Amazon Attribution is a free measurement tool that shows how non-Amazon marketing, such as search ads, social ads, email, and influencer links, drives clicks, detail page views, add to carts, and sales on Amazon. It works by appending Amazon-issued tracking tags to the destination URLs you already use.
Key Takeaways
- Amazon Attribution measures off-Amazon channels driving on-Amazon conversions, closing the blind spot between ad platforms and marketplace sales.
- Access requires a Seller Central or Vendor Central account with a registered brand.
- Reporting covers clicks, detail page views, add to carts, purchases, and sales, by channel, campaign, and ASIN.
- Tags are per creative or per link, so plan a naming convention before you generate hundreds of them.
- Attribution is click-based with a fixed lookback, so it undercounts view-through and cross-device journeys.
What Problem Does Amazon Attribution Solve?
When you send paid traffic to a marketplace listing, the ad platform sees the click and then loses the trail. Amazon sees the sale but not the source. Marketers respond by guessing, usually by correlating spend spikes with sales spikes, which breaks down as soon as more than one channel is active. Amazon Attribution replaces the guess with per-link reporting for the on-Amazon funnel.
That funnel visibility is the real value. Knowing that a campaign produced clicks but few detail page views points at a link or landing mismatch. Detail page views with few add to carts point at listing content, images, reviews, or price. Add to carts without purchases point at checkout friction such as shipping cost or availability. The tool turns an opaque channel into a diagnosable one.
Who Is Eligible and How Do You Get Access?
Eligibility centers on brand ownership. Sellers enrolled in Brand Registry can access Attribution through the advertising console, and vendors can access it through their existing account. Agencies typically request access on behalf of the brand rather than under their own account, so reporting stays with the brand if the relationship changes.
Availability varies by marketplace, so confirm support in each country you sell in before designing a global tagging scheme. If a marketplace is unsupported, you will need a fallback such as coupon codes, dedicated landing pages, or storefront-level analysis for that region.
How Do You Set Up Tags Correctly?
| Setup element | What to do | Why it matters |
|---|---|---|
| Advertiser | Create one advertiser per brand | Keeps reporting clean when managing several brands |
| Campaign | Mirror the campaign names used in the ad platform | Lets you join Amazon results to ad spend |
| Publisher | Select the closest match, or use custom for channels not listed | Drives channel-level rollups in reporting |
| Creative or link | One tag per creative, placement, or newsletter slot | Without granularity you cannot tell winners from losers |
| Destination | Product detail page or a store page, tagged consistently | Store pages often convert better for multi-ASIN campaigns |
| Naming | Adopt a fixed convention such as channel-campaign-audience-creative | Tags cannot be renamed retroactively without losing continuity |
What Is the Implementation Order?
- Confirm access through Brand Registry or your vendor account, and decide who owns the advertiser entity.
- Map the channels you intend to measure, and list every distinct creative or placement that deserves its own tag.
- Define the naming convention and record it somewhere shared before generating a single tag.
- Generate tags and replace the destination URLs in each channel with the tagged versions. Bulk upload templates save hours once you exceed a few dozen links.
- Test one link per channel by clicking it and confirming it lands on the correct listing, then confirm the click appears in reporting.
- Set a reporting cadence that pulls Amazon Attribution metrics alongside ad platform spend so cost per on-Amazon sale is computed in one place.
Testing before scale is the step teams skip. A single malformed tag pushed across an entire paid social account produces a week of untracked spend, and the data cannot be recovered retroactively.
How Do You Read the Reports Without Fooling Yourself?
Amazon Attribution reports click-based conversions within a fixed lookback window. That means it credits journeys where the shopper clicked your tagged link and then bought, and it does not credit journeys where the shopper saw an ad, searched Amazon later on a different device, and bought without ever clicking. So reported sales are a floor for the channel's contribution rather than a complete picture.
Two practical habits keep the numbers honest. Compare tagged campaign performance against a holdout or against periods when the channel was paused, so you can sanity-check whether total marketplace sales move in the direction the tool implies. And separate branded from non-branded traffic, because branded clicks often would have converted anyway, which inflates apparent efficiency.
How Does It Compare to Other Amazon Measurement Options?
Amazon Attribution is the right tool for off-Amazon channels you buy yourself. Brand Referral Bonus builds on the same tagging to return a percentage of qualifying external sales as a credit, so tagging your external traffic can lower effective fees, not just improve reporting. Amazon Marketing Cloud serves a different need: a privacy-safe clean room for advanced cross-channel and audience analysis, aimed at larger advertisers with analyst resources.
Outside Amazon's own tooling, teams use unique landing pages, one-time coupon codes, and post-purchase surveys as triangulation methods. None of these are as precise as tagged links, but they help estimate the view-through contribution that click-based attribution structurally misses.
What Are the Most Common Mistakes?
The biggest is tagging at too coarse a level. One tag for all of paid social tells you the channel works but never which audience or creative works, which is the decision you actually need to make. The second is inconsistent naming across marketplaces and channels, which makes joined reporting impossible without manual cleanup every month.
Others are quieter. Sending traffic to an out-of-stock ASIN wastes spend and depresses conversion metrics for the entire campaign. Failing to tag influencer and affiliate links leaves the largest external channel invisible. And judging a campaign on purchases alone, rather than on the detail page view to add to cart step, hides listing problems that are cheaper to fix than ad targeting.
How Do You Use the Data to Change Bids and Budgets?
The output only matters if it moves money. Build a simple weekly view that joins ad platform spend to Amazon Attribution sales at the creative level, then compute cost per on-Amazon purchase for each row. Rank the rows, cut the bottom decile, and shift that budget to the top performers. Because Attribution undercounts, set your acceptable cost per purchase threshold slightly looser than you would for a directly measured on-site conversion.
Watch the funnel ratios as well as the cost. A creative with a strong click to detail page view rate but a weak detail page view to add to cart rate is not a bad ad, it is a good ad pointed at a weak listing, and pausing it destroys demand you already paid to create. Fix the listing first, then re-evaluate.
How Should External Traffic Fit the Wider Channel Mix?
External traffic to Amazon competes with your own direct-to-consumer store, so decide deliberately where each campaign should land. Marketplace destinations usually convert better for low-consideration, review-dependent products and for shoppers already inside Amazon's trust and shipping expectations. Your own site is better where you need email capture, subscription enrollment, bundle configuration, or first-party data for retargeting.
A common resolution is to split by funnel stage: prospecting and social discovery campaigns point to Amazon for frictionless conversion, while retargeting and email point to the owned store where margin is higher and data is retained. Tagging both paths lets you compare them with real numbers instead of internal preference.
Frequently Asked Questions
Does Amazon Attribution Cost Anything?
No. Amazon Attribution is offered at no cost to eligible brand-registered sellers and vendors. The cost is operational rather than financial: someone has to design a tagging convention, generate and deploy the tags across every channel, and maintain them as campaigns change. Budget setup time and a recurring reporting cadence instead of a license fee.
What Metrics Does Amazon Attribution Report?
Reporting covers clicks on your tagged links plus downstream on-Amazon actions: detail page views, add to carts, purchases, units ordered, and sales revenue, broken out by campaign, publisher, creative, and ASIN. Reading these as a funnel is the point, because the step where the drop-off occurs tells you whether to fix the ad, the link, or the listing itself.
Can Agencies Manage Amazon Attribution for a Client?
Yes. The usual arrangement is that the brand owns the advertiser entity and grants the agency access, so historical measurement data stays with the brand if the agency changes. Agencies should still standardize naming conventions across clients, since consistent tag structures are what make automated reporting and cross-client benchmarking possible.
Does Amazon Attribution Track View-Through Conversions?
Attribution is click-based within a fixed lookback window, so a shopper who sees an ad and later buys on Amazon without clicking the tagged link will not be credited. Expect reported sales to understate the true channel contribution, and use holdout tests, paused-period comparisons, or survey-based methods to estimate the uncredited portion.
Should Tagged Links Point to a Product Page or a Brand Store?
Both work, and the right choice depends on the campaign. Single-product ads should land on the product detail page to keep the path short. Campaigns promoting a range, or prospecting traffic that is still browsing, often perform better on a brand store page, where multiple products and brand content reduce bounce. Test both with separate tags rather than assuming.