What an Amazon Marketing Firm Actually Does — and How to Check It's Working
An Amazon marketing firm manages your Sponsored Ads, DSP and AMC reporting instead of your in-house team. The question that actually separates them: do they hold real DSP seats and measure incrementality, or just resell access and report last-click numbers?
What this looks like in a real account
What an Amazon marketing firm actually does
An Amazon marketing firm is a company you pay to run some or all of your Amazon advertising and account management instead of doing it in-house. That can mean Sponsored Products, Sponsored Brands and Sponsored Display (the auction-based search ads), Amazon DSP (programmatic display and video bought on and off Amazon), Amazon Marketing Cloud reporting, listing and A+ content work, and catalog operations like FBA/FBM management.
Most firms specialize even when their homepage lists everything. Some are built around creative and listing optimization with ads bolted on. Others are built around media buying — Sponsored Ads and DSP — with creative and catalog work handled by the brand or a separate team. Neither model is wrong. The mismatch happens when a brand needs one and hires the other.
- Full-service firms bundle strategy, creative, storefront, ads and inventory ops into one account team.
- Media-buying specialists run Sponsored Ads and DSP as a managed service, usually plugging into a brand's existing creative and catalog.
- In-house teams keep everything internal and hire a firm only for DSP seat access or AMC analysis they can't build themselves.
The one distinction that predicts almost everything else: does the firm hold its own DSP seats, or does it resell access through someone else's? A seat holder can build audiences, run holdout tests and pull log-level data in AMC. A reseller is limited to whatever dashboard the seat owner gives them.
A worked example: reading a real DSP performance book
Take a firm that reports "6.04x ROAS" on a DSP program. That number alone tells you nothing — it could be one hero campaign out of twenty. It becomes evidence once you can see it against the whole book it came from.
Here's what that looks like at scale: across 30 advertisers reMKTR ran in July 2026, the portfolio — not the best line item — delivered a 6.04x return on ad spend, on 78.4 million impressions bought at a $4.00 CPM, with a blended $1.42 cost-per-click. Blended cost per acquisition landed at $5.49 across 57,137 attributed purchases, and 20.1% of those purchases came from shoppers new to the brand.
Notice the blended CPC — $1.42 — sits well above the $0.41 figure you'll see quoted around this category. That $0.41 number is real, but it's an online-video-only rate, not what the whole book paid. This is the exact substitution to check for in any case study a firm shows you: is the headline number the whole account, or the best channel inside it?
When you ask a firm for their numbers, ask for these four in the same sentence: ROAS, CPM, blended CPC, and new-to-brand percentage, all pulled from the same period and the same set of advertisers. If they can only give you one, that's the one they're proud of — not necessarily the one that matters.
How Amazon marketing firms charge — and what each structure hides
Ask three firms how they charge and you'll often get three different structures, sometimes with no number named until a call. That's not automatically a red flag — Amazon media buying is variable enough that flat pricing rarely survives contact with a real account — but you should know which structure you're looking at before the call, and check the firm's current pricing directly rather than trusting a third-party listicle.
None of these structures is inherently better. Percentage-of-spend pricing, done honestly, aligns the firm's incentive with growing your spend responsibly. The problem isn't the model — it's when a firm publishes a low base fee and never states the percentage, so the real cost only shows up on the first invoice at scale.
Last-click reporting is the industry's biggest blind spot
Most Amazon reporting — inside Seller Central, inside most DSP dashboards — defaults to last-click attribution. It gives the sale credit to whichever ad the shopper clicked right before buying, even if that click just intercepted a purchase they were already going to make.
Last-click can't tell you whether DSP added a sale or took credit for one Sponsored Products would have closed anyway. It never could — it isn't built to answer that question. Amazon Marketing Cloud, reconciled properly, can show DSP and Sponsored Ads without double-counting each other, which is the first honest step. The second step is a holdout or matched-control test: run DSP to part of the audience and withhold it from a comparable part, then compare outcomes. That's the only way to prove incrementality rather than assume it.
If a firm's entire reporting deck is ROAS and ACoS with no mention of holdouts, matched controls, or AMC reconciliation, ask directly how they know DSP added anything. A confident, specific answer is a good sign. A pivot back to ROAS is your answer.
The mistakes that waste the budget — ours included
The most common mistake is scaling DSP budget before the brand's Sponsored Ads foundation can convert the traffic it sends — display drives awareness and consideration, but if the listing, reviews or price aren't ready, that spend shows up as impressions with nothing behind them. We've made this mistake ourselves: pushing budget into a program before checking that the catalog underneath it could hold the traffic.
The second is judging DSP on ROAS alone without checking new-to-brand share. A campaign can post a strong ROAS by retargeting shoppers who already knew the brand and were going to buy anyway — that's not growth, it's a receipt. New-to-brand percentage is the number that tells you whether the spend actually expanded the customer base.
The third is treating a firm's case study as proof rather than a floor. A single campaign's numbers, especially a screenshotted one, tell you what's possible under ideal conditions, not what your account will do. Ask for the portfolio number, not the best line item — that's the one that will predict your result.
Where reMKTR fits
reMKTR runs Amazon DSP as a managed service for brands that already have their sponsored ads and catalog fundamentals working and want display media measured on incrementality, not last-click — we hold our own DSP seats, reconcile through AMC, and are part of the Full Circle group, which has managed more than $500M in Amazon spend across 100+ brands including HexClad, Ridge, BK Beauty, Beardbrand, Epic Gardening, The Woobles and Walkize. If you need creative, listings and inventory ops bundled with ads under one team, a full-service firm like Darkroom is built for that job and we aren't trying to be. If you just need to know whether your current DSP spend is doing anything, that's the question we're built to answer.
| Structure | How it works | What to watch for |
|---|---|---|
| Flat retainer / per-seat | Fixed monthly fee regardless of spend | Fine at low spend, but watch for scope creep instead of underpricing |
| Percentage of ad spend | Fee scales with what you spend on ads | The rate matters more than the base fee — a hidden point of spend is real money once budgets grow |
| Banded by spend tier | Fee steps up at spend thresholds | Ask what happens right at a boundary — some firms reprice the whole account, not just the increment |
| Quote-only / custom | No published rate, priced after a discovery call | Not a red flag by itself — get the full structure, including any percentage, in writing before signing |
Which one you should actually pick
Full-service firms suit brands that want creative, listings, storefront and inventory managed alongside ads under one team — Darkroom is built for that. Media-buying specialists suit brands with a working catalog and creative that need DSP run and measured properly. In-house teams suit brands with staff to build AMC and creative internally but no DSP seat access.
Shortlist on the job, not the feature grid. Pull your search-term report for the last 90 days and total the spend against terms that produced no orders — 33.6% on the account above. Then ask each vendor on your list what they would do about it in week one, and see who answers with a process rather than a screenshot.
Common questions
Is an Amazon marketing firm the same as an Amazon agency?
Yes — the terms are used interchangeably. Neither implies a specific scope, so always ask whether the firm runs Sponsored Ads, DSP, both, or neither, and whether they hold their own DSP seats or resell access.
How much does an Amazon marketing firm cost?
It depends on the structure, not just the rate. Most charge a flat retainer, a percentage of ad spend, a banded fee that steps up with spend, or a custom quote after a discovery call. Check their current pricing directly and get the full structure — including the percentage, if that's the model — in writing before you sign.
What's a red flag when evaluating one?
No mention of DSP seat ownership, no attribution methodology beyond ROAS/ACoS, and case studies that show one campaign's numbers instead of the whole account's. Ask for the portfolio-level figure, not the best line item, and ask how they test for incrementality rather than infer it from last-click.
Can a firm run both Sponsored Ads and DSP well, or should I split vendors?
Some do both well, especially full-service firms built for it. The practical test isn't the org chart — it's whether they reconcile DSP and Sponsored Ads in AMC so the two aren't double-counting the same sale. If they can show you that reconciliation, splitting vendors usually isn't necessary.
How do I know if DSP is actually working, not just reporting well?
ROAS alone won't tell you. Ask for the new-to-brand percentage and whether they've run a holdout or matched-control test. A campaign that only retargets existing buyers can post a great ROAS while adding zero incremental sales.
We show the method before the number.
Claim the free auditRead next
- Amazon DSP Partner Comparison: Four Routes InComparison · amazon dsp partner comparison
- Criteo Pricing: The Fee Stack Inside Your BudgetPricing · criteo pricing
- Pacvue Pricing: No Public Number — What to AskPricing · pacvue pricing
- Skai vs Pacvue: Contracts, Not Feature GridsHead to head · skai vs pacvue