Amazon DSP Advertising in 2026: The Enterprise Guide to Amazon DSP Targeting

Amazon DSP advertising changed in 2026, and the biggest change wasn’t to the platform. It was how Amazon counted what your ads did.

If you’re comparing this year’s Amazon DSP performance against last year, attributed revenue is probably lower. 

That doesn’t mean your campaigns got worse. On January 1, 2026, Amazon introduced a shopping signal enhanced last touch model for eligible Amazon Store ads, changing how purchases get credited. 

Numbers reported today aren’t directly comparable with 2025. Rebuild your baseline using Amazon’s Purchases (All Views) metric before you draw any conclusions.

At Sequence Commerce we manage Amazon DSP advertising and Streaming TV campaigns for enterprise and Fortune 500 brands. 

Across the accounts we manage, measurement changes have moved reported performance more than strategy changes have.

This guide covers what changed, how it affects your reporting, and what to do next. It also looks at what happens to DSP as Amazon’s AI reshapes how shoppers find products.

What Is Amazon DSP Advertising?

Amazon DSP is Amazon’s demand-side platform that lets advertisers programmatically buy display, video, audio and streaming TV ads across Amazon-owned and third-party inventory. Pricing is primarily CPM-based, although some campaign types may support other optimization models. 

One point of confusion is worth clearing first. The letters DSP mean two different things at Amazon.

A demand side platform is digital advertising technology. An Amazon Delivery Service Partner is a logistics business that runs delivery routes. Search results mix the two constantly. The DSP amazon meaning on this page is always the advertising platform.

In this blog read about what a DSP platform does. 

It centralises media buying across many inventory sources and bids in real time on your behalf. That’s the DSPs meaning across the whole industry, not just at Amazon.

What Makes Amazon DSP Different From Other Programmatic Advertising Platforms

Many programmatic advertising platforms use combinations of demographics, contextual signals, cookies, device identifiers, and other audience data. Amazon DSP differentiates itself through Amazon’s first-party shopping and streaming signals. 

Amazon Ads reports that its authenticated audience reaches approximately 90% of U.S. households. Amazon’s first-party shopping signals are a key differentiator of Amazon DSP compared with many other demand-side platforms. 

The practical result is that DSP targets audiences by what they actually bought, not what a model assumes about them. Amazon DSP helps advertisers reach shoppers who have demonstrated category behaviour, and it delivers ads to the right moment in their journey rather than the right keyword.

If you want the beginner level breakdown, our guides on what Amazon DSP is and the Amazon demand side platform cover the fundamentals. This article assumes you already run Amazon Sponsored Products well and are deciding how to scale programmatically. Our Amazon DSP management team works with brands at exactly that point.

What’s New in Amazon DSP in 2026

Three things shifted this year. All three change how Amazon DSP campaigns should be run.

Amazon’s advertising business reached $19.8 billion in Q2 2026, up 26% year over year (Amazon Q2 2026 earnings, July 30, 2026). 

Attribution Got Stricter on January 1, 2026

• Amazon retired the blind 14 day view through the window. In its place came a shopping signal enhanced last touch model driven by machine learning.
• The old system asked one question. Did the shopper see the ad within 14 days of buying?
• The new system asks for a harder one. Did that ad view actually influence the purchase?
• The model looks at exploratory browsing, general category searches, past shopping patterns, and whether the shopper was discovering the brand or already intending to buy. 
• If it decides the shopper was going to purchase anyway, the view gets no credit.
• This applies to Amazon Store ads, which includes Amazon DSP ads, Sponsored Brands and DSP display placements, and Sponsored Display. 
• Offsite delivery is the exception and still runs the old 14 day click and 14 day view window. Click attribution didn’t change at all.

Automation Turned On by Default in July 2026

• On July 20, 2026, Amazon shipped 10 new capabilities across Brand+ and Performance+. Three of them activate without any advertiser action.
• Brand Outcome Optimization is now live on all Brand+ Reach line items. It optimises toward branded search, detail page views and add to cart instead of reach alone. 
• Prime Video Signal Integration feeds streaming behaviour into Brand+ ranking, which makes ads on Prime Video more relevant to what the viewer actually shops for. 
• Automatic Deal Selection is on by default for new Brand+ Streaming TV prospecting lines.
• Audience Signals lets your first party DSP data enter Brand+ and Performance+ as an optimisation input rather than a hard targeting fence. 
• The AI uses it to learn faster while staying free to find shoppers outside your uploaded segments. It runs through Ads Data Manager and costs nothing extra.
• Audio also moved to general availability for Brand+ and entered open beta for Performance+. 
• Amazon reports audio drives 43% incremental reach compared with Streaming TV campaigns, and improves ROAS efficiency by 6.7% to 8.7% when paired with video (Amazon internal data, July 2026). 

Supply Consolidated Around Amazon DSP

• Amazon spent the last year absorbing inventory that used to sit somewhere else.
• Netflix and Spotify are now direct integrations available through Amazon DSP. 
• Spotify brings a global audience of 777 million monthly active users (Spotify Q2 2026 earnings, reported August 2026).
• SiriusXM Media and iHeart added streaming audio, and the Podcast Audience Network is now integrated. 
• Amazon DSP also became the preferred transition partner for Microsoft Invest customers as that platform winds down.
• Complete TV arrived to manage an entire streaming investment in one place, across publishers and including linear.
• Amazon DSP in 2026 is becoming the primary omnichannel buy, and Amazon is pricing and positioning it that way.

Why Your Amazon DSP Performance Looks Worse Than It Is

Since Amazon replaced the 14 day view through window with a machine learning model on January 1, 2026, fewer ad views qualify as attributed conversions. 

Reported revenue falls even when real sales hold steady. 

Agencies across the market are seeing double digit declines in attributed Amazon DSP performance this year. 

Shopper behaviour didn’t change on January 1. The measurement did.

Here’s how it plays out inside a real brand. A household goods brand runs the same Streaming TV and display plan in Q1 2026 that it ran in Q4 2025. 

Sell through is flat. Detail page traffic is flat. Attributed DSP revenue drops 20%. Finance sees a 20% efficiency loss and asks the media team to defend next quarter’s budget.

All of it is a reporting problem.

Amazon saw this coming and shipped a fix. It’s called Purchases (All Views) and almost nobody uses it.

That metric keeps the original 14 day view methodology and counts every ad view in the window, ignoring the machine learning filter. 

You’ll find it in Amazon Ads unified reporting and in the Reporting tab of the DSP console. 

It exists for one job, which is an honest year over year comparison against performance before 2026.

How to Rebuild Your Amazon DSP Baseline

1. Pull Purchases (All Views) for the full comparison period. Use it as your baseline against 2025.
2. Rebuild any year over year deck that crosses the January 1 boundary. Any chart spanning that date without this metric compares two different definitions of a conversion.
3. Brief your internal stakeholders before they spot the gap themselves. Explaining a measurement change ahead of a budget review is a very different conversation from explaining it during one.
4. Track standard ROAS and multi touch ROAS side by side. Amazon launched Multi Touch Attribution in beta for US advertisers at roughly the same time. It spreads credit across the journey instead of handing all of it to the final interaction.
5. Read the gap between the two numbers. If multi touch ROAS runs well above standard ROAS, your upper funnel is doing work last click reporting can’t see.

Run the same maths on your own account with our Amazon ROAS calculator before and after applying Purchases (All Views). The gap between the two figures is the size of your reporting problem.

Here’s the pattern I keep seeing. The accounts that lost DSP budget this year were rarely the ones performing worst. They were the ones whose reporting nobody rebuilt.

Should Incrementality Testing Become Standard in 2026

Yes, and this is the year it stops being optional. Attribution tells you what Amazon’s model believes influenced a sale. 

Incrementality tells you what would have happened without the ad at all. Those are different answers, and only one of them survives a CFO conversation.

Geo holdouts and audience holdouts are the practical starting points. Hold back ads in a comparable geographic region or for a randomly selected group of shoppers, run the campaign everywhere else, and compare total sales rather than attributed sales. 

It’s slower than reading a dashboard, but it’s one of the most reliable ways to measure the true impact of your advertising because it’s not affected by attribution model changes. 

Why Amazon DSP Ads Work Even Without Ad Clicks

Most Amazon DSP ads never get clicked, and that’s normal. Display advertising and connected TV ads work through exposure and recall, not immediate response. 

A shopper sees your brand on Prime Video on Tuesday and searches for it on Saturday. The click that eventually happens looks like organic branded search.

According to Amazon Ads, Streaming TV campaigns in Canada delivered the following results compared with industry streaming benchmarks (Lucid, 2025; cited by Amazon Ads in its 2026 Upfront Canada announcement):

• 1.2× higher ad recall
• 2.3× stronger brand favourability
• 9.8× greater purchase intent

That’s why judging an awareness campaign on last click ROAS gives you the wrong answer every time. The metric was built for demand capture. Awareness does demand creation. Track these instead:

• Reach
• Frequency
• Video completion rate
• Detail page view rate
• Branded search movement

If a competitor retargets shoppers who viewed your listing, they’re not stealing a click. They’re intercepting the consideration window before your branded search ever happens. You’ll see it as flat branded search volume, not as a lost sale.

Amazon DSP vs Amazon Sponsored Ads: What Each One Actually Buys

For most enterprise brands, DSP and Sponsored Ads aren’t alternatives. Sponsored Ads primarily harvest demand that already exists. 

DSP creates demand and recovers it. Running one without the other caps growth at the size of current search volume.

Factor	Amazon Sponsored Ads	Amazon DSP
What it buys	Clicks on Amazon search and detail pages	Impressions across Amazon and the open internet
Targeting basis	Keywords, ASINs, categories	Audiences, behaviour, context, remarketing, third party data
Where ads appear	Mostly Amazon search and product detail pages	Amazon.com, Prime Video, Twitch, Fire TV, Kindle, Alexa, Netflix, Spotify, third party sites and apps
Core metrics	CPC, CTR, ACoS, ROAS, conversion rate	Reach, frequency, DPVR, new to brand, view through conversions, ROAS, AMC insights
Attribution in 2026	Click attribution unchanged. Sponsored Brands and Sponsored Display view credit now filtered	On Amazon delivery filtered by the new model. Offsite delivery still on the 14 day window
Pricing model	Cost per click	Cost per thousand impressions
Best for	Brands with strong existing search demand and margin discipline	Brands whose category search volume is capped, or who need retargeting and retention at scale

That attribution row is the one most comparisons leave out. It also explains why the two channels now move differently even when the same shopper journey drives both.

Sponsored Ads are strongest when a shopper is already searching on Amazon for what you sell. 

DSP is better when you need to reach that shopper earlier, or bring them back after they left without buying.

Should You Use Amazon DSP Before or After Amazon PPC?

After, almost always. Amazon PPC tells you which keywords convert, which ASINs carry the account, and what your real conversion rate looks like. 

That’s the input DSP needs. Going programmatic first means spending CPM budget to learn things CPC budget teaches faster and cheaper.

There’s one exception. A brand entering a category with almost no existing search volume has nothing to harvest. 

In that case DSP and streaming TV ads create the demand that Sponsored Products later captures. Launches into new categories fit this. Launches into established categories don’t.

Amazon DSP vs DV360, The Trade Desk and Viant

Amazon DSP is not automatically the right demand side platform. It is strongest when Amazon sales and retail signals sit at the centre of the strategy. When they don’t, another platform usually fits better.

Platform	Strongest for	Main advantage	Limitation versus Amazon DSP
Amazon DSP	Retail and commerce led brands	First party shopping, browsing, streaming and purchase signals	Weaker fit when Amazon is not central to sales
Google DV360	YouTube and Google led programmatic	Deep Google inventory access and YouTube scale	No Amazon retail purchase signals
The Trade Desk	Open internet buying at scale	Broad independent reach across channels	Commerce data must be integrated separately
Viant	People based omnichannel campaigns	Identity led targeting with strong CTV options	Less directly connected to Amazon retail behaviour

• If the outcome you are optimising toward is an Amazon purchase, Amazon DSP has data no competitor can replicate. 
• If the outcome is a site conversion, a lead or an offline sale, the advantage narrows fast and the other three deserve a serious look.

Plenty of enterprise brands run two. Amazon DSP for commerce outcomes, DV360 or The Trade Desk for broader brand campaigns. 

The mistake is running both against the same KPI and expecting the numbers to reconcile. They measure different things.

If Amazon’s AI Runs Your DSP Campaign, What Is Left for You?

Amazon now automates bidding, audience discovery, deal selection, and increasingly creative production. 

Brand+ and Performance+ optimize themselves. Ads Agent writes AMC queries from plain English.

Four advantages survive automation, and they’re the only four worth investing in.

1. The Quality of the First Party Data You Feed It

Audience Signals, the July 2026 feature that lets your customer file act as an optimisation input, is free and most brands still haven’t connected it. A model trained on your high value repeat purchasers learns faster than one starting cold. Your customer file is the one input competitors can’t copy.

2. Creative Volume and Variety Across Every DSP Ad Format

Automation optimises across what you give it. Give it three assets and it optimises across three assets. Dynamic formats and every display ad variant reward brands with deep approved libraries. Most enterprise brands are creative constrained rather than budget constrained.

3. Retail Readiness Before You Scale Amazon DSP Spend

No amount of AI fixes a detail page that doesn’t convert, a price that isn’t competitive, or a hero ASIN that goes out of stock in week three. Amazon DSP sends more traffic into whatever your product page already does, including its weaknesses.

4. Measurement Literacy Under the 2026 Attribution Model

This is the newest advantage and the most underrated. When the platform optimises itself, your edge is knowing what its numbers mean. January’s attribution change proved it. Two brands with identical performance had completely different internal conversations, and the difference was entirely who understood the metric.

More automation doesn’t reduce the need for strategy. It moves strategy upstream, from campaign settings to inputs.

What Amazon DSP Becomes When AI Assistants Do the Shopping

Product discovery is moving.Amazon has its own shopping assistant. Google puts AI Overviews above the results. The search box is losing its monopoly on discovery.

That has a real consequence for Amazon DSP advertising, and it cuts in an unexpected direction.

If AI systems narrow 10 options down to three, keyword bidding gets less useful. 

You can’t outbid your way into a recommendation the way you outbid your way into a search result. 

What decides inclusion is whether the shopper already knows and wants your brand on Amazon.

DSP stops being a media buy and becomes a demand generation function. Its output isn’t clicks. Its output is branded demand that shows up later as branded search, subscribe and save enrolment, and repeat purchase.

To be clear about what we’re claiming and what we aren’t. Amazon doesn’t publish any direct link between DSP spend and organic ranking. 

• The mechanism we can actually observe is simpler. 
• DSP exposure lifts branded search. 
• Branded search brings higher converting traffic to your product page on Amazon. 
• Higher conversion improves organic performance. 

The connection is real but indirect, and it takes quarters rather than weeks.

The planning version: if your category is heading toward assistant led discovery, budget that only harvests existing demand has a shelf life. Brands building recognition now will be the ones assistants have something to recommend later.

How to Build an Amazon DSP Strategy That Survives the Automation Layer

The strategy question in 2026 is what you feed the system and how you keep it honest.

Four inputs decide outcomes now.

1. Connect Your First Party Data Through Audience Signals

Audience Signals is the highest value free feature Amazon shipped this year. Hashed customer lists connect through Ads Data Manager and activate as advertiser audiences, or as custom segments built in Amazon Marketing Cloud. Every week you run Brand+ or Performance+ without it, budget funds cold start learning your own data could have shortened.

2. Build Audience Exclusions Before You Build Audiences

Audiences in the Amazon DSP program overlap constantly. One shopper can qualify as a product viewer, a category shopper, a past buyer and a competitor audience at the same time. Set these exclusions first:

• Exclude recent purchasers from new to brand campaigns
• Exclude converters from cart abandoner campaigns
• Exclude overexposed users from awareness lines

Without this, retargeting inflates prospecting results and you lose the ability to tell whether your top of funnel works at all.

3. Separate Prospecting From Retargeting Campaigns

This is usually taught as a best practice. It’s closer to a prerequisite. Combined campaigns look stronger than they are because warm audiences carry the average.

4. Govern Frequency Across Accounts, Not Within Them

Cross account reach and frequency reporting now shows audience overlap across multiple advertisers, marketplaces or business units. 

Enterprise brands running advertising at scale through several agencies or regional teams often pay three times to reach the same household. 

Frequency groups apply a shared cap across campaigns and stay badly underused.

Amazon’s AI Generated Image Tag Requirement for 2026 

Amazon now requires sellers to label listing images and A+ Content that use photorealistic AI-generated people. 

Three things matter operationally:

• The requirement follows New York’s synthetic performer law, effective June 9, 2026, plus EU AI Act obligations
• It applies at the file level before the asset reaches Amazon Seller Central
• It cannot be added retroactively once a listing is live

Brands producing AI generated lifestyle imagery at volume should audit their asset pipeline now.

Why Amazon Marketing Cloud Stopped Being Optional in 2026

Amazon Marketing Cloud is where all of this gets verified. AMC now holds 25 months of ad traffic signals, which finally makes a long window path to purchase analysis possible. Audiences activate once they reach 2,000 users.

If DSP is your demand generation layer, AMC is the only place you can prove it worked. Our work on media mix modelling and Amazon brand lift studies covers how to validate upper funnel contribution beyond platform reporting.

Amazon DSP Targeting Strategies and Where Your Ads Appear in 2026

Amazon DSP targeting strategies start with audiences, not keywords. You decide who you want to reach first, then decide how to show your ads to them.

Amazon DSP Audience Types That Actually Matter

Amazon DSP offers six main audience types:

• In market and behavioural audiences, built from shopping and browsing signals
• Contextual targeting, which matches the content environment rather than the individual user
• ASIN and competitor targeting, for shoppers comparing similar products
• Remarketing audiences, covering detail page viewers and cart abandoners
• Lookalike audiences, modelled on your existing buyers
• AMC custom audiences, built from event level data in Amazon Marketing Cloud

Three things are worth saying plainly about Amazon DSP audience selection, because they decide where the budget goes.

Most brands over invest in retargeting

It reports well because it reaches people already close to buying, so it always wins the ROAS comparison. That’s exactly why it’s a poor guide to budget allocation. If retargeting is more than half your spend, you’re mostly paying to close sales that were already closing.

Lifestyle and in market audiences serve different jobs

In market audiences work harder in consideration, because interest already exists. Lifestyle audiences work harder at the top, where you’re building recognition in people who haven’t entered the category yet. Judging a lifestyle audience on conversion rate will always make it look worse than it is.

ASIN and competitor targeting only works with a genuine comparison advantage: 

That means better reviews, clearer pricing, stronger creative or a real feature gap. Without one, you’re paying to send shoppers back into a competitor’s consideration set.

Amazon DSP Ad Types and Where They Run

Amazon DSP supports 5 creative formats, and the right DSP ad depends on the job, not on what assets you happen to have approved.

Format	Best used for	2026 notes
Static display	Retargeting, education, efficient reach	Still the volume workhorse on and off Amazon
Dynamic eCommerce ads	Product led retargeting	Pulls listing assets and returns shoppers to the detail page
Online video	Consideration, launches, demonstration	Strongest where the Amazon product needs explaining
Streaming TV	Upper funnel reach	Prime Video, Twitch, Fire TV, live sports, Netflix, plus Complete TV for cross publisher planning
Audio	Reach, recall, and now mid funnel performance	General availability for Brand+, open beta for Performance+, with Spotify, SiriusXM, iHeart and podcast supply

Amazon DSP inventory spans owned and operated placements, Amazon Publisher Direct, and premium third party sites and apps. Availability varies by region (Amazon Ads). 

On placement saturation, no public data ranks Amazon inventory by competitive pressure. What you can watch inside your own account is CPM drift against flat delivery. 

If CPMs climb while reach stays still, competition for that placement goes up and your money buys less of it. Track it per placement, quarter over quarter.

Audio is the format enterprise brands invest in most. 

It reaches commutes, workouts and household tasks that screen based formats can’t. With Performance+ audio still in open beta following the July 2026 update, auction competition is lower now than it will be at general availability. 

That makes Q4 2026 a sensible window to test it.

What Amazon DSP Costs and Who Should Use Amazon DSP

Access model	Built for	Minimum spend
Self service	Advertisers who want full control of setup, pacing and optimisation	Amazon does not publish a minimum for self service access
Managed service	Advertisers who want consultative support or have limited programmatic experience	Amazon states the managed service option typically requires a minimum spend of $50,000 USD, and notes this may vary by country
Agency or partner led	Brands that need strategy, execution and measurement support	Set by the partner, not by Amazon. Varies by scope, market and service model

Source: Amazon Ads

Amazon publishes only the managed service figure. Treat any number quoted as an official minimum for the UK or Canada with suspicion unless it links to Amazon’s own documentation. 

Amazon DSP Readiness Checklist for Enterprise Brands

Cost is the easier question. The harder one is whether you’re ready, and for enterprise brands that have almost nothing to do with listings. It’s about what you can measure. Work through these five before you scale Amazon DSP access:

• Baseline rebuilt using Purchases (All Views), so your year over year comparison is honest
• Customer lists cleaned, hashed and connected through Ads Data Manager
• Creative library stocked with more approved assets than dynamic and video formats will burn through
• Scheduled reports migrated ahead of the December 31, 2026 deadline covered below
• Hero ASIN detail pages converting the traffic you already send them

As a rough guide, we typically see brands above $1M in annual Amazon revenue get enough audience volume for DSP to work properly. Below that, Sponsored Display retargeting often does a similar job for less.

The last item on that list is the gate, not a checkbox. DSP amplifies whatever your product page already does. If conversion is the constraint, programmatic spend makes that constraint more expensive.

How Should You Split Amazon Ad Spend Across PPC, DSP and Streaming TV?

There’s no correct percentage, and anyone quoting one is selling something. There is a correct method.

Fund the bottleneck. Work out which constraint is actually capping growth.

What you’re seeing	The binding constraint	Where the budget goes
Losing impression share on converting keywords	Demand capture	Sponsored Products. Amazon ad spend on DSP won’t fix this
You own your keywords and growth has flattened	Demand creation	DSP and Streaming TV

Then hold the split for at least one full quarter. Upper funnel spend reported monthly always looks like waste, because returns arrive after the reporting period closes. 

Brands that reallocate every month never let demand generation compound. Effective advertising at this level is a quarterly discipline, not a monthly one.

Amazon DSP Advertising Across the US, UK and Canada

Access is broadly similar across these three markets. Inventory, audience availability and measurement maturity are where they differ.

	United States	United Kingdom	Canada
Supply breadth	Widest available. New features land here first	Mature, with established third party supply	Strong streaming inventory including Bell Media, CBC, Rogers, Netflix and Roku
Measurement	Multi Touch Attribution beta available to US advertisers only	Standard reporting. MTA not currently announced for this market	Standard reporting. MTA not currently announced for this market
Audience signals	90% authenticated household reach makes cross device measurement most reliable here	GDPR shapes how first party data is collected and activated	Deterministic signals from Amazon.ca, Fire TV and Prime Video sign ins
Streaming TV position	Large but highly competitive	Established Prime Video ad inventory	Top position in unique CTV reach among adults 18 plus versus other DSPs
What’s coming in 2026	Continued MTA rollout	Prime Video ads extend into Belgium, Denmark, Norway and Turkey, widening European reach	NBA arrives on Prime Video in October, plus expanded WNBA coverage and Canadian originals

Canada CTV position based on Numeris data, reported by Amazon Ads, June 2026.

Brands running all three should plan frequency centrally. 

Regional teams buying independently is the most common source of wasted reach we see in multi market accounts.

The December 31, 2026 Reporting Deadline Most Brands Haven’t Planned For

Amazon is shutting down the legacy sponsored ads and Amazon DSP report centres on December 31, 2026. 

Saved reports, scheduled reports and historical data still sitting on those pages will be permanently deleted. Unified reporting reached general availability on June 8, 2026 and is the replacement.

Date	What happens
June 8, 2026	Migration opens. Manual moves to the new reporting page begin
August 17, 2026	Self migration tooling becomes available
November 16, 2026	45 day warning issued
December 17, 2026	New report creation and editing disabled. Existing reports go read only
December 31, 2026	Old report centres decommissioned. Remaining reports, schedules and historical data deleted

Enterprise teams are the most exposed group. Scheduled sends, feeding BI (Business Intelligence) stacks, shared agency dashboards and automated stakeholder reports all break silently when the source disappears. 

Three actions before the end of Q3. 

1. Inventory every saved and scheduled report, including ones built by people who’ve left. 
2. Identify every downstream consumer of that data. 
3. Run the new unified schema in parallel for two weeks before switching, because the export shape changed even though the underlying data didn’t.

When Amazon DSP Is Not the Right Investment 

Amazon DSP isn’t the right fit for every brand. In some situations, you’ll see stronger returns by investing elsewhere. 

• The strict attribution model might simply be right: Under the old blind 14 day view window, a shopper who glanced at an ad and bought 13 days later handed DSP full credit. 

A lot of that credit was probably unearned. Purchases (All Views) restores comparability with 2025. 

It doesn’t restore accuracy. If your 2025 numbers were inflated, the honest move is to reset expectations rather than defend the old figure.

• Default on automation isn’t automatically bad: Brand Outcome Optimization swaps impressions for commerce behaviour as the optimisation target, which is a better target than most teams were using. 

Automatic Deal Selection removes manual deal work many in house teams were losing to anyway. 

The problem is reading the performance change they cause as your own optimisation work.

• DSP is the wrong buy for some brands, including large ones: If retail conversion is capped by pricing, review quality, or stock availability, DSP simply sends more traffic into an existing bottleneck. Increasing ad spend won’t solve problems on the product page. 

Conclusion

Amazon DSP didn’t fundamentally change in 2026. The way it’s measured did.

On January 1, Amazon introduced a new attribution model. In July, automation expanded, and DSP inventory continued to grow across streaming and other channels. Many brands are still comparing 2025 results with 2026 reporting, leading to misleading conclusions.

At the same time, Amazon DSP is becoming more than a programmatic advertising platform. As AI-powered shopping experiences evolve, DSP plays a bigger role in building demand before shoppers actively search for a product.

At Sequence Commerce, we’ve found that many performance issues come down to outdated measurement rather than poor campaign execution. Understanding the new attribution model and using the right baseline is essential for evaluating DSP performance accurately.

Frequently Asked Questions