Amazon Attribution, Ads Reporting, or AMC? How to Choose the Right Measurement Setup

Amazon measurement can look more complicated than it actually is.

There is Amazon Ads reporting. There is Amazon Attribution. Amazon Attribution can be implemented manually, in bulk, or through an API. Then there is Amazon Marketing Cloud.

They are related, but they do not solve the same problem.

The easiest way to choose the right setup is to start with one question:

Where did the media run?

That decision separates most use cases immediately.

Start with where the media ran

If the advertising ran through Amazon Ads, you already have native Amazon advertising reporting.

If the advertising ran somewhere else - Meta, Google, email, an influencer campaign, another publisher - and sends shoppers to Amazon, the relevant Amazon measurement product is generally Amazon Attribution.

Amazon itself describes Amazon Attribution as its measurement solution for non-Amazon marketing and lists channels including search, social, display, video, and email.

A simplified decision tree looks like this:

Where did the media run?

Amazon Ads
|
|-- Standard campaign reporting
|      -> Amazon Ads reporting / Unified Reporting
|
'-- Deeper custom analysis, audiences or data joins
       -> Amazon Marketing Cloud


Outside Amazon -> traffic goes to Amazon
|
'-- Amazon Attribution
       |
       |-- Ads Console
       |     |-- Individual campaigns -> Manual tags
       |     '-- Facebook/Instagram or Google Search at scale -> Bulk tags
       |
       '-- API integration
             |-- Publisher supports dynamic macros -> Macro tag
             '-- No dynamic macro support -> Non-macro API tag


Need external media cost + Amazon sales in one analysis?
-> Combine publisher data with Amazon Attribution data

The second question - how should the Attribution tag be created? - only matters once you know that Amazon Attribution is the measurement setup you need.

If the media ran through Amazon Ads

For campaigns bought through Amazon Ads, the starting point is Amazon's own advertising reporting rather than Amazon Attribution.

Amazon made Unified Reporting generally available in June 2026. It allows advertisers to build reports across multiple Amazon Ads products, accounts, countries, metrics, and dimensions in one reporting environment.

This includes Sponsored Ads and Amazon DSP. Amazon's current Unified Reporting documentation also states that the former Sponsored Ads and Amazon DSP reporting pages are scheduled to be shut down by December 31, 2026.

For a brand simply asking:

• How much did we spend?
• What sales were attributed to our Sponsored Products campaigns?
• How did campaigns perform across Amazon ad products?
• Which placements or audiences performed best?

native Amazon Ads reporting is the logical starting point.

You do not need Amazon Marketing Cloud just to retrieve ordinary campaign performance.

When AMC becomes relevant

Amazon Marketing Cloud is a different layer.

Amazon describes AMC as a privacy-safe cloud-based clean room where advertisers can analyze pseudonymized signals, build audiences, combine Amazon Ads signals with advertiser-provided inputs, and run more flexible custom analyses.

Typical questions are therefore more sophisticated than ordinary campaign reporting:

• How do different Amazon Ads touchpoints interact across the customer journey?
• What incremental reach exists between different Amazon Ads products?
• How can first-party transaction data be analyzed alongside Amazon Ads signals?
• Can we create a custom audience based on a specific sequence of advertising interactions?
• How do customer groups behave across a longer journey rather than within one campaign report?

Importantly, AMC is not just a DSP product anymore. Amazon introduced self-service AMC access for Sponsored Ads advertisers in 2025.

That still does not make AMC a replacement for Amazon Attribution.

The two products answer different primary questions.

If the media ran outside Amazon

Suppose you spend money on Meta and send shoppers to an Amazon product detail page.

Meta knows what happened on Meta:

• spend
• impressions
• reach
• clicks
• campaign structure
• ad structure
• other publisher-side delivery metrics

But your actual purchase takes place on Amazon.

This is exactly the use case Amazon Attribution is designed to address.

An Amazon Attribution tag is added to the link leading to Amazon. Amazon can then associate eligible Amazon shopping activity with the tagged click.

According to Amazon's Attribution guide, reporting can include metrics such as clicks, detail page views, add-to-cart activity, purchases, units sold, product sales, and brand-halo activity.

Understand the attribution model before using the numbers

Amazon Attribution uses a 14-day, last-touch attribution model.

Amazon documents this directly in its Attribution methodology:

• A conversion must occur within 14 days of an eligible Attribution click.
• Credit is assigned to the most recent eligible click.

This matters.

If Amazon Attribution reports $20,000 in attributed sales, the statement you can make is that those sales received credit according to Amazon's documented attribution rules.

It is not automatically evidence that all $20,000 would have disappeared without the advertising.

That distinction becomes important when Attribution data is later used for budget optimization or incrementality testing.

Four ways to implement Amazon Attribution

Once you have decided to use Amazon Attribution, there is another decision to make.

You can create the measurement setup through the Ads Console or through the Amazon Attribution API.

Within each approach, Amazon documents two different workflows.

These are not just four different ways to create the same URL. They affect how campaigns are implemented and how easily the resulting data can be connected to publisher data later.

Option 1: Manual Amazon Attribution tags in the Ads Console

This is the most straightforward workflow.

Amazon documents the process in its Amazon Attribution guide and current campaign-creation documentation.

The basic process is:

1. Open Amazon Attribution under Measurement & Reporting in the Amazon Ads Console.
2. Create an Attribution campaign.
3. Select the products you want to measure.
4. Create an ad group for the specific marketing tactic.
5. Select the publisher and channel.
6. Enter the Amazon product page or Store destination.
7. Amazon generates a unique Attribution tag.
8. Place that tag in the corresponding external ad, link, post, or button.

Amazon recommends structuring ad groups based on where and how the individual tag will be used. Each ad group receives its own unique Attribution tag.

This works well when the number of placements is manageable.

For example, a brand running:

• five Meta ads
• one newsletter
• three influencer links

can create a controlled tagging structure without building an API integration.

The trade-off is operational effort.

As the number of campaigns, creatives, or destinations grows, creating and maintaining individual tags becomes increasingly cumbersome.

Option 2: Bulk tags for Facebook, Instagram, and Google Search

Amazon also offers a console-based bulk workflow.

This is an important distinction because you do not necessarily need an API simply because you have many ads.

Amazon currently documents bulk creation for:

• up to 8,500 Facebook or Instagram ads in one file
• up to 100,000 Google Search keywords in one file

The workflow is documented in Amazon's Attribution guide.

For Facebook and Instagram, the process is conceptually:

1. Export campaign/ad information from Meta Ads Manager.
2. Add the required data to Amazon's bulk template.
3. Upload the file to Amazon Attribution.
4. Let Amazon generate the corresponding Attribution tags.
5. Download the processed file.
6. Import the setup back into Meta Ads Manager.

Google Search uses a similar export-template-upload process.

This can be a very practical middle ground for teams that have too much volume for manual tag creation but do not want to maintain an API integration.

There is one important limitation:

Amazon's public documentation currently describes the console bulk workflow specifically for Google Search and Facebook/Instagram.

It should therefore not be assumed that the same bulk workflow exists for every external publisher.

Option 3: API macro tags

The API workflow is where Amazon Attribution becomes significantly more scalable.

Amazon describes the Amazon Ads API as a solution for advertisers, agencies, and technology providers that need programmatic access and have the engineering resources or an integrated tool to support it. Amazon also states that there is no additional Amazon Ads fee for using the API.

For Amazon Attribution, the most interesting implementation is the macro tag.

Instead of generating a separate static tag manually for every ad, the API can return a publisher-specific tag containing dynamic parameters.

Amazon's Attribution API documentation explains that macro-enabled tags can dynamically populate campaign information when the publisher prepares the ad click.

The important fields include:

• aa_campaignid
• aa_adgroupid
• aa_creativeid

Amazon then uses those identifiers to create the corresponding Attribution campaign hierarchy.

Why this makes data stitching easier

This solves one of the most annoying problems in cross-platform reporting: matching entities between two systems.

Imagine you have Meta campaign data containing:

Campaign ID: 12345
Ad set ID: 67890
Ad ID: 24680
Spend: $4,250

Amazon Attribution has the Amazon-side outcome:

aa_campaignid
aa_adgroupid
aa_creativeid
Amazon purchases
Units sold
Product sales

With the macro workflow, identifiers from the publisher are carried into the Amazon Attribution structure.

Amazon itself describes one benefit of the API workflow as "No manual mapping" in its Amazon Attribution guide.

That does not mean there is suddenly only one data source.

You still have Meta data and Amazon data.

What changes is that software can join the records using stable identifiers instead of relying on a person to match campaign names between spreadsheets.

That is a major operational difference.

Do not hard-code a list of macro publishers

Amazon's API provides a publisher resource that indicates whether each currently supported publisher uses macros.

Its technical documentation gives Google Ads, Facebook, Microsoft Ads, and Pinterest as examples of publishers with macro-containing tags.

A production integration should use Amazon's current publisher resource rather than assuming a permanently fixed list of supported publishers.

Option 4: API non-macro tags

A common misunderstanding is that a publisher without dynamic macro support automatically forces you back to manually creating tags in the Ads Console.

Amazon documents another API option:

/attribution/tags/nonMacroTemplateTag

The Amazon Attribution API guide states that the integration can retrieve a non-macro template and supply its own values for:

• aa_campaignid
• aa_adgroupid
• aa_creativeid

The difference is therefore not API versus no API.

It is:

Who supplies the identifiers?

With a macro-compatible publisher, the publisher can insert the dynamic values.

With a non-macro tag, the integrating system supplies the values itself.

That still allows a completely programmatic measurement workflow.

Amazon also documents the use of non-macro tags for eligible links in environments such as emails and blog posts, provided the traffic ultimately lands on a supported Amazon product detail page or Store page.

Reporting depends on how you built the Attribution setup

This is where implementations are often described too casually.

There are clear documented reporting routes for each setup.

Console-created campaigns

For manual and bulk campaigns created through the Ads Console, Amazon documents:

• a visual Attribution reporting dashboard
• downloadable Excel reports
• reports that can be generated on demand
• reports that can be scheduled ahead of time

Amazon describes four main report families in its Attribution reporting guide:

1. Campaign reports
2. Channel/publisher reports
3. Keyword or creative reports for bulk-created campaigns
4. Product reports

For a brand operating a console-based workflow, this provides a fully documented extraction route without requiring an Attribution API integration.

API-created campaigns

For API workflows, Amazon documents two primary report types:

• PERFORMANCE
• PRODUCTS

The PERFORMANCE report can include:

• aa_campaignid
• aa_adgroupid
• aa_creativeid
• publisher
• click-throughs
• promoted conversion metrics
• total conversion metrics

Amazon documents reporting at campaign, ad-group, or creative level and up to 13 months of reporting history.

The PRODUCTS report provides product-related dimensions including ASIN and brand information, but there are important constraints.

According to Amazon's API documentation:

• it is always aggregated at ad-group level
• it returns only the top 100 products per campaign under Amazon's documented ranking logic
• it does not contain click-throughs
• its documented reporting history is 90 days

These details matter when designing a reporting data model.

You should not design a dashboard first and discover later that the underlying API grain does not support what the dashboard assumes.

Do not treat Amazon Attribution data as real-time final data

Another implementation detail becomes important once reporting is automated.

Amazon documents that Attribution data can change after the initial conversion.

For API reporting, Amazon states that:

• clicks can take up to 24 hours to appear
• conversions can take up to 48 hours to appear
• historical reports are re-aggregated after 1, 7, and 28 days
• returns and refunds can affect later values
• Amazon recommends retrieving the previous 30 days every day

There is also a privacy threshold.

Rows representing tags or tag groups with fewer than 10 clicks can be masked and display zero until the threshold is reached.

This means a reporting pipeline should not simply fetch yesterday once, store it permanently, and assume the number is final.

For automated reporting, historical restatement has to be part of the data process.

Why you still need data stitching

Amazon Attribution solves the Amazon side of off-Amazon measurement.

It does not eliminate the fact that two different systems hold different parts of the business question.

Consider a Meta campaign.

Meta holds information such as:

Campaign
Ad set
Ad
Spend
Impressions
Reach
Meta-side clicks

Amazon Attribution can supply:

Attribution campaign/ad hierarchy
Amazon Attribution clicks
Detail page views
Add to carts
Purchases
Units sold
Product sales
Brand-halo outcomes

If the question is:

How much Meta spend generated how much attributed Amazon revenue?

neither dataset alone is sufficient.

This is what data stitching means in this context.

The media platform owns the cost and delivery information. Amazon Attribution owns the Amazon shopping outcome information.

The two sides need to be combined if you want unified economics.

Macro tags reduce the mapping problem, not the number of systems

This distinction is particularly important.

A macro-tag integration does not magically cause Meta spend to become an Amazon Attribution metric.

Instead, it makes the relationship between the two datasets much easier to maintain.

The stable publisher identifiers carried into the Amazon aa_* fields give you a programmatic joining mechanism.

So:

Without a structured identifier strategy:
You may end up manually matching campaign names, ad names, or exported files.

With a properly implemented API macro workflow:
The publisher and Amazon datasets can be joined programmatically using identifiers.

That is what makes the API approach so powerful at scale.

Which setup makes sense for which use case?

There is no documented Amazon rule saying that a company of a certain size must use one particular setup.

A better way to choose is based on operational complexity.

For many small and mid-sized Amazon brands, this means there is no reason to jump directly to the most technically sophisticated architecture.

If you can reliably maintain twenty external links, an API integration may add unnecessary complexity.

If you manage thousands of ads across accounts and need continuously updated cross-platform reporting, manually maintaining individual Attribution tags becomes a very different proposition.

The correct question is therefore not:

Which Amazon measurement product is the most advanced?

It is:

What is the simplest setup that gives us the measurement granularity, reliability, and automation our decision process actually requires?

Where Amazon Marketing Cloud fits

AMC becomes much easier to understand once Amazon Attribution has been separated from it.

Amazon Attribution is designed to measure non-Amazon marketing that sends customers to Amazon.

AMC is a privacy-safe analysis environment for deeper custom analysis using pseudonymized signals.

Amazon says AMC can combine signals from:

• Amazon Ads
• advertiser-provided inputs
• onboarded third-party providers

and only exposes aggregated, anonymous outputs.

Amazon also documents first-party signal ingestion into AMC, allowing advertisers to analyze their own pseudonymized information alongside Amazon Ads signals.

That makes AMC valuable for questions that go beyond a conventional campaign report.

But if your immediate problem is simply:

"Which Meta campaign drove purchases on Amazon?"

Amazon Attribution is the Amazon product explicitly designed for that measurement problem.

AMC should not be introduced simply because it is technically more sophisticated.

Attribution is not the same as incrementality

There is one final distinction that matters once the data starts influencing budget decisions.

Amazon Attribution assigns conversion credit according to Amazon's documented 14-day last-touch methodology.

That is attribution.

Incrementality asks a different question:

What would have happened if the advertising had not run?

Those are not interchangeable questions.

For example, assume a Meta campaign receives credit for $100,000 of Amazon sales.

The Attribution data tells you which eligible campaign clicks received credit for those purchases.

It does not by itself tell you how many of those customers would have bought anyway.

That is why attribution data can be extremely useful for:

• operational reporting
• campaign comparisons
• creative optimization
• identifying strong traffic sources
• building test hypotheses

while causal budget decisions may require an additional experimentation or incrementality framework.

A mature measurement setup therefore does not replace Attribution with testing.

It uses each method for the question it can actually answer.

Where Azulens fits

Our view at Azulens is that getting access to Amazon Attribution data is only the first step.

The real value comes from turning that data into a repeatable decision process.

A brand may already have:

• Meta campaign cost and delivery data
• Google campaign data
• Amazon Attribution outcomes
• Amazon Ads campaign reporting
• additional sales or business signals

but these datasets live in different systems, use different structures, and answer different parts of the performance question.

This is where we focus.

Azulens can bring the available publisher and Amazon measurement data into a structure that can be used for:

• unified cross-platform reporting
• optimization analysis
• campaign and creative comparisons
• decision-support workflows
• pre-test analysis
• incrementality test design
• post-test evaluation

We do not see Amazon Attribution as a replacement for incrementality testing.

We see it as one important input into a broader measurement and optimization system.

For smaller brands, that can mean working with scheduled Amazon Attribution reports rather than introducing unnecessary API infrastructure.

For more automated setups, it can mean using API-based Attribution data and publisher identifiers to create a much more scalable reporting process.

The architecture should follow the actual use case.

The simplest way to think about the Amazon measurement stack

If you remember only four things, remember these:

1. Start with where the media ran. Amazon Ads media starts with Amazon Ads reporting. External media driving to Amazon starts with Amazon Attribution.
2. Do not confuse Amazon Attribution with one specific implementation method. It can be implemented manually, in bulk, or through macro and non-macro API workflows.
3. External-media economics still require data stitching. Your publisher has the media cost. Amazon Attribution has the Amazon outcome. Unified ROAS requires both.
4. Do not confuse attribution with causality. Attribution tells you which touchpoint receives credit under a defined model. Incrementality requires answering what happened because the advertising ran.

Once these distinctions are clear, Amazon's measurement stack becomes much easier to design around.

And more importantly, you can choose the measurement infrastructure based on the decisions your team actually needs to make - instead of simply implementing the most complex tool available.

Primary sources

• Amazon Attribution product overview
• Amazon's complete guide to Amazon Attribution
• Create an Amazon Attribution campaign
• Apply an Amazon Attribution tag
• Amazon Attribution API technical guide
• Amazon Attribution methodology
• Amazon Attribution troubleshooting
• Amazon Ads API overview
• Amazon Unified Reporting announcement
• Amazon Unified Reporting documentation
• Amazon Marketing Cloud
• AMC self-service access for Sponsored Ads advertisers
• First-party signals in Amazon Marketing Cloud