Simplifying Omnichannel Data Enrichment
When we first launched Tealium AudienceStream as the visitor profile engine for our customer data platform, we first focused on delivering end-to-end real-time data collection and distribution. After raising the bar for data enrichment and actionability, we tackled bringing in omnichannel data, including offline data such as point of sale data. The biggest problem was the variety of data sources and elements that customers wanted to import into the customer profile.
Our first step to solve this problem was to create a ‘many to one’ approach to visitor identification, which you can think of as a wallet. A wallet has numerous cards that all identify an individual. If any of those unique identifiers is accessed and confirmed, the individual can be connected. This is far more powerful and flexible than a traditional ‘primary key’ approach, which relies on a single user ID in order to tie devices and data sets to a person.
Using this approach, we now had a way to not only stitch devices across our user base, but also ‘stitch’ disparate data sources and silos to the profile using any of those unique IDs.
There was one final step, however, that we needed to take in order to truly modernize the data supply chain. This is something we refer to as “profile onboarding,” which focuses on importing legacy CRM and profile-based data that most customers have sitting in various systems. This enables customers to start with any genesis data set and then evolve it in any method they wish. So on Day One, when we launch the customer data platform for a client, we can begin with a rich set of profiles ready for action.
So, today we announce the completion of our omnichannel data enrichment capabilities, where customers can immediately onboard legacy data and then begin enriching it with real-time behavioral insights and attributes for what we think is the most powerful customer data in the industry. All of this, it should be noted, is available within a convenient point-and-click interface. Simple, indeed.