Amazon DSP vs Sponsored Products: Why the Best Advertisers Use Both

At some point, every Amazon advertiser runs into the same problem. More budget goes in. Sales don't move. The campaigns look fine. The issue isn't the campaigns.

It's that Search has a ceiling — and most advertisers don't realize they've hit it until they've been stuck there for months.

‍

What Is the Search Ceiling on Amazon?

SP and SB do one thing well: they reach people who are already looking. Someone searches, your ad shows up, they buy. That model works — reliably, at scale — right up until it doesn't.

The pool of people actively searching for your product today is finite. Once you've captured most of it — high impression share, competitive bids, solid conversion rate — spending more just costs more. You're not growing the pool. You're paying to stay in it.

That's the ceiling. You can't bid your way out of it. To grow past it, you need to reach people before they search.

‍

How Amazon DSP Reaches Audiences Beyond Search?

DSP runs display and video ads on Amazon and across the web. The targeting isn't keywords — it's people. Specifically, people who haven't found you yet, people who visited your listing and left, and people who bought once and didn't come back.

Three use cases drive most of the value:

‍

Retargeting is usually where advertisers start. Someone landed on your product page and didn't buy. You already paid to get them there — through Search, through organic ranking, through whatever brought them in. DSP brings them back. The intent is already there. You're not rebuilding it from scratch.

Worth clarifying one thing here: Sponsored Display also has a retargeting function. The difference is reach. SD stays on Amazon. DSP follows users off-platform — other sites, other apps — after they've left. The audience controls are also more granular: lookback windows, purchase behavior, category interest. They're not the same tool doing the same job.

Prospecting goes further upstream. DSP can target audiences who look like your existing buyers — same demographics, same purchase patterns, same category behavior — but who've never heard of you. This is where new demand actually gets created. It converts slower than retargeting. It's harder to attribute. But it's the only lever that genuinely expands your market.

Retention targets past buyers for repurchase or cross-sell. The trust is already there. The barrier is low.

‍

Why Last-Click Attribution Undervalues Amazon DSP

DSP gets a bad reputation for being hard to justify. That's usually a measurement problem, not a performance problem.

Last-click attribution gives DSP almost no credit. Here's what actually happens: a customer sees a display ad on Tuesday, searches on Thursday, buys through a Sponsored Products ad on Friday. The Search ad gets the conversion. DSP looks like it did nothing.

The purchase path was real. The attribution just wasn't looking at it.

There's a pattern we see consistently: advertisers launch DSP prospecting, then their Search conversion rate goes up. Looks like a Search win. It usually isn't. DSP pulled new people into the funnel, they searched by brand name, and the Search ad got the credit. Without cross-channel visibility, that signal is invisible — and DSP keeps looking expensive on paper.

‍

The Right Metric for DSP Isn't ROAS

Most advertisers evaluate DSP by its attributed ROAS. That's the wrong test.

Attributed ROAS only counts revenue the attribution model can see — and last-click attribution, as we covered, misses most of what DSP actually does. More fundamentally, a strong DSP ROAS doesn't tell you whether DSP grew your business or just claimed credit for sales you'd have made anyway. You could have impressive ROAS numbers and zero incremental growth. The metric doesn't surface that distinction.

The honest test is at the total business level. That's where TACoS comes in.

TACoS, Total Advertising Cost of Sale, measures your combined ad spend across all channels (Sponsored Ads and DSP) against your total revenue, including organic. Not just ad-attributed revenue. Everything. When DSP pulls new audiences into your funnel and they eventually convert through Search or organically, TACoS captures that lift. When DSP spend goes up without real revenue growth behind it, TACoS captures that too. It doesn't let you hide behind channel-level numbers.

The test is simple: did total revenue go up after you turned DSP on? Did TACoS stay under control? If sales rise while TACoS holds roughly flat, that's incremental volume — not reshuffled credit.

‍

AMC Attribution: Seeing the Real Purchase Path

Amazon Marketing Cloud is what fixes the attribution problem. It's a clean room environment — free to use, anonymized at the user level — that brings your Sponsored Ads data, DSP data, and Amazon retail data into one place. Not individual journeys. Statistical reality: how your touchpoints actually interact before a purchase happens.

In practice: you can see what percentage of DSP-exposed customers eventually converted through Search, how long the path was, which audiences moved fastest, and where there's room to shift budget.

One honest caveat: AMC requires SQL to get meaningful data out of it. The access is free. The analysis isn't plug-and-play. That's a real barrier for most advertisers — which is why m19 Pro surfaces AMC insights directly without requiring you to write queries yourself.

‍

Amazon DSP Performance: What the Data Shows

We ran exactly that test across 168 advertisers live on m19's DSP automation, comparing the 60 days before launch to the 60 days after.

The median advertiser: +8% total revenue, TACoS up just 1pp. Across the full cohort: +30% in sales, TACoS up 2pp.

Sales up, cost of sale close to flat. That's the signature of incremental growth.

TACoS is the right lens because it doesn't let you hide behind channel-level attribution. When DSP drives brand search and organic lift, TACoS picks it up. When DSP spend grows without real revenue behind it, TACoS picks that up too. It's a harder metric to game — which is exactly why it's the right one.

DSP performance still depends on category, competition, creative, and audience strategy. But the directional signal across 168 advertisers is consistent: advertisers who added DSP to a mature Search account grew. Those who stayed in Search alone didn't.

‍

Conclusion

These aren't three separate strategies running in parallel. They're one loop.

Search captures existing demand. DSP creates new demand and recovers the intent Search missed. AMC measures what actually drove results — not what the attribution model says drove results — and feeds that back into both channels.

The ceiling Search alone creates is real. Most advertisers hit it without knowing what it is. Once you see it for what it is — a structural feature of the channel, not a campaign problem — the path forward becomes clearer.