Almost everything written about Amazon DSP assumes a brand with a team, a media budget, and an agency on retainer. That is who the platform was built for. The question for everyone else, the seller doing a few hundred thousand a year with one person on ads, is whether any of it applies.
Some of it does. This guide is the version for sellers below the big-brand line: how access works at smaller scale, what to run if you get in, and the honest case for not getting in yet.
Why the standard advice does not fit
DSP is a programmatic platform that buys display and video placements on and off Amazon, targeting audiences built from Amazon's shopping data. Amazon's own managed service, where their team runs the campaigns, has carried a monthly minimum that is more than many sellers' entire ad budget. The advice written for that tier, multi-audience funnels, video creative, off-Amazon awareness campaigns, assumes a budget that can afford to learn slowly.
A smaller seller cannot. The question is not "how do I run DSP well" but "can DSP do anything for me at a budget where mistakes are expensive," and the answer is narrower than the platform's marketing suggests.
How access works below the minimum
Agency seats. Many Amazon advertising agencies hold DSP access and run campaigns for clients at lower minimums than Amazon's direct service, charging a management fee or a percentage of spend. This is the most common route for mid-size brands. It works if the agency's DSP practice is real and not a reseller arrangement; ask how many DSP accounts they run and who runs them.
Self-service platforms. A growing number of third-party tools offer DSP access through their own interface, with lower minimums and a software fee. The trade-off is that you are now learning both DSP and the platform, and the support for a small account varies.
In both cases, the minimum still exists. It is lower, and it is real. This breakdown of DSP costs covers the structures.
What to run if you get in
One thing. Retargeting your own product viewers, purchasers excluded, with a lookback window of 14 to 30 days. Optimize for conversions. Run it continuously for at least a month.
This is the campaign with the smallest audience, the highest intent, and the clearest baseline. It tells you whether DSP creative and your product pages work together for your catalog, before you spend on audiences that have never heard of you. The targeting options guide lays out what comes after, and the honest answer for a smaller seller is that "after" may not come for a while.
Do not start with in-market or lifestyle audiences. They are larger, cheaper per impression, and much harder to read. A small budget spread across broad audiences produces results that look like noise, because they are.
The case for waiting
Three questions, and if any answer is no, the budget is better spent elsewhere.
Is Sponsored Products optimized? ACoS where you want it, structure clean, search terms current, scheduling in place. If not, the cheapest growth available is finishing that work. DSP on top of a leaky Sponsored Products program adds a second, more expensive leak.
Is there a retargeting pool worth reaching? DSP retargeting needs product views to build an audience from. A catalog with a few hundred views a month does not have enough people to retarget for the spend to deliver. Check the traffic numbers in your business reports first.
Is reach the constraint? If you are winning the searches that exist and the ceiling is shoppers who have not searched, DSP is the next tool. If you are not yet winning the searches that exist, more reach brings more shoppers to a page that is losing the ones it already gets.
For most small and mid-size sellers, at least one of these is a no, and the right move is to come back in two quarters.
The in-between: Sponsored Display
If the three questions are a yes and the minimums are still a stretch, Sponsored Display does most of what a small seller would use DSP for. Views-based retargeting, in-market audiences, contextual targeting on competitor pages, and some off-Amazon placements, all in the regular console, billed per click, with no minimum.
What it lacks relative to DSP: fewer placements, less audience control, limited frequency management, and no video. Those are scale problems. A seller who has saturated Sponsored Display's retargeting and wants more is a seller who has outgrown the small-seller question, and DSP is the next step for them. For everyone else, Sponsored Display is enough.
If you do go in
Treat DSP spend with the same discipline as sponsored ads. Budgets paced, event windows ramped and restored, performance read weekly against the goal the campaign was set to. The monitoring checklist applies, and the one new habit is reading DSP on the metrics it is built for (detail page views, new-to-brand, lift in branded search) rather than on ACoS alone, which understates it. The DSP versus sponsored ads comparison covers how to read the two together.
Frequently asked questions
Can a small seller get access to Amazon DSP?
Yes, through an agency with a DSP seat or a third-party platform that offers self-service access, both at lower minimums than Amazon's managed service. Each adds its own fee on top of media spend, and the minimums still exist; they are lower, not gone.
What is the smallest useful DSP budget?
Enough to run one retargeting campaign long enough to read, which means several weeks of continuous delivery to a views-based audience. Below that, the data is too thin to tell whether the channel works for the catalog, and the spend is better put into Sponsored Products.
Is Sponsored Display a substitute for DSP for small sellers?
For most, yes. It covers views-based retargeting, in-market audiences and some off-Amazon placements in the regular console with no minimum. The brands that outgrow it need more placements, frequency control or video, and that is usually a scale question rather than a small-seller question.
Off Hours is for the first question above: getting Sponsored Products optimized and running on rules before the next channel is the question. Start a free 14-day trial.