Note: Adtriba Core is a legacy feature that we continue to support but no longer sell. The latest version of this feature is available as Funnel Measurement.
If you are already using Funnel Measurement, find more information here. To upgrade, contact us.
Overview
Cross-device tracking refers to identifying and connecting a user’s interactions across multiple devices into a single, unified customer journey. For example, a user might click on a mobile ad during their commute and later complete a purchase on a desktop device. Without cross-device tracking, these two interactions would appear unrelated, reducing visibility into the effectiveness of the first ad.
Cookies, the most common web tracking mechanism, are limited to a single device and browser. This makes connecting events across devices challenging without additional methods.
Tracking Approaches
1. Deterministic Tracking
Deterministic tracking relies on a unique identifier that is explicitly linked to a user, typically when they log in. Platforms like Google, Facebook, and Amazon can do this easily because users are almost always signed in while using their services.
Any advertiser with an account-based system can also implement deterministic tracking by capturing user identifiers at key points such as login, sign-up, checkout, or newsletter subscription.
Advantages:
High accuracy with low false matches.
Transparent matching process.
Limitations:
Requires users to be logged in.
If using third-party platforms, data may be restricted to that platform (“walled garden”).
2. Probabilistic Tracking
Probabilistic tracking uses statistical models to infer whether multiple devices belong to the same user. This inference is based on signals such as:
IP address
Location data
Browser type and version
Operating system
Ad interaction patterns
To train these models, vendors often combine probabilistic methods with deterministic data as a baseline.
Advantages:
Expands potential reach beyond deterministic matches.
Can be applied across different ad platforms.
Limitations:
Trade-off between precision and match rate:
Precision = proportion of identified matches that are correct.
Match rate = proportion of actual matches that are identified.
Vendor algorithms are often proprietary (“black box”), making it difficult to validate accuracy.
Privacy considerations.
Considerations for Marketing Attribution
When probabilistic cross-device data is used in attribution modeling, it combines two statistical systems — the tracking model and the attribution model. This layering can make it difficult to determine which part of the system is contributing to errors or inaccuracies.
Deterministic data is generally preferred for attribution because it provides verifiable matches. However, deterministic data may underestimate the true cross-device effect because not all users log in.
Adtriba Core and Sphere Approach
In Adtriba Core and Adtriba Sphere, deterministic tracking is prioritized wherever possible. This is achieved by offering user-ID tracking that can link interactions across devices when a user:
Logs in
Signs up for an account
Completes checkout
Clicks through from an email
Subscribes to a newsletter
These linked journeys are then processed by machine learning-based attribution models to give a more accurate view of campaign performance across devices.
Encouraging users to log in is a best practice for improving match rates. When combined with marketing attribution reports, this approach can reveal the real influence of mobile campaigns on desktop conversions, often changing the perceived ROI of certain channels.