Amazon’s retail media network crossed $62 billion in ad revenue in 2025, according to eMarketer’s Q1 2026 report — and the infrastructure behind that number looks almost nothing like it did two years ago. The company’s rollout of its AI-driven Demand-Side Platform 2.0, combined with a new audience signal layer called Amazon Marketing Cloud Insights (AMC Insights), has fundamentally changed how both first-party sellers and third-party marketplace operators need to think about paid visibility.
If you’re still running Sponsored Products campaigns the way you did in 2024 — bidding on keywords, adjusting manually, reviewing ACoS weekly — you are already behind. The sellers pulling 4x ROAS in mid-2026 are doing something structurally different. This guide breaks down exactly what that looks like.
What changed in Amazon’s ad ecosystem in 2026 that operators need to know about?
Three overlapping shifts happened between Q4 2025 and mid-2026 that most sellers underestimated at launch.
First, Amazon consolidated its campaign types under what it internally calls “Unified Performance Bidding” (UPB). UPB algorithmically blends Sponsored Products, Sponsored Brands, and Sponsored Display signals into a single optimization loop. You no longer set bids per placement in isolation — the system dynamically reallocates budget across formats in real time based on predicted conversion probability.
Second, AMC Insights became available to sellers at the $50K/year ad spend threshold (down from $250K). This democratized access to path-to-purchase data that was previously only available to brands like P&G and Unilever running full Amazon DSP campaigns.
Third, Amazon quietly sunsetted the old “Bid+” top-of-search multiplier in April 2026, replacing it with predictive placement scoring. Sellers who didn’t notice the deprecation have been paying blended CPCs 15–30% above the efficient frontier since May.
“The sellers who got hurt in Q2 weren’t outspent — they were out-structured. They had no idea Bid+ was gone, and their campaigns kept bleeding at placements they thought they’d capped.” — Megan Lau, Head of Marketplace Strategy at Tinuiti
How do you set up Amazon Marketing Cloud Insights for actionable data?
AMC Insights is a SQL-based analytics environment that sits on top of your Amazon ad data and links it to anonymized shopper behavior signals. Setting it up correctly is a prerequisite for everything that follows.
- Step 1: Get AMC access through your Amazon Ads console. If you’re above $50K annual ad spend, navigate to the “Measurement & Optimization” tab. You’ll see an AMC Insights activation card. Accept the data processing agreement — this takes 48–72 hours to process.
- Step 2: Connect your Brand Registry account. AMC pulls richer signal data when your storefront is linked. Organic browse events, wishlist adds, and subscribe-and-save conversions all feed into the attribution layer.
- Step 3: Run the new-to-brand vs. repeat purchase query first. This is the most immediately useful report. It tells you what percentage of your sponsored conversions are actually growing your customer base versus cannibalizing organic repeat buyers you would have kept without the ad spend.
- Step 4: Build an audience suppression list. Export customers who purchased in the last 90 days and suppress them from your Sponsored Display retargeting. This alone typically reduces wasted spend by 8–12% according to agency benchmarks from Pacvue’s Q2 2026 performance report.
- Step 5: Set up a weekly AMC Insights dashboard. Use Amazon’s pre-built query library — specifically the “Multi-Touch Attribution” and “Path to Conversion” templates. Schedule these to refresh every Monday morning before your weekly media review.
How does Unified Performance Bidding actually work — and how should you structure campaigns to work with it?
UPB is essentially Amazon’s answer to Google’s PMax: a single-budget, multi-format campaign type where the algorithm decides how to allocate spend across placements. The key difference is that Amazon’s version retains keyword-level targeting as a constraint, which gives sellers more structural control than PMax ever offered.
The mistake most sellers make is running UPB campaigns with overly broad match types and hoping the algorithm sorts it out. It doesn’t. The algorithm amplifies your structure — good structure gets better, bad structure gets worse, faster.
“We spent three weeks restructuring a home goods client’s campaign architecture before we even touched bids. Moved from 12 broad campaigns to 40 tightly themed exact-match clusters. UPB’s win rate on top-of-search went from 18% to 34% within six weeks.” — Jordan Fiske, VP of Retail Media at Bobsled Marketing
Here’s the campaign structure framework that’s working in Q3 2026:
- Cluster by purchase intent, not by product category. Group keywords by where a shopper is in the funnel (“best [category]” vs. “[brand name] [product]” vs. “[product] for [use case]”) rather than by SKU family.
- Use exact match for your top 20 converting terms. Keep these in isolated campaigns with manual CPC floors. UPB can still optimize placement, but you retain cost control on your highest-leverage terms.
- Let broad match run in a separate budget bucket — maximum 20% of total spend. This is your discovery layer. Review the search term report weekly and graduate high-converting terms to exact match clusters.
- Set portfolio-level budgets with daily caps per cluster. Portfolio budgets prevent UPB from over-indexing on one campaign type during high-competition periods like Prime Day recovery windows.
What does a healthy retail media P&L look like for a $5M Amazon seller in 2026?
The benchmark numbers have shifted significantly. Based on aggregated data from Intentwise’s Q2 2026 seller analytics report and conversations with agency operators, here’s what a well-run mid-market Amazon account looks like:
- Total Advertising Cost of Sale (TACoS): 10–14% of total Amazon revenue (blended across all ad types)
- Sponsored Products ACoS: 18–25% for established categories; 30–40% acceptable during launch windows
- Sponsored Display share of total ad spend: 12–18% (up from 6–8% in 2024, driven by AMC-informed audience targeting)
- New-to-brand conversion rate from sponsored: 40%+ is healthy; below 30% suggests significant cannibalization of organic repeat traffic
- DSP as a % of total ad budget: 15–20% for brands at $5M+ revenue; below this threshold, DSP is typically inefficient
Sara Glick, founder of the DTC consulting firm Cartograph, recently shared internal benchmarks from her agency’s book of business: “The sellers at 12% TACoS are not spending less — they’re spending more efficiently. They’ve got AMC suppression lists running, they’re not burning budget on existing customers, and their creative rotation on Sponsored Brands Video is actually tied to what’s converting in their A/B tests.”
How do you use Amazon’s new AI creative tools without losing brand control?
In March 2026, Amazon rolled out generative AI creative tools directly inside the Ads console — specifically for Sponsored Brands and Sponsored Display. The system can auto-generate lifestyle imagery, headline copy variations, and video stills from your product ASIN data.
The tools are genuinely useful for velocity — you can produce 20 creative variants in the time it used to take to brief an agency for one. But sellers who deploy AI-generated creative without governance are seeing brand consistency problems and, in some cases, imagery that doesn’t match their actual product colorways.
Here’s a practical governance protocol:
- Lock your brand kit in the Amazon Brand Registry creative hub first. Upload your brand color palette, approved font stack, and a library of approved lifestyle images. The AI creative tool will use these as constraints when it generates variants.
- Never auto-publish AI creative without a 48-hour human review window. Set campaign end dates two days out from launch to force a review checkpoint.
- A/B test AI creative against your control creative, not against nothing. Use Amazon’s built-in campaign split testing to measure click-through rate and conversion rate lift. AI creative tends to win on CTR but sometimes loses on conversion — especially in premium categories where brand trust is a purchase driver.
- Use AI creative aggressively for retargeting (Sponsored Display). Personalized lifestyle imagery in retargeting contexts outperforms static product shots by 22% on average, per Amazon’s own published benchmark data from May 2026.
What are the biggest mistakes sellers are making with Amazon’s AI ad tools right now?
The most common operational failures in Q3 2026 aren’t strategic — they’re executional. Sellers are adopting the new tools without updating their measurement frameworks.
“Everyone’s excited about UPB and AMC, but I’m seeing the same mistake everywhere: sellers are measuring success with the old ACoS lens. You cannot evaluate an AI-optimized campaign with a 2022 reporting framework. You have to move to contribution margin by traffic source, or you’re flying blind.” — Megan Lau, Tinuiti
The three most damaging mistakes operators are making right now:
- Measuring UPB campaigns on ACoS instead of TACoS. Because UPB influences organic rank lift through sales velocity, its true ROI is invisible if you’re only looking at attributed ad revenue.
- Not suppressing existing customers from retargeting pools. This is free money left on the table. AMC makes this easy — there’s no excuse for running it without suppression lists in place.
- Treating AI creative as a cost-cutting measure rather than a testing velocity tool. The sellers winning with AI creative are running 5x more tests than their competitors, not spending 5x less on creative production.
Amazon’s retail media platform in 2026 is genuinely more powerful than anything that existed two years ago — but that power is asymmetric. Sellers who invest in AMC setup, campaign architecture discipline, and creative testing infrastructure will compound advantages quarter over quarter. Those who default to the auto-campaign and hope for the best will find that the algorithm is very efficient at spending their budget without returning it.
The window to build structural advantage before Q4 2026 is roughly eight weeks. That’s enough time to get AMC configured, restructure your top 10 campaigns into intent-clustered exact-match groups, and run your first suppression-list test. Start there.
