Amazon DSP Partner Comparison: Compare Routes First

"Partner" is a defined term here, not a synonym for supplier

In this market the word carries a specific meaning, and knowing it stops you comparing apples to filing cabinets.

Amazon runs a Partner Network, described on its own site as a self-service hub for agencies and tool providers to manage their business relationship with Amazon Ads. Three kinds of organisation are eligible: agencies that support advertisers in managing campaigns, tool providers offering software, and companies providing advertising training and information. So a listed "partner" may be a media buyer, a piece of software, or a training business — three very different purchases sharing one label.

On top of that sits partner status. Amazon's own description is that advanced or verified status is earned by demonstrating expertise with Amazon Ads and delivering growth for advertisers over time, and that status holders can display the badge in their own marketing and receive an enhanced directory listing. Amazon's guidance to advertisers is measured and worth quoting in spirit: treat partner status as one input into your selection process, alongside your own business objectives and each firm's track record.

Before going further — Amazon DSP here means the demand-side platform, the programmatic product for display, video and audio. Amazon Delivery Service Partner is the unrelated courier franchise with routes and vans. Both get abbreviated to DSP and the search results mix them.

The four routes, and what each one actually costs you

Most "partner comparison" articles compare companies. The larger decision is which route you are taking, because it sets your floor, your control and your staffing need before any company is chosen.

• Amazon's managed service. Amazon's own team plans and runs the campaigns. Their product page describes this option as designed for advertisers who want access to Amazon DSP inventory with advisory service, and as typically requiring a minimum investment of USD 50,000, noting that the figure varies by country. Costs you the least internal effort and the most budget certainty at the bottom end.
• A self-service seat. You hold the account and keep full control of campaigns. Costs you a competent trader on payroll, permanently. Cheapest per dollar of media if — and only if — that person already exists and is good.
• An agency holding seats. The most common route for mid-market brands. Costs you a fee, usually a percentage of media, and a dependency you should structure your exit around from day one.
• A packaged product. The same work bought with a defined scope rather than a retainer relationship. Costs you flexibility at the edges and buys you predictability in the middle.

Pick the route on your own constraints — internal capability, budget scale, how much control you want — and only then shortlist companies inside it. Doing it the other way round is how brands end up comparing a $50,000-floor managed service against a boutique retainer and concluding that the market is confusing.

What a partner badge does and does not certify

Badges are genuinely useful information. They are also routinely over-read, so here is the honest boundary.

What it does tell you. That the firm has a real, ongoing relationship with Amazon Ads; that it has demonstrated expertise and delivered advertiser growth over a period, on Amazon's own assessment; that it maintains certifications and appears in a directory Amazon controls rather than one it bought a listing on. That is a meaningful filter, and it is checkable by you in a browser rather than taken on trust from a pitch deck.

What it does not tell you. Whether the firm runs DSP specifically, as opposed to sponsored ads — plenty of badged partners are excellent at search and light on programmatic. Whether it has ever tested incrementality. Whether the team that earned the status is the team that will run your account. Whether its fees are reasonable, capped, or inclusive of creative. Whether your category is one it has ever worked in.

What to do with it. Use status as a gate, not a ranking. Anyone badged clears the first bar; the comparison then happens on the seven things that actually differ. And check current criteria on Amazon's own pages rather than in a third-party article, because programme details in this area change and most write-ups are describing an older version of the rules.

The three things that genuinely differ between partners

Once route and credential are settled, partners differ on a surprisingly short list. Everything else is presentation.

Audience architecture. Anyone can select Amazon's standard segments. The difference is in how remarketing pools are defined and refreshed, how in-market and lifestyle audiences are layered, how conquesting is bounded so it does not simply buy your own existing customers back, and whether your first-party data is genuinely being used or merely mentioned. Ask to see a segment map from a live account, redacted.

Inventory discipline. Which placements run on Amazon properties and which run across the wider web and apps, what the default exclusion list contains, and who has authority to change it mid-flight. A partner who cannot show you an exclusion list has not thought about where your brand appears.

Measurement design. The one that determines whether the other two can be assessed at all. Ask what gets withheld, from whom, for how long, and how the control group is matched. Then ask how display and sponsored ads are stopped from being credited with the same purchase — because if both channels claim it, your blended return is inflated and the inflation grows with your budget.

A comparison sheet you can fill in on a call

Eight questions, all answerable in a first conversation, all producing a comparable answer rather than a narrative.

• Which route am I buying, and whose seat is it?
• What is your minimum monthly media spend, and is there a minimum fee underneath it?
• What is the fee basis and the exact rate, and what is it calculated on?
• What is inside the fee — creative production, clean-room analysis, audience development?
• Who runs my account weekly, and how many other accounts do they hold?
• Describe a holdout you have run and what it showed.
• What do I keep on exit — audiences, creative, clean-room instance, historical data?
• What notice period applies, and what happens to my rate if the firm is acquired?

Score each candidate on the same eight. Any answer that arrives as reassurance rather than as a number or a document is a gap, and gaps are what you are hunting.

Where reMKTR sits on this map

We are route three: an agency holding its own seats, focused on Amazon DSP and nothing else. reMKTR is the DSP arm of Full Circle, a full-service Amazon management company with $500M+ in managed spend across 100+ brands. We run 109 live Amazon DSP advertiser seats.

Across 30 of those advertisers in July 2026 — the scope stated because a figure without one is not evidence — the book returned 6.04x on ad spend, ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost per click, and delivered $5.49 cost per acquisition across 57,137 attributed purchases, 20.1% of them from shoppers new to the brand.

Our position on the third differentiator above is the reason to hire us rather than the numbers. Last-click attribution cannot prove incrementality and never could; holdouts and matched controls can. We reconcile in Amazon Marketing Cloud, where display and sponsored ads stop double-counting each other. Where we are not the right route: if you have a strong trader in-house, buy the seat; if you clear the managed-service minimum comfortably and want Amazon's own team, that is a legitimate choice we would not argue you out of.

Two siblings and when they apply. Dr. DSP — route four, the same capability packaged as a product with a defined scope, for teams who would rather buy an outcome than manage an agency. Dr. Stock — when stockouts and fees are the real constraint, because driving display demand into inventory that runs out mid-flight wastes the media and damages rank.