Amazon Advertising Management

TL;DR — Key takeaways

Amazon’s ad revenue crossed $56B in 2024 — but the brands winning aren’t the ones spending most. They’re spending at the right layer.
ACoS measures campaign efficiency. TACoS measures whether you’re actually growing. Most brand teams track the wrong one.
Running Sponsored Products, Sponsored Brands, and Sponsored Display as three separate budgets with separate logic guarantees funnel leakage — full-funnel coordination typically closes a 25–30% performance gap.
Rufus, Amazon’s AI shopping assistant, now influences which sponsored results surface for conversational queries — and it indexes A+ Content signals, not just bids.
The most overlooked fix: audit listing readiness before scaling spend. Ad budget against a weak listing is money deposited directly into a competitor’s account.

A brand manager once shared their Amazon dashboard with me the week before Black Friday. Campaigns live. Impressions strong. ACoS at 18% — well within targets. Net Amazon revenue had been flat for three consecutive quarters. What surprised me wasn’t the stagnation. It was how cleanly the numbers hid it.

This is the central problem with how most brands run Amazon advertising: they optimize the instrument panel while the engine runs rough. ACoS tells you whether your ads are efficient. It says nothing about whether they’re working.

Why ACoS Is the Most Misleading Metric on the Dashboard

A 15% ACoS in a category where your organic rank is collapsing is a disaster dressed up as a win. Advertising Cost of Sales divides ad spend by ad-attributed revenue — it tells you nothing about organic cannibalization, nothing about per-unit margin, and nothing about whether you’re building real demand or renting shelf space.

The metric that matters is TACoS — Total Advertising Cost of Sales — which divides total ad spend by total revenue, organic and paid. When TACoS falls as revenue grows, your advertising is doing its job: building demand that converts without paid support. When TACoS rises and ACoS looks clean, your ads are propping up listings that can’t hold their own weight.

What we see at Epinium is that most brand managers inherit a framework built for direct-response performance marketing — CPL, CPA, ROAS — and bolt it onto Amazon, where the rules are different. A click that doesn’t convert on Amazon doesn’t just waste budget. It sends a negative conversion-rate signal that actively suppresses your organic rank on the very keywords you just bid on. That feedback loop — spend → no convert → rank drop → spend more to recover — is the engine behind most stagnating Amazon P&Ls.

62% of Amazon Ad Budgets Fund Listings That Can’t Convert

Analysis from Amazon Advertising’s own performance benchmarks and third-party audits presented at Amazon Unboxed 2025 points to the same uncomfortable finding: the majority of Sponsored Products spend flows to product detail pages that fail basic conversion readiness — fewer than 15 reviews, no A+ Content, main images missing lifestyle shots, or bullet points that lead with features instead of outcomes.

Ad spend is not a substitute for listing quality. It’s an amplifier. If your listing converts at 8% with organic traffic, advertising pushes more visitors through an 8% filter — and you pay for every one who decides against buying. If your listing converts at 22%, every dollar compounds. Amazon’s own published data shows A+ Content lifts conversion rates by 3–10% compared to standard listings. Sponsored Brands video drives 108% higher click-through rates than static Sponsored Brands on mobile — where more than 60% of Amazon purchases originate.

The implication is uncomfortable: before increasing advertising budget, run a listing readiness audit. Not once a year. Before every significant spend increase. This is the step most performance-focused teams skip because it doesn’t show up in the campaign dashboard.

Vendor vs. Seller: The Advertising Reality Nobody Maps

Most Amazon advertising guides treat all sellers as one category. They’re not. Brands on Vendor Central — manufacturers and distributors who sell wholesale to Amazon — operate in a fundamentally different advertising environment than Seller Central brands, and conflating the two is a persistent strategic error.

Vendors have exclusive access to Amazon Marketing Cloud’s full SQL attribution workbench, Demand-Side Platform inventory for programmatic targeting, and vendor-funded coupons. They also face a problem Seller Central brands don’t: they don’t control pricing. Amazon sets retail prices algorithmically. A PPC-heavy vendor can find themselves advertising a product at a price point Amazon has marked down 30% — where every ad conversion generates negative margin.

The correct approach for vendors is to tie advertising strategy to AMC data — specifically, understanding new-to-brand rates and customer lifetime value by ad format, not just immediate ROAS. Vendors running DSP alongside Sponsored Products typically see new-to-brand percentages 30–40% higher than Sponsored Products alone. That’s audience compounding over time. For a deeper breakdown of how vendors should approach their full Amazon presence, our analysis of what AI tools actually mean for Amazon brand management covers the structural shifts now in play.

For Seller Central brands, the gap is different: most lack vendor-grade attribution tooling. The workaround is building a TACoS dashboard at the ASIN level, and setting targets by lifecycle stage — high (15–20%) for new launches, medium (10–12%) for growth, low (6–8%) for mature catalog items where organic rank is strong.

30%

reduction in wasted ad spend when AI bid automation replaces manual CPC management

Source: Amazon Advertising, 2025

Amazon Advertising in 2025–2026: What Actually Changed

Rufus Reshapes Demand Signals

Amazon’s AI shopping assistant Rufus, rolled out across the US and EU through 2024–2025, now handles a material share of product discovery. Rufus doesn’t match keywords — it interprets intent from conversational queries and surfaces Sponsored Products based on A+ Content descriptions, reviews, and backend attributes. Brands that optimize A+ Content with use-case language — not just feature specs — are seeing incremental impression share from Rufus-mediated results that never appears in keyword-level reports.

Amazon Marketing Cloud Goes Broad

AMC’s SQL-based attribution workbench opened to all Sponsored Ads brands in 2024, previously restricted to DSP spenders. Brands can now run multi-touch attribution queries — understanding whether a Sponsored Display touchpoint 14 days before conversion was incrementally valuable, or noise. The brands already building AMC query libraries are compounding a data advantage their competitors don’t see yet.

AI Creative Tools Arrive — With a Trap

Amazon’s Creative Studio and AI image generation became broadly available in 2025. The trap, flagged at Amazon Unboxed 2025: brands that wholesale replace tested creative with AI-generated assets see conversion rate drops in the first 30 days. The correct use is augmentation — generate variations, split test, promote winners. Not wholesale replacement overnight.

Portfolio-Level Budget Controls Ship

Amazon rolled out portfolio-level budget caps and intelligent pacing in 2025, fixing years of campaign-level daily budget exhaustion at 2pm on high-traffic days. Brands that haven’t migrated their architecture to portfolios are still operating with a blunt instrument where they could have a scalpel.

The Three-Layer Campaign Structure That Actually Scales

The highest-performing Amazon advertising accounts are defined by structure, not scale. Here’s the architecture that works across brand size and category:

Layer 1 — Harvest. Automatic Sponsored Products at conservative bids. Purpose: discover converting search terms you didn’t know existed. Run for 3–4 weeks, pull the Search Term Report, promote winners to Layer 2. This is a listening layer, not a revenue layer.

Layer 2 — Precision. Manual Sponsored Products with Exact and Phrase match on proven converters, plus Sponsored Brands in headline position for branded and category-intent queries. This is your revenue engine. Active bid optimization, weekly review, tight targets.

Layer 3 — Brand Build. Sponsored Brands video and Sponsored Display for retargeting cart abandoners, cross-sell to existing buyers, and reaching shoppers viewing competitor pages. TACoS contribution here is lower; its job is repeat purchase rate and consideration.

Here’s where most brands get it wrong: they treat Layer 3 as optional overflow when budget remains. For any brand in a competitive category with more than 50 SKUs, Layer 3 is how you prevent slow erosion of repeat customers to private label. Cutting it to defend a short-term ACoS target is a structurally expensive decision that never shows up in the quarterly report that triggered it.

Epinium data

Across brands managed on Epinium Platform, products with an A+ Content quality score above 80% achieve a 22% lower ACoS at equivalent bid levels compared to identical SKUs with basic listings. The effect is most pronounced in beauty, nutrition, and home categories — precisely where Rufus query traffic is highest. Raising content quality before increasing bids is consistently the highest-ROI action available to our clients’ catalog teams.

Precision vs. Automation: Which Bidding Approach Fits Your Account

Approach	Best For	Main Tradeoff	Typical ACoS
Manual Exact Bidding	Mature catalogs, known converters	Time-intensive; misses emerging terms	8–18%
Dynamic Bidding (Down Only)	Budget-constrained, margin-sensitive SKUs	Limits impression share in auctions	10–20%
Dynamic Bidding (Up and Down)	Growth-phase products, aggressive rank targets	Higher spend variance, needs daily watch	12–25%
Portfolio Automated Rules	Large catalogs (100+ SKUs)	Needs 8+ weeks of data to stabilize	10–22%
AI-Powered Rules (Epinium)	Multi-market brands, vendor/seller hybrid	Setup investment upfront; compounds over time	6–14%

FREE TRIAL

See How Epinium Manages Amazon Advertising at Scale

AI-powered bid rules, catalog scoring, and cross-ASIN attribution — all connected to your live Amazon data. No spreadsheets. No guesswork.

Explore Platform → ✓ 7 days free ✓ No credit card ✓ Your own data

9 Questions Brands Actually Ask About Amazon Advertising

What is the difference between ACoS and TACoS, and which should I report?

ACoS is campaign-level: ad spend divided by ad-attributed revenue. TACoS is strategic: total ad spend divided by total revenue including organic. Report ACoS to your campaign manager for tactical decisions. Report TACoS to your leadership — it’s the only metric that shows whether advertising is building organic momentum or masking a listing that can’t survive without paid support. A falling TACoS as revenue grows is the actual proof of concept leadership wants.

Should established brands start with automatic or manual campaigns?

The standard advice — always start with auto — is correct for new sellers with no conversion history. It is wrong for established brand manufacturers who already know their category search terms and have 12+ months of product performance data. In that case, launch Exact Match manual campaigns on your top 10 converting terms first. Run auto in parallel at 20% of your manual budget as a discovery tool — not as your primary vehicle. Defaulting to auto when you already have data is burning weeks of budget to relearn what you already know.

My ACoS looks healthy but Amazon revenue is flat. What is wrong?

Three likely culprits. First, you’re spending predominantly on branded keywords — capturing demand that would have converted organically anyway. Pull your Search Term Report and check what share of attributed revenue comes from your brand name. Second, your Sponsored Products may be running against ASINs with weak organic rank, creating an artificial visibility dependency for listings that don’t earn their placement. Third, you have no Sponsored Brands or Display presence, meaning every euro you spend reaches only people who already intended to buy from you. That’s not growth — that’s recirculation.

How do I know if a listing is ready for advertising spend?

Before scaling any campaign, verify: main image on white background, secondary images include lifestyle and infographic, title has primary keyword in first 80 characters, bullet points lead with outcomes not feature specs, A+ Content is active (not text-only default), review count is above 15, and star rating is above 4.0. If three or more of these fail, the listing fix generates better ROI than any bid adjustment. Every click on a weak listing teaches Amazon’s algorithm that the page doesn’t satisfy its traffic — a signal that persists after you eventually fix the content.

What is Amazon Marketing Cloud and when does a brand actually need it?

AMC is Amazon’s SQL-accessible data clean room for multi-touch attribution, audience overlap analysis, and customer journey mapping across ad formats. You need it when: spend exceeds €5,000/month on Amazon Ads, you’re running both Sponsored Ads and DSP, or you want to understand what percentage of your conversions are genuinely new-to-brand versus repeat purchasers already in your funnel. Below that threshold, the standard Attribution dashboard and Search Term Report are sufficient for monthly optimization cycles.

How should vendor brands handle the pricing control problem in their ad strategy?

Vendors who can’t control retail price need campaign rules that pause or reduce spend when Amazon’s algorithmic pricing drops products below margin-safe thresholds. Set up price-monitoring alerts via Epinium Platform or Helium 10 and connect those alerts to portfolio budget rules. A vendor spending €2,000/day on Sponsored Products for an item Amazon has temporarily discounted 35% below MAP is actively subsidizing their own margin destruction. The fix isn’t a larger budget — it’s smarter rules that respond to retail price in real time.

What is the right budget split across Sponsored Products, Sponsored Brands, and Sponsored Display?

A defensible starting structure for most brand manufacturers: 65–70% Sponsored Products as the revenue layer, 20–25% Sponsored Brands for headline visibility and branded recall, 10–15% Sponsored Display for retargeting and cross-sell. In highly competitive categories, the Sponsored Brands share rises. In categories with long consideration cycles — supplements, electronics accessories — Sponsored Display earns a larger portion because repeat exposure meaningfully drives conversion within 14-day attribution windows.

Can AI fully automate Amazon advertising management?

Substantially, not completely. AI excels at bid optimization, negative keyword harvesting, and budget pacing — the repetitive data-dense work humans do slowly and inconsistently. Amazon’s Dynamic Bidding combined with portfolio automated rules handles 70–80% of routine campaign management reliably. What AI cannot replace: category strategy, launch sequencing for new products, competitive response when a rival runs a flash promotion, and the judgment call of whether a declining TACoS reflects a healthy catalog or a shrinking category. AI is the execution layer. Humans own the strategic frame.

What happens to Amazon advertising when an AI agent does the buying?

This is the question most advertising teams aren’t asking yet — which is precisely why it matters. Rufus and emerging AI shopping agents don’t scan results the way humans do. They synthesize product information from A+ Content, reviews, Q&A, and structured attributes to generate recommendations, then surface Sponsored Products when targeting aligns with conversational query intent. Brands that invest in structured, use-case-driven content alongside bid management will have compounding advantage in an AI-mediated commerce environment. Brands optimizing purely for keyword density will find themselves invisible to the next generation of shopper. The Epinium Platform tracks content quality scores per ASIN, flagging listings at risk before you lose the impression share.

The temptation in Amazon advertising management is always to do more: more campaigns, more targets, more rules, more creative variants. The data from thousands of accounts and hundreds of millions in managed spend consistently points the other direction. The highest-performing accounts are defined by ruthless simplification — fewer ASINs with excellent listings, two or three proven campaign structures, TACoS targets aligned to actual strategic objectives per product, and the discipline to pause spend when the catalog layer isn’t ready to convert it.

Amazon advertising is entering its most structurally complex phase: Rufus reshaping demand signals, AMC enabling real attribution for the first time, AI creative generation arriving faster than most teams can evaluate it. The brands building the right foundation now are not the ones who will be chasing competitors in 2027. They are the ones competitors will be studying.

EPINIUM PLATFORM

Manage Your Entire Amazon Ad Stack Without Losing a Day to Spreadsheets

Brand managers using Epinium have cut wasted ad spend by up to 30% with AI-powered bid rules, catalog scoring, and cross-ASIN attribution.

Start free →

No credit card · 7 days free · Your own data