Amazon Advertising Attribution Explained: Windows, Models, and Why Your Reports May Be Wrong

Amazon Advertising Attribution Explained

Attribution is the process of connecting an advertisement click to a resulting purchase. It sounds simple, but Amazon's attribution model is nuanced, frequently misunderstood, and can lead sellers to make bad decisions when they do not understand how it works.

This guide provides a technical explainer on Amazon's attribution windows, how conversions are counted, why your reports change after the fact, and how to make better decisions once you understand the system.

What Attribution Means in Amazon Advertising

When a shopper clicks your Sponsored Products ad and then buys your product, Amazon "attributes" that sale to the ad click. The sale shows up in your advertising reports as an attributed sale, and it factors into your ACoS, ROAS, and other performance metrics.

Attribution Windows by Ad Type

Amazon uses different attribution windows for different ad formats:

Sponsored Products: 7-day click attribution

A sale is attributed to a Sponsored Products ad if the purchase occurs within 7 days of the ad click. If the shopper clicks your ad on Monday and buys on the following Sunday, that sale is attributed. If they buy on the following Monday (8 days later), it is not attributed.

Sponsored Brands: 14-day click attribution

Sponsored Brands ads have a longer 14-day window. This makes sense because Sponsored Brands are often brand awareness plays — the shopper may click on a Sponsored Brands ad, discover your brand, browse your store, and come back to purchase days later.

Sponsored Display: 14-day click attribution, 14-day view attribution

Sponsored Display has both a click attribution window (14 days) and a view-through attribution window (14 days). View-through attribution means that if a shopper sees (but does not click) your Sponsored Display ad and then purchases your product within 14 days, that sale is still attributed to the ad.

View-through attribution is controversial because the causal link between seeing an ad and making a purchase is much weaker than clicking an ad and making a purchase. The shopper may have bought your product regardless of seeing the ad.

Amazon DSP: Customizable windows

Demand-Side Platform campaigns offer configurable attribution windows. Advertisers can set click-through attribution from 1 to 14 days and view-through attribution from 1 to 14 days. This flexibility is useful for testing how different windows affect reported performance.

How Attribution Affects Your Reports

Understanding attribution windows is essential because they directly affect the numbers in your advertising reports — and those numbers change over time.

#### The Data Maturation Problem

When you pull an advertising report today for campaigns that ran last week, the data is incomplete. Here is why:

On Monday, your Sponsored Products campaign received 100 clicks. By Tuesday, 60 of those clicks have resulted in purchases. By Friday, 70 have. By the following Monday (7 days later, when the attribution window closes), 75 have.

If you pull a report on Tuesday, you see 60 sales. If you pull the same report the following Monday, you see 75 sales. Same campaign, same day of performance, but 25% more attributed sales once the data fully matures.

This means:

Reports pulled before the attribution window closes are understating performance. Your campaigns look worse than they actually are if you check too early.

ACoS improves as data matures. ACoS = spend / sales. If sales increase while spend stays the same (because spend is recorded immediately but sales take days to attribute), ACoS decreases as the data matures.

Comparing different time periods requires matured data. Comparing last week's campaigns (fully matured) with this week's campaigns (still maturing) makes this week look worse, creating a false impression of declining performance.

#### Last-Touch Attribution

Amazon uses a last-touch attribution model. If a shopper clicks on two different ads for the same product, the sale is attributed to the last ad clicked.

Example: A shopper clicks your Sponsored Brands ad on Monday. On Wednesday, they click your Sponsored Products ad. On Friday, they buy. The sale is attributed to the Sponsored Products ad (the last click), not the Sponsored Brands ad.

This has practical implications:

Sponsored Brands performance looks worse than it is. Sponsored Brands often serve as an introduction to your brand. The shopper discovers you through the Sponsored Brands ad, then returns later via a Sponsored Products ad to purchase. The Sponsored Products ad gets the credit.

Upper-funnel ads are systematically undervalued. Any ad that introduces your brand but does not result in an immediate purchase is undervalued in a last-touch model.

The ad that gets the sale is not necessarily the ad that drove the sale. This is a fundamental limitation of last-touch attribution that advertisers in every channel (not just Amazon) struggle with.

Same-Product vs. Cross-Product Attribution

Amazon attributes sales at the product level, but the attribution rules for what counts differ:

Sponsored Products: Attributes sales of the specific advertised ASIN. If a shopper clicks an ad for Product A but buys Product B (even within the same brand), the sale is not attributed to the Sponsored Products ad for Product A.

However, if the shopper clicks the ad for Product A and then adds both Product A and Product C to their cart and checks out, only Product A's sale is attributed.

Sponsored Brands: Attributes sales of any product in the brand's catalog that the shopper purchases within the attribution window after clicking the Sponsored Brands ad. This broader attribution makes Sponsored Brands appear to drive more sales because cross-selling within your brand counts.

Sponsored Display: Similar to Sponsored Brands — can attribute sales of the advertised product or related products depending on targeting type.

This difference means you cannot directly compare Sponsored Products ACoS with Sponsored Brands ACoS as apples-to-apples metrics. They are measured differently.

New-to-Brand (NTB) Metrics

For brand-registered sellers, Amazon provides New-to-Brand metrics. These identify whether the customer who made a purchase has bought from your brand before (within the last 12 months) or is a first-time brand customer.

NTB metrics include:

New-to-brand orders: Orders from customers who have not purchased from your brand in the last 12 months
New-to-brand sales: Revenue from those orders
New-to-brand order percentage: What fraction of attributed orders are from new customers

Why NTB matters:

A Sponsored Products campaign with 30% ACoS that is 80% new-to-brand is fundamentally different from a campaign with 30% ACoS that is 10% new-to-brand. The first is acquiring new customers. The second is mostly recapturing existing customers who might have purchased anyway.

NTB is particularly important for evaluating Sponsored Brands campaigns. If your Sponsored Brands ads primarily drive repeat purchases from existing customers, the brand awareness value is limited. If they are 60%+ new-to-brand, they are genuinely expanding your customer base.

The Attribution Window Illusion

One of the most misunderstood aspects of Amazon attribution is what happens at the boundary of the attribution window.

Consider this scenario:

A shopper clicks your Sponsored Products ad on Day 1. They spend a week researching, reading reviews, comparing alternatives. On Day 8, they decide to buy your product and purchase it organically (they search for your brand name and click the organic listing).

In Amazon's reporting:

The Sponsored Products ad gets zero credit for this sale (outside the 7-day window)
The organic sale is recorded as a pure organic sale

But in reality, the ad played a crucial role in the eventual purchase. The shopper discovered your product through the ad, and without that initial click, they may never have purchased.

This means:

Your advertising is likely more effective than your reports suggest. Sales driven by ads but occurring outside the attribution window are invisible to Amazon's reporting.

TACoS is a better metric than ACoS for this reason. TACoS (ad spend / total revenue) captures the organic lift that advertising creates, even if that lift falls outside the attribution window.

Cutting ad spend based purely on ACoS can be destructive. If you pause campaigns because ACoS looks too high, you may lose the organic sales those campaigns were indirectly supporting.

How to Account for Attribution in Your Decisions

#### Rule 1: Wait for Data to Mature

Do not make bid changes or campaign decisions based on data less than 7 days old for Sponsored Products or 14 days old for Sponsored Brands. The numbers will change as conversions continue to attribute.

Practical approach: When reviewing campaign performance, always look at data that ended at least 7 days ago. Instead of reviewing "last week" on Monday, review "two weeks ago" which has had a full attribution window to mature.

#### Rule 2: Use Rolling Averages, Not Snapshots

A single day's performance can vary wildly based on when conversions happen to attribute. Use 7-day or 14-day rolling averages for decision-making. This smooths out the attribution timing noise.

#### Rule 3: Evaluate Sponsored Brands and Sponsored Products Separately

Because they have different attribution windows and different product attribution rules, comparing their ACoS directly is misleading. Evaluate each ad type against its own benchmarks.

#### Rule 4: Factor in NTB for Upper-Funnel Campaigns

If a Sponsored Brands campaign has a 40% ACoS but 70% new-to-brand rate, it is acquiring customers at scale. Those customers have future purchase potential that the ACoS number does not capture.

A simple way to think about it: if the lifetime value of a new customer is 3x the initial purchase, and 70% of your attributed sales are from new customers, the effective ACoS is much lower than reported because you are buying future revenue.

#### Rule 5: Track TACoS Alongside ACoS

TACoS = total ad spend / total revenue (organic + advertising attributed).

TACoS captures the full impact of your advertising, including the organic lift that does not show up in ACoS. If your ACoS stays flat but TACoS is declining, your advertising is becoming more efficient because it is driving more organic sales.

#### Rule 6: Be Skeptical of View-Through Conversions

Sponsored Display view-through attributed sales should be taken with a grain of salt. A shopper who saw your display ad and then purchased your product may have purchased regardless. If you are evaluating Sponsored Display performance, focus on click-through conversions and treat view-through as a bonus.

Amazon Marketing Cloud (AMC) and Advanced Attribution

For brands spending significant budgets, Amazon Marketing Cloud provides more sophisticated attribution analysis:

Multi-touch attribution: Instead of last-touch, AMC can show you every touchpoint a customer had with your ads before converting. This reveals the true contribution of upper-funnel ads.

Path-to-purchase analysis: See common customer journeys — how many touchpoints, which ad types in which order, and how long the path takes.

Cross-channel analysis: Understand how DSP display ads interact with Sponsored Products and Sponsored Brands in the conversion process.

Custom attribution windows: Test different window lengths to see how attributed sales change.

AMC requires DSP access and is typically used by larger brands or agencies. But the insights it provides can fundamentally change how you allocate budget across ad types.

Common Attribution Mistakes

Mistake 1: Judging today's campaigns by today's data. The data is immature. Wait for the attribution window to close before making decisions.

Mistake 2: Comparing ad types by ACoS alone. Sponsored Products (7-day, same-ASIN) and Sponsored Brands (14-day, cross-ASIN) are measured differently. Comparing them directly is misleading.

Mistake 3: Ignoring view-through attribution when it matters. While view-through should be treated with skepticism, completely ignoring it understates the impact of display advertising, which is inherently more passive.

Mistake 4: Cutting campaigns that drive organic sales. If pausing a campaign causes organic sales to decline two weeks later, the campaign was providing more value than its ACoS suggested.

Mistake 5: Not accounting for NTB. A campaign that looks expensive on ACoS but drives 80% new customers is an investment in future revenue.

Building an Attribution-Aware Strategy

Use ACoS for tactical bid management on Sponsored Products. It is the right metric for keyword-level optimization.

Use TACoS for strategic campaign evaluation. It captures the full impact of your advertising program.

Use NTB for budget allocation. Invest more in campaigns with high NTB rates, even if their ACoS is above target. They are growing your customer base.

Review matured data only. Build a 7-day reporting lag into your process to ensure data accuracy.

Test incrementality. Periodically pause specific campaigns and measure the impact on total sales (not just attributed sales). This reveals the true incremental value of each campaign.

Attribution is not just an academic exercise. It determines how you interpret your advertising performance, where you allocate budget, and which campaigns you expand or cut. Getting it right means making better decisions with every dollar you spend on Amazon advertising.

Amazon Advertising Attribution Explained: Windows, Models, and Why Your Reports May Be Wrong