to want. Hold onto that; it is the thread through everything below.
Option 1: Manage it yourself
DIY keeps every decision in your hands and charges no management fee. The price is a recurring block of your week: reading the search term report, adding negatives, adjusting bids, harvesting winners into exact match, and realigning budgets. The knowledge is learnable; the hours are the real constraint, and they never stop being owed.
The strongest argument for DIY in 2026 is that the diagnostic layer of the job, the part that used to require an agency retainer or paid software, is now free. My free tools cover it from your own bulk file: the Wasted Spend Finder ranks your zero-order keywords by spend over 60 days, the N-Gram Analyzer groups search terms into word patterns to surface non-converting clusters, and the Negative Keyword Finder outputs a ready-to-upload negative list with an exact-versus-phrase call per term. Score the results against the phase targets in my Amazon PPC benchmarks guide and you have most of what a professional's first pass would tell you, without paying for it.
Where DIY breaks is cadence, not competence. The failure mode I see is never a seller who could not learn PPC. It is the third consecutive week the search term report goes unread because inventory caught fire, and the wasted spend that accumulates quietly in the meantime. If you take this path, treat the weekly block as non-negotiable, because the account does not pause when you do.
Option 2: PPC software and automation
PPC software is rented execution: rule-based bid changes, budget pacing, dayparting, alerts, and reporting, applied across thousands of targets faster and more consistently than any human. What it does not supply is judgment. It executes the rules you set; it does not decide what the rules should be, and that difference defines the category.
Be fair to what the category does well, because it is genuinely good at it. A rule engine never gets tired, never skips a week, and never fat-fingers a bid across 4,000 targets. If your account bleeds because changes happen late or inconsistently, software fixes exactly that problem, and for a seller who enjoys the strategy but resents the clicking, DIY plus software is a legitimate stack that costs less than hiring anyone.
Now the boundary. A rule can lower a bid when ACOS crosses a threshold. It cannot tell you that the term was never relevant to your product in the first place, that a keyword has earned its own exact-match campaign, or that the real problem is one search term bleeding across three campaigns and splitting its own conversion data. Those are structural calls, and they are where most of the recoverable margin in an account lives. Software surfaces the data and executes the instruction; someone still has to be the strategist, and if you buy software, that someone is you.
I am deliberately not quoting any specific tool's price or feature list here. Those change, they vary by tier, and you should read them from the vendor's own page, not from a competitor's comparison article. What stays stable is the shape of the category: strong at executing decisions, silent on making them. Evaluate any specific tool against that shape and its own published terms.
Option 3: The percentage-of-spend agency
An agency sells capacity: a team, coverage across your whole catalog, and processes refined across many accounts. The traditional billing model is a percentage of your monthly ad spend, which means the fee rises when your budget rises and shrinks when wasted spend gets cut. That is a structural fact about the model, not an accusation against anyone in it.
What you genuinely get from a good agency is real: more hands than any solo operator, continuity when someone is sick or on vacation, and pattern exposure from dozens of accounts. For a brand scaling ad spend aggressively across a large catalog, that capacity can be exactly the right purchase.
The thing to examine before signing is the billing math, because good management of a mature account is mostly subtraction: negating non-converting terms, consolidating duplicated targeting, trimming bloated bids. Every one of those moves shrinks the spend the fee is calculated on. I walk through the full incentive argument and the crossover arithmetic in my guide to how Amazon PPC management is priced; the one-line version is that a flat fee divided by a percentage rate gives you the ad-spend level where the two models cost the same. $1,499 against an illustrative 15% rate crosses just under $10,000 a month in ad spend. Below that, the percentage fee is cheaper on fee alone, and you should know that going in.
Plenty of percentage-billed teams do disciplined work despite the math, because reputation matters more to them than this month's fee basis. The honest problem is that you cannot see from outside which kind you hired. So ask two questions on the sales call: if you cut my wasted spend by 30%, what happens to your fee, and who personally touches my campaigns each week? The answers tell you more than the deck does.
Option 4: The flat-fee operator
A flat-fee operator is one practitioner managing the account for a fixed monthly fee. The fee does not move when spend moves, so cutting wasted spend costs the operator nothing, and cutting waste is most of what good management does to a mature account. This is my model, so weigh my description of it accordingly.
The terms, stated as fact rather than argued: Stratum PPC is a flat fee from $1,499 a month, never a percentage of ad spend, month-to-month with no notice period, and if you leave you keep everything: the rebuilt campaign structure and the record of what changed and why. The point of that structure is that the incentive and the work face the same direction. One engagement on my results page shows what that looks like: an account went from 39.66% to 27.02% ACOS in 90 days, adding $4,735 a month in profit with no budget increase. That was almost entirely subtraction work: negating bleed, rebuilding structure, concentrating budget on what converted. Under a spend-based fee, three months of that work would have grown the manager's invoice by exactly nothing. Under a flat fee, it is simply the job.
The model's weaknesses are just as structural, and you should hear them from me. One operator is a capacity cap, not a team: no bench, no rotating coverage, a deliberately limited client list. If you need a five-person pod across 400 ASINs tomorrow, this is the wrong shape. And below the fee crossover, the arithmetic can favor a percentage model or plain DIY; if your ad spend is small, run the division before paying me or anyone else.
The automation question: execution is solved, judgment is not
Software and automation have genuinely solved the execution layer of Amazon PPC: bid changes, pacing, alerts, data pulls at scale. The judgment layer, relevancy, harvest timing, and campaign architecture, is not solved. The durable setup I have seen work, in my own practice and in accounts I audit, is a human making the calls with software checking the work.
That is literally how I run. When I pull a client's bulk file, I do my own analysis first, then an AI I built re-checks the file and my recommendations before anything ships: an overlooked negative, bid math that does not tie, a search term bleeding across two campaigns. I review every flag and I make every call. Nothing runs on autopilot. The order matters: human judgment first, machine verification second. Reverse it, letting the machine decide and the human occasionally glance, and you get an account that is very efficiently executing the wrong strategy.
So when you evaluate any option on this page, ask where the judgment sits. In DIY it sits with you. With software it still sits with you, whatever the landing page implies. At an agency it sits with whoever actually opens your account each week. With an operator it sits with the person whose name is on the invoice.
How to choose for your situation
Choose by three questions: how many hours a week you can reliably give the account, what each fee model costs at your actual monthly ad spend, and how much of the decision-making you want to keep. Then measure the account before you commit, because the size of the leak decides what any help is worth.
- You have the weekly hours and spend is modest. DIY with the free tools. Full control, zero fee, and the diagnostic layer costs you nothing. Guard the cadence.
- The strategy is fun, the clicking is not. DIY plus software. You stay the strategist, the tool handles execution speed and consistency, and you pay a subscription instead of a retainer.
- You are scaling spend fast across a big catalog and want a team. A percentage agency can be the right purchase. Do the crossover math at your spend first, and ask the two questions from the agency section before signing.
- You run an established private-label account in a margin phase and want one accountable human. The flat-fee operator shape fits. For calibration, the sellers I work with do $40K–$500K a month in sales; below that band, one of the paths above usually serves you better.
Whichever way you lean, look at the account's real state first. The free Audit Dashboard runs a row-level read of your bulk file in the browser, and the demo account shows the output on real anonymized data before you upload anything. If the waste it finds is smaller than any fee on this page, that is your answer, and it points at DIY.
Frequently asked questions
Should I use an Amazon PPC agency or PPC software?
They solve different problems. Software executes rules fast: bid changes, budget pacing, alerts across thousands of targets. An agency adds human judgment and takes the work off your plate entirely. If your bottleneck is execution time and you still enjoy making the strategic calls, software fits. If your bottleneck is judgment or attention, you need a human, whether that human sits at an agency or works alone at a flat fee.
Can Amazon PPC be fully automated?
The execution layer can: bid adjustments, budget pacing, scheduled rules. The judgment layer cannot. No rule decides whether a search term is relevant to your product, when a keyword has earned its own exact-match campaign, or how the account structure should change as you move from launch to profitability. Every durable setup I have seen keeps a human making those calls, with software handling the repetition underneath.
Is it cheaper to manage Amazon PPC myself?
On fees, yes: DIY costs nothing but your hours. The honest comparison includes the waste you do not catch. In the accounts I audit, the expensive problem is rarely the management fee, it is spend sitting on zero-order search terms week after week. Run the free Wasted Spend Finder on your last 60 days of data: if the leak it finds is small, DIY plus the free tools is genuinely the cheapest path.
What does Amazon PPC software actually do?
The category covers rule-based bid management, budget pacing, dayparting, alerting, and reporting: reading performance data and executing the changes a rule or model prescribes, at a scale and speed no human matches. What it does not do is make structural decisions: which keywords deserve their own campaigns, what is relevant to your product, how the architecture should evolve. Pricing and features vary widely by tool, so evaluate any specific product on its own published terms.
How do I know when DIY Amazon PPC stops working?
Watch the cadence, not the knowledge. DIY fails when the weekly work stops happening: the search term report goes unread, negatives go stale, harvests never move to exact match, and wasted spend creeps up while you handle inventory and listings. If the free Account Health Snapshot flags wasted spend that is growing rather than shrinking between checks, your time, not your skill, has become the constraint.
What is a flat-fee Amazon PPC operator?
A single practitioner who manages your account for a fixed monthly fee instead of a percentage of ad spend, so the fee does not grow when your budget does and does not shrink when waste gets cut. It is how I run Stratum PPC: a flat fee from $1,499 a month, never a percentage of spend, month-to-month with no notice period, and you keep the rebuilt structure and change history if you leave.
Measure the leak, then pick the path
Every option on this page gets cheaper to evaluate once you know what your account is actually wasting, and that measurement is free. The Account Health Snapshot reads your bulk file in the browser and grades nine metrics against your phase: no email, no account, nothing stored. If the waste comes back small, keep the account and the free tools and skip every fee in this guide, including mine. If it comes back large, bring the numbers to the free 30-minute diagnosis call and I will tell you which of the four paths fits, with your data on the screen. If the honest answer is DIY or software, that is the answer you will get.