Modelled (All) metrics in Analytics
Modelled metrics are enhanced performance measurements available in Commerce Growth's Analytics module that provide a more complete view of your advertising performance by estimating the behavior of users who cannot be directly tracked.
These metrics combine observed data with statistical modelling to fill measurement gaps caused by users who reject cookies for advertising or have privacy settings that prevent tracking.
Currently available modelled metrics include:
Visits All: Total estimated visits including both directly observed visits and modelled visits from non-trackable users.
Cost Per Visit All (CPV): Cost efficiency calculated using total estimated visits (both observed and modelled).
Sales All: Total estimated sales including both directly observed conversions and modelled sales from non-trackable users.
When to use it?
Use modelled metrics when you need to:
Understand the complete impact of your advertising campaigns beyond directly trackable users.
Make informed decisions about campaign optimization and budget allocation in privacy-conscious environments.
Measure performance accurately in regions with high cookie rejection rates or strict privacy regulations.
Compare campaign effectiveness across different privacy compliance scenarios.
Demonstrate true advertising ROI to stakeholders who need complete performance visibility.
Plan and optimize campaigns for audiences that include privacy-conscious users.
Assess the real cost efficiency of your advertising spend across all user segments.
Future-proof your measurement strategy as privacy regulations continue to evolve.
Why to use it?
Understanding how Criteo estimates performance from users who do not consent to be tracked for advertising is crucial for modern marketers. Modelled metrics provide several key advantages:
Complete measurement picture: Address the growing measurement gaps caused by privacy settings and cookie rejection
Accurate performance assessment: Get closer to your true advertising impact rather than underestimating campaign effectiveness
Informed decision-making: Make optimization decisions based on comprehensive data that includes all user segments
Privacy-compliant insights: Maintain measurement capabilities while fully respecting user privacy choices and consent preferences
Strategic planning: Better understand your total addressable market and true campaign reach
Competitive advantage: Maintain sophisticated measurement capabilities as direct tracking becomes increasingly limited
ROI accuracy: Calculate a more precise return on advertising spend by accounting for all influenced users
Who should use it?
Performance marketers who need comprehensive campaign measurement for optimization decisions.
Digital advertising managers responsible for demonstrating complete campaign ROI and effectiveness.
Marketing analysts requiring full data sets for strategic insights and performance forecasting.
E-commerce marketing teams focused on understanding true customer acquisition costs across all user segments.
Campaign managers who need to report total campaign impact including non-trackable users
Marketing strategists planning for privacy-first advertising environments and evolving measurement landscapes.
Data-driven marketers who want to maintain measurement accuracy despite increasing tracking limitations.
Regional marketing teams operating in privacy-conscious markets with declining cookie acceptance rates.
How does it work?
Modelled metrics work by observing the performance of users who can be tracked and then assuming equivalent performance for those who cannot be tracked. This process is managed using user context:
Group A: These are users with a "user_context" value greater than 0, indicating that they have consented to tracking and have been tracked previously on the advertiser's website. The performance metrics for this group are directly observed.
Group B: These are users who initially have not consented (user_context = 0) but click an ad and later consent to track upon landing on the advertiser's page. The conversion and visit rates of these users are carefully observed, as they form the basis for modelling Group C.
Group C: These are users who click an ad but do not consent to tracking, either by rejecting the consent prompt on the advertiser's page or having other tracking prevention methods in place. Their performance cannot be directly observed.
The estimated performance for users that do not consent or are otherwise not trackable (Group C) is modelled off of the performance of the users that were non-consented/untracked before they clicked but are now trackable (Group B). We assume the same conversion rate and visit rate for Group C as we observe in Group B.
Final performance data now includes: Group A (observed) + Group B (observed) + Group C (estimated off on Group B performance).
How to access it?
Use the left navigation bar and go to Analytics under the Manage and Measure section. Select Reports library.
Click +New custom report button and scroll down to the Table section.
You can search and add your metrics using +Add a metric button by typing all.
