Enterprise Data Warehouses vs. BU-Specific Data Stores

Move over, content! At ABC, we feel DATA is king. Why? Because without data you can’t have content—or at least, you can’t have good, personalized content that’s relevant to your customer.

Most recently, we’ve been working on quite a few data architecture initiatives (strategy, architecture design, platform evaluation, and operating models). Which is great news! That means business leaders have prioritized building and having best-in-class data environments for their companies. 

In general, we are not fans of data siloes, but we do understand the value of separate databases to serve the needs of different Business Units (BUs). We’ve listed the pros and cons of having separate databases below. 

Centralizing all data in an Enterprise Data Warehouse: 

PROS: 

• Standardized view of the customer data 

• Significantly increases chances of providing consistent CX as all engagements are tracked in one location 

• Makes preference centers possible (when a customer opts in or out of various types of communications) 

• Allows true data governance and compliance across the Enterprise  

CONS: 

• Can be extremely expensive to maintain and refresh all data 

• Can be difficult to align and manage all stakeholder requirements for data (both to/from the MDW) 

• May increase chances of significant data disruption or risk to the business as a single point of failure (sweeping changes to the database can impact all data sets) 

• May impede BU needs for agility/speed to access and use data 

A BU-specific Data Store: 

PROS: 

• Highly specific, targeted view of the customer (based off BU requirements) 

• Robust data sets and insights tailored to the BU’s information and interactions with that customer 

• More cost-effective to maintain a smaller database (with less fields), which may not require integrations to external platforms 

• Does not require standardization or reconciling with the Enterprise Master Database or going through company-wide data governance 

 
CONS: 

• Can be difficult to have “one view of the customer” as the data is not shared back to EMD and collective insights is not possible 

• May be difficult to provide a consistent CX across their customer lifecycle (across BUs) 

• Potential for increased gap between the Enterprise’s data roadmap and the BUs 

Which is preferred?  

At the end of the day (and quarter), we believe in a hybrid approach between these two models. We recognize the need for BUs to have their own specific customer data that’s more relevant to their goals (for example, a marketing team needing to track email opens by product, which may or may not be relevant to a finance department).  

We also recognize the critical need for “core data” or fundamental customer data to be shared back to an Enterprise Data Environment, to ensure the customer’s experience with your company is kept intact, and your company continues to be able to know how the customer is engaging with your company (so you can best serve them what they want when they want it.) 

Ultimately, the goal is to design a hybrid model that serves both the need-for-speed and targeted requirements of business teams while also following the best practices/compliance/holistic approach of data/IT teams. What this means is there should still be a workflow for the BU’s data store to share back to the wider database, and what that also means is the business/IT teams will have to come together (more than once, and more than once a year) to revisit their data governance and business rules engine. 

We believe that more than one data store can exist at a business—as long as they perform at their most lean and efficient, and continue to deliver the best and consistent CX to the customer.