TNOMADS / Services / Advanced Analytics

Advanced Analytics

Amazon Marketing Cloud — answers
your reports can't give.

Privacy-safe clean room analytics that reveals true path-to-purchase, incremental reach, and audience insights — then activates them directly in DSP campaigns.

What is Amazon Marketing Cloud?

Amazon Marketing Cloud (AMC) is Amazon Ads' privacy-safe data clean room. It holds event-level, pseudonymised signals from your Amazon DSP and Sponsored Ads campaigns, which you query with SQL to analyse path to purchase, channel overlap, frequency and incrementality — then build custom audiences and activate them into DSP.

Type: Data clean roomQuery language: SQLOutput: Analysis + audiencesActivation: Amazon DSP

What is Amazon Marketing Cloud?

Amazon's privacy-safe clean room for advertisers.

Amazon Marketing Cloud (AMC) is a secure, privacy-safe data clean room that gives advertisers access to event-level, pseudonymised signals from their Amazon Ads campaigns — enabling analytics and audience creation that standard campaign reports cannot provide.

Launched in 2021, AMC has evolved significantly. The 2025 AI SQL Generator (announced at CES 2025) eliminated the need for data engineering expertise, allowing advertisers to describe audiences in plain English and receive working SQL queries in minutes. In November 2025, Amazon launched the Ads Agent for AMC — built on Amazon Bedrock — enabling full natural language interaction for both analytics and audience creation.

AMC now retains ad traffic data for 24 months (up from 13 months), enabling year-over-year analysis and seasonal audience building that was previously impossible. Custom audiences created in AMC can be activated directly in Amazon DSP and the Ads Console.

01

Custom Analytics

SQL-based reporting environment where advertisers can query pseudonymised ad signals — impressions, clicks, conversions — across Sponsored Ads and DSP to build analyses that standard reports cannot answer.

02

Audience Creation & Activation

Build custom audience segments from AMC query outputs and push them directly to Amazon DSP campaigns. The AI Audience Generator (July 2025) enables this via natural language — no SQL required.

03

First-Party Data Integration

Connect your CRM or CDP data to AMC via Ads Data Manager for secure, pseudonymised overlap analysis — enabling purchase cycle analysis, lapsed customer identification, and high-value segment targeting.

04

Ads Agent (AI Assistant)

Announced at unBoxed 2025, built on Amazon Bedrock. Generates analytics SQL, creates audience queries, and provides real-time AMC product guidance — all through conversational natural language interaction.

What AMC can answer

Questions your campaign reports cannot answer.

PATH-TO-PURCHASE

Multi-Touch Attribution

Which combination of ad touchpoints — DSP impressions, Sponsored Product clicks, Sponsored Brand views — actually leads to conversion? AMC maps the full path so you understand which channels create demand and which capture it, informing both budget allocation and creative strategy.

Media OverlapTouchpoint SequencingCross-Channel
NEW-TO-BRAND

NTB Analysis & Cohort Tracking

Amazon's standard NTB metric has limitations. AMC allows custom NTB analysis with configurable lookback windows, purchase frequency thresholds, and month/year dimensionality — giving you a true picture of how your media is growing your customer base, not just attributing to repeat buyers.

NTB RateCohort AnalysisRepeat Purchase
FREQUENCY

Reach & Frequency Optimisation

How many times did a unique shopper see your ads across DSP and Sponsored Ads combined? AMC deduplicates reach across ad types to reveal true frequency — identifying both under-exposed audiences (opportunity) and over-exposed ones (waste).

Unique ReachFrequency CapsDSP + Sponsored
5 yrretail purchase lookback for measurement use cases, expanded from 13 months [1]
Mar 2026Amazon added Ads Agent query-editor controls to AMC for SQL editing, explanation and repair [2]
No SQLrequired to start — templates, guided workflows and natural-language query generation are now available [3]
DSPaudiences built in AMC activate directly into Amazon DSP line items

The TNOMADS approach

AMC is the bridge between data and decisions.

🧬

Query library, not one-off analysis

We build a reusable AMC query library for each client — NTB tracking, geo attribution, path-to-purchase, media overlap, and frequency analysis — so insights compound over time rather than requiring fresh builds each month.

🎯

Audience activation pipeline

We close the loop between AMC insight and DSP activation — converting query outputs into live audience segments that improve targeting precision, reduce wasted impressions, and create compounding performance gains.

📈

Budget decisions from AMC, not ROAS

Standard ROAS is a correlation metric. We use AMC's path-to-purchase and media overlap data to identify which channels are causing conversions vs. claiming credit for them — and allocate budget accordingly.

The vocabulary

What are the core concepts in Amazon Marketing Cloud?

Clean room
A secure environment where event-level data can be analysed without any party accessing raw user-level records. Results are returned only in aggregate, above minimum-count thresholds, which is what makes the analysis privacy-safe.
Event-level signals
Individual impressions, clicks, detail page views and purchases, each timestamped and linked by a pseudonymous identifier. This is what makes sequencing and overlap analysis possible — and what standard reports flatten away.
Instance
Your dedicated AMC environment, holding signals from your own advertising. Data does not leave it and cannot be joined to another advertiser's instance.
Lookback window
How far back signals are available. Amazon expanded the retail purchase lookback to five years for measurement use cases, materially improving lifetime value and new-to-brand analysis for seasonal and long-lifecycle products.
Path to purchase
The ordered sequence of ad exposures preceding a conversion. The analysis that shows whether DSP is genuinely opening journeys that Sponsored Products later closes, or simply arriving late and taking credit.
Audience overlap
How much of your reach is being paid for twice across channels. Frequently the single highest-value first query in a new AMC instance.
Effective frequency
The number of exposures beyond which additional impressions stop producing incremental conversions. Identifying it is how frequency caps get set with evidence instead of convention.
Custom audiences
Segments defined by SQL against your own event data — for example, shoppers who viewed but did not purchase within a window — then activated directly into Amazon DSP.
New-to-brand (NTB)
Purchasers with no prior purchase of your brand in the lookback period. AMC lets you analyse NTB by channel, by sequence and over time rather than as a single campaign-level number.
Incrementality
Whether exposure caused the purchase. AMC supports the analysis and holdout structures that make incrementality testing possible on Amazon.

Comparisons

How does this compare to the alternatives?

AMC vs Amazon Attribution

Two different products that get conflated constantly. They answer different questions.

Amazon Marketing CloudAmazon Attribution
PurposeAnalyse event-level Amazon ad signals and build audiencesMeasure how non-Amazon traffic sources drive Amazon sales
Data scopeAmazon DSP and Sponsored Ads signals within your instanceExternal channels — search, social, email, display — tagged to Amazon outcomes
InterfaceSQL queries, templates, guided workflows and AI assistanceReporting dashboard with tagged links
OutputCustom analysis and DSP-activatable audiencesAttribution reporting by external source
When to use whichUse Amazon Attribution when the question is “is my Google or Meta spend driving Amazon sales?” Use AMC when the question is “how do my Amazon channels actually work together?” Brands running both external media and DSP typically need both.

AMC vs standard campaign reporting

Why the console cannot answer the questions that matter most for budget allocation.

AMCCampaign reports
GranularityEvent level — individual impressions and conversionsAggregated by campaign, ad group or keyword
Cross-channel viewSponsored Ads and DSP joined on the same shopper journeyEach channel reported separately
SequencingFull ordered path to purchaseLast-touch attribution only
FrequencyDistribution and effective frequency analysisAverage frequency at best
Audience creationCustom SQL-defined segments activatable into DSPPreset audiences only
When to use whichCampaign reports tell you what a campaign was credited with. AMC tells you what actually happened. If you are making six- or seven-figure allocation decisions on last-touch reporting, you are optimising to an artefact of the attribution model.

AMC vs a third-party clean room or MMM

Where AMC's advantage ends and other measurement approaches begin.

AMCMMM / third-party
ScopeAmazon advertising signals onlyAll channels, including offline and non-digital
GranularityEvent level, user-sequencedAggregate, usually weekly time series
Time to insightDaysWeeks to months
AnswersHow Amazon channels interact and overlapHow total marketing investment drives total business outcomes
When to use whichAMC cannot see outside Amazon, so it cannot answer whole-business allocation questions. Marketing mix modelling can, but at far lower granularity. Mature programmes use AMC for channel-level decisions inside Amazon and MMM for portfolio-level decisions across the business.

Real results

What does this work look like in practice?

Industry · Outdoor Consumer Electronics

Measuring customer journeys beyond last-click attribution

68.3%Of purchases followed a single-touch journey
SP+DSPThe pairing driving most multi-touch conversions

Challenge

Standard campaign reports could not explain how Sponsored Ads and Amazon DSP influenced each other throughout the purchase journey.

Our approach

  • Built customer journey analysis in Amazon Marketing Cloud
  • Evaluated multi-touch conversion paths
  • Analysed channel interactions
  • Reviewed campaign sequencing

Key findings

  • 68.3% of purchases followed a single-touch journey
  • Multi-touch conversions were primarily influenced by Sponsored Products working alongside DSP Display campaigns
  • Journey analysis identified opportunities to rebalance investment across the funnel

Business impact

Budget allocation shifted from last-click optimisation toward customer journey optimisation.

Case studies are presented by industry rather than by client name. Figures are drawn from live account analysis. Engagements marked prior agency engagement were delivered by Ana Perez Ibarz in a previous agency role; the work and results are hers, the client relationships were the agency's. TNOMADS does not identify clients or publish client performance data without written consent.

Frequently asked questions

Common questions, answered directly.

Amazon Marketing Cloud is Amazon Ads' privacy-safe clean room. It stores event-level, pseudonymised signals from your Amazon advertising, which you query using SQL to analyse the customer journey, then use to build custom audiences for Amazon DSP.
A secure environment where detailed event-level data can be analysed without any party gaining access to raw user records. Queries return aggregated results above minimum thresholds, so individual shoppers cannot be identified or re-identified.
Historically AMC was tied to Amazon DSP usage, and eligibility has since expanded. Access is arranged through Amazon Ads, and most brands with meaningful advertising volume can now request an instance. Availability varies by region and account type, so confirm eligibility with your Amazon account team.
The core instance is provided by Amazon rather than sold as a licence. Costs arise in three places: query compute at scale, paid features that Amazon charges separately for, and the analyst time required to write and interpret queries. The third is usually the largest.
Not to start. Amazon provides instructional query templates, guided no-SQL workflows and AI-assisted query generation, and in March 2026 added Ads Agent controls in the query editor for editing, explaining and repairing SQL. [2] To go beyond templates and answer questions specific to your business, SQL still matters.
How many touches precede a purchase and in what order; how much your channels overlap and where you are paying twice; the frequency at which additional impressions stop converting; how new-to-brand behaviour differs by channel; and how much attributed revenue is genuinely incremental.
Amazon expanded the retail purchase lookback to five years for measurement use cases, up from 13 months, which materially improves lifetime value and new-to-brand analysis for seasonal and long-lifecycle products. [1] Ad-traffic lookback is shorter and varies by market, so confirm the current window for your region.
Yes, and it is one of the strongest reasons to use it. You can define a segment in SQL — for example, shoppers who viewed a detail page twice without purchasing in the last 30 days — and activate it directly into DSP line items.
Amazon has extended AMC-built audience segments beyond Sponsored Display into Sponsored Products ad types. Availability continues to change, so verify current support in your account before designing a strategy that depends on it.
Partially. You can analyse Sponsored Ads signals without running DSP, which still answers questions about sequencing, new-to-brand behaviour and search overlap. But audience activation and the full cross-channel picture depend on DSP being in the mix.
Reconstructing the ordered sequence of ad exposures preceding a conversion. It shows whether a channel is genuinely initiating journeys or arriving at the end and receiving last-touch credit for demand created elsewhere.
By showing the distribution of exposures and the conversion rate at each frequency level, AMC identifies the point where additional impressions stop producing incremental conversions. That figure is the evidence base for where the cap should sit.
It supports the analysis, but incrementality still requires experimental design — holdout groups, geo splits or controlled exposure. AMC gives you the event-level data to analyse a properly designed test; it does not substitute for designing one. See incrementality.
Amazon Attribution measures how non-Amazon traffic sources drive Amazon sales. AMC analyses event-level signals from your Amazon advertising. Different questions, different products — brands running external media alongside DSP usually need both.
Simple template queries typically return in minutes. Complex multi-table queries across long lookback windows can take considerably longer, and poorly written queries can be expensive as well as slow. Query design matters for cost as much as for accuracy.
No. AMC returns aggregated results above minimum thresholds and blocks outputs that could identify individuals. This is the core privacy guarantee of the clean room and cannot be worked around.
Additional datasets and capabilities Amazon charges for separately, beyond the standard instance. Availability and pricing change, so confirm current options with your Amazon Ads representative rather than relying on published summaries.
Structural analyses like overlap and effective frequency are worth revisiting quarterly or whenever campaign architecture changes materially. Cohort and new-to-brand tracking works better as a monthly recurring output. Running everything continuously mostly generates cost rather than insight.
It depends on advertising volume, not company size. Below a certain impression and conversion volume, event-level analysis produces samples too small to support confident conclusions, and the analyst time is hard to justify. Brands running DSP at scale get the most value.
It only sees Amazon advertising signals, so it cannot answer whole-business allocation questions. Aggregation thresholds limit small-segment analysis. It shows correlation and sequence, not causation, unless you design an experiment. And it requires interpretation — the query is the easy part.
Someone who understands both SQL and your account structure. Queries written without campaign context routinely produce technically correct but strategically meaningless answers — a number that is accurate and tells you nothing about what to do next.
It is the deep-dive layer for Amazon specifically. Above it sits cross-platform retail media strategy and, for whole-business questions, marketing mix modelling. Below it sit Sponsored Ads and DSP campaign operations.

Why TNOMADS

Why bring AMC work to TNOMADS?

SQLqueries written in-house, not licensed from a dashboard vendor
13+years of retail media context behind every query
Ownthe queries — you keep them if we stop working together
1:1analysis delivered by the person who ran your campaigns

About the author

Who wrote this page?

Ana Perez Ibarz

Senior Retail Media Consultant

Amazon AdsAmazon DSPAmazon Marketing Cloud Walmart ConnectInstacartRetail Media Strategy

Last reviewed 16 August 2026

Ana Perez Ibarz is the founder of TNOMADS Consulting and has spent 13+ years in retail media and performance marketing. She manages live campaign operations across Amazon Ads (Sponsored Products, Sponsored Brands, Sponsored Display, DSP and Amazon Marketing Cloud), Walmart Connect, Instacart, Kroger Precision Marketing, Target Roundel, Loblaws Advance, DoorDash, Criteo, Google Ads and Meta.

On AMC specifically, Ana writes production SQL against live instances — new-to-brand cohort queries with time dimensionality, halo and off-catalogue attribution separation, audience overlap analysis and geo-level DSP queries. Where a query has limitations, this page says so rather than presenting AMC as a solution to every measurement problem.

She works as a subcontracted specialist for agencies as well as directly with brands, and writes on retail media measurement, incrementality and commerce media architecture. Based in Granada, Spain; operating across North America and Europe.

More about Ana and TNOMADS →

Sources

Where do these figures come from?

A note on these sourcesAMC changes faster than almost any Amazon Ads product. Eligibility rules, Paid Features, lookback windows and audience activation support are all revised regularly, and secondary sources frequently disagree with each other — published figures for the ad-traffic lookback window currently vary between 24 and 25 months depending on the source and market. Confirm anything you plan to build a strategy on against your own Amazon Ads account.

Related services

What should you read next?

Amazon DSP

Where AMC-built audiences activate, and the channel AMC most improves.

Incrementality

The experimental design AMC data supports but does not replace.

Amazon PPC

The Sponsored Ads signals that make up half of any AMC overlap analysis.

Retail Media Strategy

How AMC findings translate into cross-platform budget decisions.

Retail Media Audit

Structural review, often the right step before investing in clean room analysis.

Google Ads

External demand that Amazon Attribution, not AMC, is built to measure.

Ready to unlock your AMC?

Move from reporting to real decisions.

Book a Discovery Call