What is Attribution?
Attribution connects your ad activity to real purchases, showing you which campaigns, ad formats, and products are driving sales.
Criteo Commerce Max measures sales attribution across your onsite and offsite campaigns. It identifies which shopper interactions — clicks or ad views — led to a purchase, and assigns credit to the most relevant ad event. Attribution results are reflected in attributed sales and return on ad spend (ROAS) in your campaign reports.
Why It Matters
Measure performance accurately. See which ads are driving real sales, not just clicks or impressions.
Optimize your media strategy. Understand what's working and adjust your bidding, creatives, or placements accordingly.
Understand shopper behavior. Learn how and when shoppers engage with your ads before purchasing.
Update retroactively. Change your attribution model at any time — your historical reporting data refreshes automatically to reflect the new settings.
When to Configure Attribution Settings
Review or update your attribution settings when:
Launching new campaigns and defining your performance KPIs.
Testing different ad formats, such as Sponsored Products, onsite display, or offsite campaigns.
Comparing short vs. long purchase cycles (e.g., groceries vs. electronics).
Tracking campaign impact across both online and in-store sales.
User Matching Rules
Commerce Max uses four types of identifiers to connect a purchase to a shopper who interacted with your ad:
Identifier | Description |
|---|---|
Customer ID | The retailer's CRM ID. High reliability; available for authenticated users only. |
Hashed email | A hashed version of the shopper's email address. High reliability; available for authenticated users only. |
Retailer visitor ID | A first-party cookie set by the retailer. Persistent but device-specific. |
Criteo cookie | A third-party cookie managed by Criteo. Less persistent; cross-device. |
When a sale matches on more than one identifier type, Commerce Max credits the most reliable match.
All of our user matching rules are purely deterministic. We never use IP addresses or probabilistic models to determine attribution at the user level.
Product Matching
The attribution model works differently for Demand Brand and Demand Seller accounts. It considers products as attributable if:
Demand Brand Accounts
Match type | Description | Example |
|---|---|---|
Same product | The same product as the one promoted in the ad is purchased by a shopper who was reached. | If a shoe brand called GreatBalance is promoting their product Great Basic Shoes, same product matching will record attribution for any of the different options (quantity, sizing, color, etc.). |
Same category | One of the products that is within the same category and brand as the promoted one is purchased by a shopper you've reached. | If GreatBalance is promoting their Great Basic Shoes but the ad viewer ends up purchasing the Great Hiking Shoes, same category would still record this attribution. |
Same brand | Any product within the same brand as the promoted one is purchased by a shopper you've reached. | Same brand would record any product the audience purchases from GreatBalance, even if it does not relate to the featured Great Basic Shoes — for example, a Great Basic Shirt or Great Basic Running Socks. |
Demand Seller Accounts
Match type | Description | Example |
|---|---|---|
Same product | The advertised product and the purchased product are exactly the same, and both are sold by the same advertising seller. | An ad shows a specific Nike running shoe (e.g., Nike Air Zoom Pegasus 39). The customer clicks the ad and purchases the exact same Nike Air Zoom Pegasus 39 from the same seller. |
Same category | The advertised and purchased products are different but belong to the same product category, and both are sold by the same advertising seller. | An ad promotes Nike running shoes, but the customer purchases a Nike running jacket instead. Both items belong to the "running gear" category and are sold by the same advertising seller. |
Same seller | The advertised and purchased products are different and belong to different categories, but both are sold by the same advertising seller. | An ad promotes Nike running shoes and the customer purchases a Nike fitness tracker. The products are different and belong to different categories (shoes vs. electronics), but both are sold by the same seller, Nike. |
Lookback Windows
A lookback window defines how many days after an ad interaction a purchase can still be attributed to your campaign. You set separate windows for post-click and post-view attribution.
Post-click options: 7, 14, or 30 days
Post-view options: None, 1, 7, 14, or 30 days
Lookback windows are set at the campaign level and apply equally to all line items within that campaign. You can change your settings at any time — all reporting data updates retroactively, typically within two hours.
Choosing the right window depends on your product category. Fast-moving goods like groceries suit shorter windows; high-consideration products like electronics may benefit from longer ones. If you're setting up attribution for the first time, start with a window you've used before so you can compare results on a consistent basis.
Sales Deduplication
A shopper's journey to purchase often involves multiple ad interactions — a sponsored product view, a display click, and so on. Commerce Max attributes each sale to a single, most-relevant ad interaction, not to every campaign or line item that touched the shopper.
When multiple ad interactions compete for the same sale, Commerce Max applies the following priority rules, in order:
Clicks take precedence over impressions.
More specific product matches take precedence — same product beats same category, which beats same brand or same seller.
If two interactions are otherwise equal, the one closest to the purchase date takes precedence.
This logic applies across all your campaigns, line items, and ad formats — including onsite, offsite, sponsored products, and display. Only one event receives attribution for any given sale.
Other interactions that contributed to the purchase but did not receive direct attribution are recorded as assisted sales. For more information, see campaign metrics.