Luckily, a fallback mechanism for ID matching can be utilized Probabilistic modeling attribution. It uses publicly available parameters i. e. device name, device type, OS edition, platform, IP tackle, provider, to name just a few, to form a digital probabilistic ID that statistically fits genuine device attributes.
As such, Probabilistic modeling is less accurate than ID matching, and hence the attribution window is brief, typically 24 hours. In mobile app set up campaigns, the vast majority of the first app opens occur within 1 2 hours, through which case the Probabilistic modeling method is extremely accurate. However, some lack of attributed installs is unavoidable.