Over 90% of internet users are using more than one device to accomplish a task over time, and the average US household now possesses 5.5 connected devices. It’s becoming more challenging than ever for brands to distinctly identify a consumer interacting with their digital channels due to reasons ranging from cookie-less environments to privacy considerations.

For a brand, there are some marketing tools whose effectiveness hinges on being able to uniquely identify the target consumer. A DMP is one of those tools. For a DMP to qualify a consumer as part of a segment, it needs to be able to first stitch together all the data for that consumer. And the effectiveness of that is predicated on how well it is able to link together the ‘IDs’ of the consumer across devices and channels.

A DMP needs to identify both the who the user is and the what the user does in order to qualify them into a marketing segment. Therefore, for a DMP, close on the heels of the challenge of identifying a consumer’s cross-device-cross-channel identity is the challenge of uniquely identifying their behavior across devices.

There are three conceptual approaches to solve both these problems.

In this post we will conceptually talk about deterministic, probabilistic and hybrid approaches