Introduction to Data UnificationUpdated 26/06/2020
Understand why and how to properly automate data unification with these sets of guides
What is Data Unification
Data Unification is the process of turning raw data from multiple sources, which might be incomplete, conflicting, unclean, and discordant intoa set of unified profiles (or golden records), that you trust and rely on consistenly in every service you use.
In other words, unifying data involves merging a company’s many fragmented data sources into one, single central view that uses the best available information for every single User or Account you reference in your systems.
The benefit of data unification is that it provides a more holistic and accurate view of your many data sources.
It is not uncommon to have multiple tools, all with conflicting or incomplete data about the same users or accounts. Hull makes it easier to not only recognize these gaps and inconsistencies, but to unify them precisely according to your needs and specifications. Without reliable data unification, you are often left mapping similar attributes across various tools, which leads to inconsistent or competing attribute values and a major headache.
Hull allows you to solve this problem by allowing you to compute and map unified attributes that sync to all of your tools.
What's the best Practice for Unifying Data
Store raw data, then generate a new set of unified attributes. We won't get into complicated theory, but the best, "right" way to do data unification is to approach it in 2 separate steps:
- Collect "Raw" data for every user and every service. For instance, store both the
salesforce/company_namefrom Salesforce, and thehubspot/companyfrom Hubspot, each in their own namespace on your unified User or Account (By defining an incoming mapping for each service in Hull. Our own connectors save those attributes each in their own namespace by default) - Compute the "canonical" value of the Company Name, according to your own logic and store it into another group, for instance:
unified_data/company. There are multiple ways to define which entry is "the right one", and we'll see some approaches below. - Finally, Propagate the
unified_data/companyattribute back to your tools (Simply define the outgoing mapping for your services in Hull)
How to compute Attributes in Hull
By default, Hull separates user and account attributes into distinct groups based on the source of the attribute (Salesforce, Hubspot, Outreach, etc.) This allows Hull to cleanly separate and distinguish attributes that belong to different systems. The Hull Processor lets you implement your own custom logic (in Javascript) that uses attributes from all of your connected systems in order to evaluate and transform your data into a single unified group.
When unifying data, it’s always important to ask yourself what your leading system is. For some, this could be the leading system for all of your unified attributes, or it may change depending on the attribute.
Now that these basic concepts are out of the way, let's review some of the possible data unification strategies (of course your own mileage may vary and you might want to implement more personalized and specific variations on those). Follow the links below to learn how to perform data unification with those different strategies:
As always, if you need some help doing that, our Experts are available to help, don't hesitate to reach out!
