Welcome back to the Adobe Experience Platform (AEP) education series. In our last blog post, How Adobe Experience Platform Drives Data Transformation, we reviewed how data transformation can help drive CDP adoption.
Customer data platforms (CDP) have allowed marketers to gain a unified view of a customer across online (digital interactions) and offline interactions. CDPs allow for the ingestion of large and small datasets to build a unified view of the customer. Bringing these datasets from different data sources (data warehouse, digital analytics, point of sale, decision interaction tools) is not always a turnkey solution.
Bringing data into the CDP without any transformation may work in some scenarios, but as noted in the report, A Collection of CDP Fails: 5 Stories to Learn From, a CDP should not be viewed as CRM replacement nor a database. Instead, it is a collection of capabilities. Merkle’s view of data into a CDP solution is what we refer to as data fabric. The data fabric is comprised of three layers: data lake, conformed layer, and the consumption layer.
As noted in the image below, the data lake is the initial layer when data is captured in its raw form. Data in this format is typically unstructured but could be structured if coming from a CRM platform. CDPs offer a data lake for consumption of data that can pull data into this layer. If a CDP does not have a data lake offering, one might consider leaving this raw data within their own data lake.
The second layer of data fabric is the conform layer. Data in this layer is conformed to common data definitions. The conform layer is used to create a common data layer from the raw data layer. Conform objects can be address standardization or customer purchase standardization from different sources in the raw data layer. As an example, within an Adobe Experience Platform (AEP) CDP deployment, the query service can be used to create common data definitions.
Lastly, is the consumption layer. This layer is what marketers will be interfacing with and creating audiences from within the CDP. This layer will display data which may have certain data rules applied for the customer’s unified profile, for example, average transactions over the past year. Oftentimes, this customer-focused view of data is overlooked which can lead to unrealized technical debt within a CDP implementation.
In conclusion, bringing your data from source systems can be achieved within CDP platforms however, your first-party data in from both structured and unstructured data sources is not always a straightforward source table to CDP data ingest mapping. Moving into a data design format to ingest data from a consumption layer will allow for faster realization of use cases.