Amazon PPC management services are one of the few marketing purchases where the buyer usually cannot tell, from the outside, whether the work being done is good. The dashboard shows ACOS. ACOS goes down. Everyone is pleased. Meanwhile the account has quietly stopped bidding on the generic queries that bring new customers, revenue has flattened, and the only thing that improved was a ratio.
That failure mode is so common it's worth building your whole evaluation around avoiding it. This guide covers what a competent Amazon ads team actually does month to month, how the three pricing models change their incentives, which numbers to hold them to, and the questions that reliably expose a shop that will just turn your bids down and call it optimisation.
What Amazon PPC management services actually cover
Strip away the deck and the work falls into five buckets.
Account and campaign architecture. Deciding how many campaigns exist, how they're segmented, and what each one is for. On Amazon the useful segmentation axes are ad type, match type, product margin band, and lifecycle stage — not product category, which tells the bidding nothing actionable. A 400-SKU catalogue does not need 400 campaigns, and the agencies that build them are usually charging by complexity.
Search term harvesting and negation. This is the engine room. Auto and broad campaigns discover queries; the search term report tells you which ones converted; profitable terms get promoted into exact-match campaigns with dedicated bids; unprofitable ones get negated at the right level. Done weekly, this compounds. Done quarterly, it doesn't.
Bid and placement management. Setting bids against a target that comes from your margins, then layering placement adjustments. Top-of-search placement on Amazon converts materially better than rest-of-search or product pages, and the placement modifier is often a bigger lever than the base bid. An account with every placement modifier sitting at zero has not been managed.
Targeting expansion. Product targeting against competitor ASINs, category targeting with brand and price refinements, Sponsored Display for retargeting and defensive placement on your own listing pages. Each has a distinct job and a distinct efficiency profile.
Listing and creative feedback. Ads send traffic; the listing converts it. A good team will tell you when the problem is your main image, your price against the competitive set, or a review count that has slipped. If your agency never raises listing issues, they are optimising around a bottleneck they've decided not to mention.
Notice what isn't on that list: reporting. Reporting is the byproduct, not the deliverable. Plenty of retainers are 80% report assembly.
The three ad types and what each is for
Sponsored Products is where the overwhelming majority of seller budget goes and where the return usually is. It's keyword and ASIN targeted, appears in search results and on product pages, and is the only ad type most sellers under about $50k monthly revenue need to think hard about.
Sponsored Brands puts your logo, a headline, and a product set above the search results, or runs a video unit inside results. It requires Brand Registry. Its real value is the new-to-brand metric, which tells you what proportion of orders came from shoppers who hadn't bought from you in the previous year. That is the closest thing Amazon gives you to a customer acquisition figure, and it's the number that should justify a lower efficiency target on these campaigns.
Sponsored Display targets audiences and products both on and off Amazon. It is the easiest of the three to waste money on and the easiest to defend a listing with. Used well, it's a retargeting and defensive tool. Used badly, it's a broad audience blast with a flattering last-click attribution window.
Amazon DSP sits above all three and is a genuinely different product with its own minimums and its own agency skill set. If a prospective partner pitches DSP to a seller spending $8,000 a month on Sponsored Products, they are selling capability rather than solving your problem.
Get your break-even ACOS before you get quotes
You cannot brief an agency, let alone judge one, without knowing the advertising cost of sale at which a unit stops making you money. On Amazon this is unusually fiddly because the fees stack: referral fee, FBA fulfilment fee, storage, returns processing, and any coupon or deal funding sitting on top.
Break-even ACOS equals your contribution margin as a percentage of sale price, after all of those. Work it out per product, not blended. Use the profit margin calculator to get the true margin once fees come out, then flip it: a product carrying a 28% contribution margin breaks even at 28% ACOS, and every point below that is profit. If you think in return on ad spend instead, the ROAS calculator gives you the same picture from the other side — 28% ACOS is roughly a 3.6x return.
Three adjustments matter enough to make explicitly:
- Returns. A category returning at 20% is not earning the ACOS Amazon reports. Effective revenue is 80% of reported revenue, and the break-even threshold tightens accordingly.
- Repeat purchase. Consumables and replenishables justify a first-order ACOS well above break-even. One-off durables usually do not. This is where new-to-brand data earns its keep.
- Launch periods. Deliberately unprofitable spend during a launch buys sales velocity, which buys organic rank. That is a real investment with a real payback, but it should be a decision with an end date, not a permanent state. Model how long you can carry it with the break-even calculator before you commit.
Bring those numbers to the first agency call. The response tells you a great deal. A good team will interrogate them. A weak one will accept whatever you say and quote a target ACOS on the spot.
Pricing models, and the incentives hiding inside them
Percentage of ad spend. Typically in the low-to-mid teens, often with a monthly minimum. It's the market default and it's transparent, but the incentive points the wrong way: the agency earns more when you spend more, regardless of whether that spend earns. Workable if the contract pairs it with an efficiency floor.
Flat monthly retainer. Predictable for both sides, and the incentive is neutral rather than perverse. The risk is the opposite one — a fixed fee rewards doing less work, so retainers need a defined scope: how many campaigns, how often search terms are harvested, what the reporting cadence is, who is actually in the account.
Percentage of attributed revenue or profit share. Aligns interests best in theory. In practice the argument is always about attribution. Amazon's default windows will credit ad-driven revenue that would have happened anyway, and the share of sales that are genuinely incremental is the thing nobody can agree on. If you go this route, define the attribution basis in writing before signing, not after the first invoice.
Whatever the model, two contract terms matter more than the rate. First, you own the Amazon advertising account, because it lives inside your Seller Central and the agency gets user permissions on it. Confirm that in writing anyway, and make sure the permissions are granted to their users rather than the account being handed over wholesale. Second, agree what happens to campaign structure, negative keyword lists, and bulk files at the end of the relationship. Those artefacts are the value you paid for. The same logic applies across paid channels, and it's covered in more depth in our guide to Google Ads management services.
The metrics to hold them to
ACOS alone is not a performance metric. It's a ratio, and any account can improve it by shutting off the top of the funnel. Never review it without volume alongside.
TACOS — total advertising cost of sale — is the better headline number. It measures ad spend against total revenue including organic. If TACOS is falling while total revenue grows, ads are buying organic rank and the flywheel is working. If TACOS is flat while ad-attributed revenue climbs, you're mostly buying sales you'd have had anyway.
New-to-brand orders and their cost. The single most useful number for anyone building a brand on Amazon rather than clearing stock. It converts Amazon advertising into a customer acquisition cost you can compare against every other channel. Work out what you can afford to pay with the customer acquisition cost calculator, and read our breakdown of how to think about customer acquisition cost if you want the wider framing.
Organic rank on your priority keywords. Track it independently of the ad platform. Advertising that lifts organic position is doing something a report on ad-attributed revenue will never show you.
Share of voice on branded search. Competitors bidding on your brand name is a tax you either pay or lose sales to. Defensive campaigns look inefficient in isolation and are usually worth running anyway.
Ask for these five in a single monthly view. An agency that can't produce TACOS and new-to-brand cost without a week of notice is not looking at them.
Red flags worth walking away over
- A target ACOS quoted before they've seen your margins. Efficiency targets are a function of your economics, not their template.
- No access to your search term reports during the pitch. They should want to see them. If they don't ask, they aren't planning to work at that level.
- Automation software presented as the strategy. Rules-based bid tools are fine and most good teams use one. The tool is not the service.
- Refusal to discuss listing quality. Traffic into a weak listing is the most expensive mistake in the account, and it's usually the client's fault, which is exactly why an agency avoiding the conversation is a bad sign.
- Reporting that only shows ad-attributed revenue. Total revenue, TACOS, and organic rank are what tell you whether the business improved.
- Junior-only account handling. Ask who is in the account weekly, by name, and how many accounts that person carries. Twenty is a lot. Forty is a content farm.
When to keep it in-house instead
Below roughly $10,000 in monthly ad spend, agency fees and minimums often exceed the improvement a competent generalist could get by spending four focused hours a week in the account. Amazon PPC rewards consistent small adjustments more than brilliance, and consistency is learnable.
Keep it in-house when your catalogue is narrow, your margins are well understood, and someone internal genuinely has those hours. Hire out when the catalogue is broad enough that harvesting and negation becomes a real operational job, when you're running across multiple marketplaces, or when the internal person you'd assign is your only merchandiser and their time is worth more elsewhere.
The middle path that works well: hire an agency for a three-month build and handover, then run it internally with a quarterly audit. You pay for the architecture rather than for someone to keep pressing buttons.
And before you increase spend at all, check that the conversion side is holding up. Paid traffic into a page that doesn't convert is a margin problem dressed up as a media problem — the same lesson that applies to ecommerce conversion rate optimisation and to paid search for ecommerce generally.
FAQ
What do Amazon PPC management services typically cost? Most agencies charge either a percentage of ad spend in the low-to-mid teens with a monthly minimum, or a flat retainer that commonly starts around $1,500 to $3,000 for a single marketplace and a moderate catalogue. Profit-share arrangements exist but require an agreed attribution basis to work.
What is a good ACOS on Amazon? There is no universal answer, and any agency that gives you one without seeing your numbers is guessing. Good means below your break-even ACOS, which equals your contribution margin after referral, fulfilment, storage and returns costs. For many sellers that lands somewhere in the 20% to 35% range, but the range matters far less than your own figure.
How long before an agency should show results? Expect structural work and search term harvesting in the first 30 days, measurable efficiency change by 60 days, and a fair judgement at 90. Amazon's data needs volume to be meaningful, so anyone promising a transformation in three weeks is either restricting spend or got lucky with seasonality.
Do I lose control of my Amazon advertising account? No. The advertising account sits inside your Seller Central, and agencies work through user permissions you grant and can revoke. Make sure access is granted that way rather than by sharing login credentials, and agree upfront that campaign structures and negative keyword lists remain yours on exit.
Should I run Sponsored Display or Amazon DSP? Sponsored Display is worth testing once Sponsored Products is stable and profitable, primarily for retargeting and for defending your own listing pages. DSP is a separate product with higher minimums and a different skill set; it makes sense for established brands with meaningful budget, not as a first expansion.