Running Amazon DSP When Most of Your Revenue Is DTC
A brand where most revenue comes from its own site can still use Amazon DSP productively — either pointed at Amazon to grow that smaller channel, or link-out to the DTC site directly, using Amazon's shopping-signal targeting the way any other programmatic buy uses audience data. What changes is which metric matters: total-account revenue, not Amazon-specific ROAS, since the goal is the whole business.
What this looks like in a real account
Why DSP is worth considering even when Amazon is the smaller channel
Amazon DSP's targeting is built on shopping and purchase-intent data at a scale most programmatic platforms can't match — that value doesn't disappear because a brand's revenue happens to be weighted toward its own site. A DTC-majority brand can use Amazon DSP the same way any advertiser uses a demand-side platform: to reach an audience, based on real purchase signal, wherever that audience is most efficiently converted — which, for a DTC-majority brand, is often the brand's own checkout rather than Amazon's.
What actually changes about the strategy
Two things shift meaningfully. First, the destination decision matters more — a DTC-majority brand has a real incentive to run more spend as link-out rather than Amazon-destination, since Amazon's own listing carries a referral fee and a customer relationship the brand doesn't own the way it does through its own site. Second, and more subtly, the measurement conversation has to include whether Amazon DSP spend is genuinely additive to total revenue or partly redirecting demand that would have converted on the DTC site anyway — the cannibalisation question, which is real enough to deserve its own dedicated treatment on this site and isn't something to wave away with an assumption either way.
A worked example of the whole-account view
Take a brand doing $400,000 a month in total revenue, 85% DTC ($340,000) and 15% Amazon ($60,000). They run $18,000 a month in Amazon DSP, split $6,000 Amazon-destination and $12,000 link-out to their DTC site. The Amazon-destination portion reports a clean 4.5x attributed ROAS ($27,000). The link-out portion, measured through the DTC site's own analytics, shows a 2.1x ROAS ($25,200). Summed naively, that's $52,200 against $18,000 spent — a 2.9x blended figure. But the real question for a DTC-majority brand isn't the blended ROAS on the DSP line; it's whether total-account revenue, DTC plus Amazon combined, grew by more than $52,200 worth relative to what would have happened without the spend — which requires an incrementality check the attributed numbers alone can't provide.
The one-account-measured field pattern worth knowing
On one publishing brand we've worked with, organic order share on Amazon rose from 13.5% before DSP spend to 27.9% after it started, and was still climbing four weeks after every dollar of DSP spend had stopped — while branded sponsored-ads efficiency on the same account compressed from 48% to 37% TACOS and held. That's evidence, from one account, that DSP spend can lift performance across an account beyond its own attributed line — a genuinely useful data point for a DTC-majority brand weighing whether DSP's value shows up entirely inside its own attributed numbers or partly bleeds into everything else the brand runs. It's one account, one window — worth testing on your own numbers, not assuming as a rule.
The common mistake, including ours
The mistake is judging Amazon DSP spend purely against Amazon-side metrics when the brand's actual business goal is total-account growth. We've had a DTC-majority client nearly cut a DSP line because its Amazon-attributed ROAS looked mediocre next to their Meta and Google numbers, without checking whether the same spend was contributing to DTC site traffic and conversion through the link-out portion, or to overall brand lift that showed up in neither channel's own attributed report. The fix wasn't a better Amazon report — it was reframing the measurement question around the whole business rather than one channel's own dashboard.
Why DSP audience data is still valuable even if you never send a click to Amazon
It's worth being explicit about the strongest version of the DTC-majority case: a brand could, in principle, run 100% of its Amazon DSP spend as link-out, never sending a single click to the Amazon listing, and still get real value from the platform — because the targeting signal (shopping behaviour, browse history, purchase-intent data) is what's valuable, not the destination. That's a legitimate strategy, not a compromise, for a brand whose margin math favours keeping every conversion on its own site. The trade-off is giving up the Amazon-side benefits — sales rank, the natural pace of new reviews that comes from real order volume, the native one-click purchase experience — in exchange for owning the full customer relationship and margin.
What to check before scaling or cutting
Before scaling DSP spend for a DTC-majority brand, run the cannibalisation check specifically — a geo-lift test comparing total-account revenue (both channels combined) in test versus control markets is the honest way to see whether the spend is genuinely additive. Before cutting on the strength of a weak Amazon-attributed ROAS alone, check the link-out portion's own conversion data and any available cross-channel signal — a DSP line judged only on its Amazon-side numbers, for a brand where Amazon is the smaller channel, is very likely being judged on the wrong metric.
| Revenue mix | DSP destination emphasis | Metric that matters most |
|---|---|---|
| Amazon-majority | Amazon-destination, standard attribution | Amazon-attributed ROAS, new-to-brand |
| DTC-majority | More link-out weighting | Total-account revenue, incremental lift |
| Roughly balanced | Mixed, tested regularly | Both, reconciled — plus the cannibalisation check |
Which one you should actually pick
A DTC-majority brand with basic cross-channel analytics can build this whole-account measurement view themselves — the framework here is the whole method, and nothing about it requires a managed-service relationship. reMKTR runs the same destination-split and cannibalisation-testing discipline for DTC-heavy clients as for Amazon-majority ones, treating total-account revenue as the real scoreboard rather than a single channel's own attributed number, as part of the practice behind Full Circle's $500M+ in managed Amazon spend across 100+ brands.
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 Amazon DSP worth running if Amazon is only a small part of my revenue?
Often yes — DSP's real asset is Amazon's shopping-signal targeting, which works regardless of destination. A DTC-majority brand can weight spend toward link-out campaigns pointed at its own site rather than assuming DSP only makes sense for Amazon-heavy brands.
How do I measure Amazon DSP's impact on my whole business, not just Amazon?
Track Amazon-destination and link-out spend separately, and consider a geo-lift test measuring total-account revenue — both channels combined — as the outcome metric, which is the most direct way to see whole-business impact.
Should I worry that Amazon DSP is just stealing sales I'd get on my own site anyway?
It's a legitimate concern worth testing directly rather than assuming either way — see the dedicated cannibalisation page on this site for the specific test design.
What's the biggest measurement mistake DTC-majority brands make with DSP?
Judging the spend purely on Amazon-attributed ROAS, when the actual business goal is total-account growth across both channels — a metric mismatch that can make genuinely useful spend look like it's underperforming.
We show the method before the number.
Claim the free auditRead next
- Tinuiti Pricing: How the Quote Gets BuiltPricing · tinuiti pricing
- Acorn vs Podean: Two Amazon Agencies ComparedHead to head · acorn vs podean
- Pacvue Pricing: No Public Number — What to AskPricing · pacvue pricing
- Skai vs Pacvue: Contracts, Not Feature GridsHead to head · skai vs pacvue