Amazon Performance+ Explained: How AI Full-Funnel Campaigns Work

A Bit of Context Before We Dive In

I remember the first time a client asked me about Performance+. It was early 2025, and they had just sat through an Amazon rep presentation that made it sound like a magic button. Set it up, let the AI run, and watch your ROAS climb. The reality, as with most things in Amazon advertising, is a bit more nuanced than that.

That said, after running Performance+ campaigns across multiple brand categories over the past year, I can say with confidence that it is genuinely one of the most interesting tools Amazon has released in a long time. When used in the right context, it does things that manual campaign management simply cannot match at scale.

This article walks you through exactly what Performance+ is, how the AI behind it actually works, where it fits in a brand’s broader advertising strategy, and what kind of results you can realistically expect going into 2026. No hype, no oversimplification — just a clear picture so you can make a good decision for your brand.

What Is Amazon Performance+?

Performance+ is an AI-powered campaign type that lives inside Amazon DSP. Unlike traditional DSP campaigns where you manually define audiences, set bids, and choose placements, Performance+ hands most of those decisions to Amazon’s machine learning models.

The system is built on something Amazon calls AdRelevance — a predictive targeting model that draws on Amazon’s first-party data from shopping, streaming, and browsing behavior. Instead of asking you to pick who to target, it continuously analyzes signals across hundreds of millions of Amazon interactions to predict which users are most likely to convert, and when.

It was officially launched globally in early 2025 alongside its awareness-focused counterpart, Brand+. The two tools are designed to work together — Brand+ builds the top of the funnel, Performance+ closes the bottom of it.

How the AI Actually Works

This is where most articles skip the important part. Let’s be specific about what Performance+ is actually doing.

When you launch a Performance+ campaign, you provide a few key inputs: your product ASINs, your target goal (typically ROAS or new-to-brand purchases), and your budget. From there, the system takes over and handles the following automatically.

The key thing to understand is that Performance+ is not a black box. Amazon describes it as a ‘glass box’ approach — you can see insight cards that explain which audience signals are driving delivery, how long users typically take to convert after seeing an ad, and how the creative is rotating. You’re not flying blind; you’re just handing the controls to a system that can process signals faster than any human team.

The Signal Stack

What makes AdRelevance different from other programmatic platforms is the depth of data Amazon has access to. It’s not just browsing behavior — it’s purchase history, search queries, category interest patterns, Subscribe & Save subscriptions, Prime Video viewing habits, and Alexa interaction signals, all stitched together into a probabilistic conversion model.

That combination is genuinely hard to replicate anywhere else. Meta has social graph data. Google has search intent. Amazon has purchase intent — the most commercially valuable signal that exists in digital advertising.

Performance+ vs. Manual DSP: What Actually Changes

If you’ve run traditional DSP campaigns before, the shift to Performance+ feels significant. Here’s how the two approaches compare in practice.

The honest trade-off is this: Performance+ gives up some control in exchange for speed, scale, and the ability to act on signals no human team can process manually. For brands with strong data signals — good conversion rates, established purchase history, meaningful ad spend — that trade is often worth it.

For brands that are very early stage, or running very niche product categories with limited data, the AI has less to work with and Performance+ may take longer to find its rhythm.

Performance+ and Brand+: Two Tools, One Funnel

Amazon designed Performance+ and Brand+ as a pair. Understanding how they work together is important because using just one of them is a bit like running a PPC campaign with no awareness layer — you’re capturing demand that someone else built.

Amazon’s own case study with H&R Block showed what happens when both tools are used together. By running Brand+ and Performance+ in tandem during tax season, H&R Block saw a 144% lift in full-funnel conversion rate and a 35% improvement in cost per acquisition. (Source: Amazon Ads, November 2025)

PepsiCo reported 4x ROAS in prospecting and 2x ROAS in remarketing when running Performance+ campaigns across their in-aisle shopper segments. (Source: Amazon Ads official announcement)

Which Brands Are Actually a Good Fit for Performance+?

Performance+ is not for everyone, and I’d rather be honest about that than oversell it.

The brands that get the most out of it tend to share a few characteristics. They’re not always the biggest brands on Amazon — but they do have the right data signals for the AI to work with.

Good candidates

• Established brands with 6 or more months of Amazon purchase history and a track record of conversions
• Products with strong listing quality — good images, A+ content, 50+ reviews, conversion rate above category average
• Brands in competitive categories where manual audience testing has become expensive and slow
• CPG, beauty, health, and household brands with repeat purchase signals that the AI can model
• Advertisers already running DSP who want to test a more automated layer alongside manual campaigns

Brands that should wait

• New-to-Amazon sellers with fewer than 90 days of sales history — the AI has too little to learn from
• Very niche or technical products with a small total addressable audience on Amazon
• Brands where listings still need work — low review counts, below-average conversion rates, thin A+ content
• Sellers looking for granular manual control over every targeting decision

How to Set Up a Performance+ Campaign

The actual campaign creation process is simpler than traditional DSP setup, which is partly the point. Here’s how it looks in practice.

1. Open Amazon DSP and create a new campaign order.
2. Select Performance+ as the campaign type.
3. Choose your campaign goal — typically ROAS, new-to-brand purchases, or detail page views.
4. Select the ASINs you want to promote. Start with your top-converting products, not your full catalog.
5. Set your budget. Amazon recommends a daily budget that allows meaningful delivery — too little and the learning phase stretches out.
6. Configure brand safety settings and any suppression lists (for example, excluding existing customers if you want a pure acquisition focus).
7. Choose your inventory preferences — Amazon-owned supply, Prime Video, third-party premium publishers, or a mix.
8. Upload your creative assets. Performance+ can also auto-generate creatives from your product listings, but manually designed assets almost always outperform auto-generated ones.
9. Launch and let the system run for at least 14 days before drawing any conclusions.

Creative: Where Most Brands Leave Money on the Table

The AI handles targeting and bidding. It cannot fix bad creative. And in my experience, creative quality is the single biggest variable separating Performance+ campaigns that run well from those that underperform.

What tends to work in Performance+ creative

• Clean product-forward images where the product is immediately recognizable at small sizes
• A benefit-focused headline (what the product does for the user, not what it is)
• A clear call-to-action — ‘Shop Now’, ‘See It On Amazon’, ‘Get Yours Today’
• Lifestyle context images showing the product in use — these outperform plain white-background shots in display placements
• For video formats: front-load the key message in the first 3 seconds, as many placements are skippable or scroll-past

Common creative mistakes to avoid

• Overcrowding the ad with multiple messages or product claims — pick one thing
• Using the same creative for awareness and retargeting audiences — someone who has already seen your product 3 times doesn’t need the same introductory message
• Low-resolution images that look blurry on high-DPI screens
• Generic ‘best seller’ claims with no proof — they’ve become invisible to shoppers

Measuring Performance+: The Metrics That Actually Matter

This section matters a lot. Performance+ is often misjudged because brands look at the wrong numbers.

Direct attributed ROAS from Performance+ will typically look lower than what you see in a Sponsored Products campaign. That’s partially because display advertising operates earlier in the shopping journey — a click on a display ad does not carry the same purchase intent as a click on a search result. Judging the two by the same ROAS standard is like comparing a TV spot to a Google search ad and declaring the TV spot a failure.

A note on Amazon Marketing Cloud (AMC): If you’re running Performance+ without AMC, you’re missing half the picture. AMC lets you run custom queries that show how Performance+ touchpoints influenced purchases that were ultimately attributed to Sponsored Products. In most cases, we find that DSP plays an assist role that the standard attribution window simply doesn’t capture.

Realistic Expectations for the First 90 Days

Let’s be specific, because generic timelines don’t help anyone plan a budget.

Brands that cut Performance+ campaigns after 30 days because ‘ROAS looks low’ are consistently the ones that come back six months later wishing they hadn’t. The learning phase is real. It costs money. But the alternative — never building a scalable DSP engine — costs more in the long run.

The Most Common Performance+ Mistakes Brands Make

1. Launching with too small a budget

The AI needs delivery to learn. A $500/day budget spread across multiple ASINs gives each one almost nothing to work with. Concentrate your initial budget on 2-3 products and let those campaigns generate real data.

2. Including your full SKU catalog from day one

Starting with 40 ASINs sounds logical — more products, more chances to convert. In practice, it dilutes the signal. Performance+ learns faster when it can focus. Start with your top 3-5 converters and expand once the campaign is proven.

3. Changing settings every week

If you adjust budget, swap creative, or change campaign settings during the learning phase, you essentially restart the algorithm’s optimization cycle. Set it up well, then leave it alone for at least 14-21 days before making any meaningful changes.

4. Judging it against Sponsored Products ROAS

These are fundamentally different campaign types doing different jobs. Comparing their direct ROAS is misleading. Use AMC to look at assisted conversions and full-funnel impact — that’s the only fair comparison.

5. Running Performance+ with no awareness layer

Performance+ is built to convert — but it needs people who have some awareness of your brand or category to work with. Running Brand+ or Sponsored Brands Video alongside it gives the conversion campaign a much warmer audience to work with.

Frequently Asked Questions

The Bottom Line

Amazon Performance+ is not a shortcut. It’s a tool that rewards brands who come to it prepared: with strong listings, an established conversion history, good creative assets, and the patience to let the algorithm learn.

What it offers in return is genuinely valuable — targeting precision at a scale that no manual team can match, real-time bidding on conversion signals that go far deeper than anything available in Sponsored Ads, and a path toward building a full-funnel advertising system that grows your brand beyond the search results page.

Used well, alongside Brand+ for awareness and AMC for measurement, it can change the trajectory of a brand on Amazon. Used badly — with too little budget, poor creative, or impatience during the learning phase — it will look like it doesn’t work, even when the potential is there.

If you want to understand whether Performance+ makes sense for your specific situation, and how to structure it within your existing campaign mix, that’s a conversation we’re happy to have.