On this page
In the context of AI-led shopping, AI commerce attribution means assigning credit for a purchase or other commercial action to an earlier AI interaction using observed evidence and defined rules.
For example, a retailer may connect a visit from an AI assistant with a later purchase. This differs from AI-powered attribution, where machine learning helps calculate credit across marketing channels. Attribution assigns credit under a model; it does not by itself prove that an interaction caused an additional sale.
Choose what to measure
Measure AI-sourced visits
Use website analytics and campaign evidence to investigate visits from AI tools. Missing referrer data limits what can be inferred.
Review campaign URL measurementTrace an AON Offer outcome
Connect an AI app’s offer referral with a reported purchase and the resulting commission record.
See the AON tracking pathAttribution to AI vs attribution using AI
The difference is whether AI is the source being measured or the method doing the measuring. Awin’s discovery announcement describes measuring AI visibility alongside conversion signals. Google Analytics’ attribution documentation explains how machine learning can assign credit across a recorded journey.
| Meaning | Question it answers | Example |
|---|---|---|
| Attribution to AI interactions | Which purchases received credit for earlier AI-assisted discovery or referrals? | A retailer links an AI assistant referral to a later order. |
| AI-powered attribution | How should a model distribute credit among observed marketing interactions? | A data-driven model assigns credit among search, email, and other measured touchpoints. |
AI referral attribution example
A shopper follows an AI assistant's headphone recommendation to a retailer's website. Assume the retailer's analytics records both the referral source and the later purchase in the same observed customer journey. The retailer can report a purchase associated with that AI referral under its chosen attribution model.
If the link is also part of an affiliate program, a separate tracking system evaluates the referral for commission. The retailer's analytics report and the affiliate report can credit the same order differently.
First-touch vs last-touch attribution
In this hypothetical journey, a shopper clicks an AI assistant's headphone recommendation on Monday, returns through a search ad on Wednesday, and clicks a retailer email before buying on Friday. Assume the measurement system observed all three visits and includes them in the same attribution window.
The purchase is the same; the chosen credit rule changes the result. The table illustrates general first-touch and last-touch rules, not AON's actual attribution rules or an affiliate program's commission policy.
| Illustrative rule | Who receives purchase credit? | What the rule emphasizes |
|---|---|---|
| First-touch | The AI assistant referral receives all credit as the first observed interaction. | The interaction that began this recorded journey. |
| Last-touch | The retailer email receives all credit as the final observed interaction before purchase. | The interaction closest to the recorded purchase. |
How AI referral attribution works
Google Analytics' attribution documentation explains credit assignment and its currently supported models. The first-touch example above illustrates a general rule; GA4 no longer offers that model. For a tracked commercial referral, the evidence path follows these steps.
Recommendation or discovery
The user sees a commercial option. A mention or citation alone measures exposure, not a referral visit or a sale.
Tracked referral
The user follows the supported link or action. The integration keeps the information needed to connect this referral to a later purchase.
Reported purchase or other action
The merchant or authorized provider reports the purchase, signup, booking, or other defined event with its referral information.
Matching and validation
The system matches the event to the earlier interaction and checks timing and other rules. For an affiliate referral, it also checks whether the purchase qualifies for commission.
Website analytics vs affiliate attribution
Website analytics assigns purchase credit using the interactions it observed and the selected model. An affiliate system applies the program's referral identifiers, attribution window, eligible products, and commission rules. Differences in evidence or rules can produce different results for the same order.
| Dimension | Website analytics | Affiliate attribution |
|---|---|---|
| Purpose | Assign purchase credit to marketing interactions. | Identify referrals that qualify for commission. |
| Evidence | Recorded visits, campaign data, and events. | Referral identifiers and reported purchases. |
| Rules | Selected model and lookback window. | Program attribution and commission terms. |
| Result | Purchase value credited to a source. | Commission earned by the publisher. |
Key AI commerce metrics
Use distinct measures for discovery, visits, and commerce. Commission and payment are additional measures when an affiliate relationship exists.
AI visibility
Mentions and citations in a defined set of AI answers measure appearance. Record the query set and testing period.
Referral traffic
Count identified AI-source visits and their landing pages. Calculate a purchase rate only from consistently measured visits and purchases.
Attributed conversion
Report purchase credit and value under a named model and lookback window, so another report can be compared meaningfully.
Validated commission
For affiliate referrals, reconcile qualifying outcomes and net commission with the provider's records, then track received payments.
What AON attributes
AON tracks referrals from AI apps to commercial offers. When the business supplying an offer reports a purchase or another agreed action, AON matches it to the earlier referral and checks the commission rules.
This answers “Which app referred this purchase?” Brand mentions in AI answers are a separate visibility measure; they do not show whether a shopper followed a referral or bought something.
Read AON Tracking APICommon questions about AI commerce attribution
Can AI influence be measured when a user does not click?
It can be investigated through methods such as surveys, aggregate analysis, or experiments. Those methods have different limits from a direct tracked referral and should not be reported as the same metric.
Why can an AI referral appear as direct traffic?
An analytics tool may classify a visit as direct or unknown when it receives no usable referrer or campaign information. A copied URL or a link opened without a referrer can lose the original source. Direct traffic alone cannot tell you whether an AI interaction influenced the visit.
How is attributed revenue different from affiliate commission?
Attributed revenue is the purchase value credited to a source under a reporting model. Affiliate commission is the amount the publisher earns under the program’s terms. A USD 100 purchase might contribute USD 100 of attributed revenue and USD 5 of commission at a 5% rate.
Sources
- Google Analytics: attribution models and credit assignmentOfficial documentation
- Google Analytics: campaign URL parametersOfficial documentation
- Awin: AI discovery and conversion measurementIndustry source
- AON Tracking APIAON documentation