In Fusion, there are three tables: raw, fused, and links. Since Fusion automates how you combine multiple data sets, it converts disparate schemas to a single, unified schemas, maps fields, matches records across different databases to create a single record, de-duplicates data within a single data source, and more.
Below we'll talk about what fused and raw data contain exactly, and the importance both have when analyzing your data. For more information about how the links tables ties tables together, read this article here.
For each system you connect to your Fusion account, there are a set of objects and fields that we can see and read from the system. When we first start to pull out the data from each system, we are building this Raw dataset.
In this Raw dataset you'll find all records and all fields for the objects that are supported for the given system. For example, if we support reading Contact records from a given marketing system, then all contacts in that marketing system along with all of the fields tied to those records where values are present will get fed down into the Raw dataset. At this layer, the data is still siloed within each given system and no relationships have been made across the connected systems.
Once the Raw dataset has been created, the next layer is the Fused layer. This is where we do a few key processes:
- Find related records across each system and Fuse them together as one unique record (i.e. firstname.lastname@example.org in my CRM and email@example.com in my Marketing Automation system is seen as 1 record in this layer) to match records across different databases and create a single record by using the most recently updated value for a like-field
- Find the field data based on your mappings to build the Fused records dataset
- Find the relationships and apply them to the Fused record so that we know what other records across each system should be seen as associated with the Fused record.
The last point is perhaps one of the most important aspects of the Fused dataset. The relationships allow an analyst to build complicated reports across their entire dataset with more flexibility than what is often afforded him or her within the systems where data originates.
Let’s take a quick look at what this might look like once you feed in your Fused database into your favorite BI Tool. Below is a screen shot taken from Amazon QuickSight. In this example we are pointing out where the Fused Company records live and where the Raw Company records live for our Dynamics CRM.