Start With Data Governance for Better BI

Data governance may not have been your dream job as a child. But, it is a critical piece of your data integration and business intelligence strategy. Data governance “defines roles, responsibilities, and processes for ensuring accountability for and ownership of data assets across the enterprise.”

When done right, data governance defines who owns the data and the processes for how they’ll be used. In this article, we’ll discuss the foundational processes you need to have in place to ensure effective data governance for your organization. Is your organization well along the path of data integration powering your business intelligence (BI)?  Or are you starting out and getting ready to import data into your BI engine? No matter the situation, there are some things you need to do right now to ensure your organization is able to get a rewarding experience from your BI. 

This advice applies to all types of data. You may be working with structured and organized data. Or, you could be working with unstructured data, like from marketing or sales tools. Either way, establish data governance practices up front. Ensure you can get useful intelligence from your data and that your work to incorporate it is not wasted.

When beginning or revising a data governance process for onboarding new data, there are three critical paths to consider: naming conventions, implementation plan, and leadership buy-in.

Start with Naming Conventions

Naming conventions are the foundation of your data governance practices. There are many theories out there about the right way to name your data. Look at them if you’d like, but just know the best way to incorporate naming conventions is whatever works best for your team. And, it’s important to do them first, before you begin importing data

When you’re considering a naming convention, first understand the systems you’re going to bring into your organization and how they will connect to your business intelligence. Understand what they bring to the table and what the whole user story will look like. 

Then, decide on a naming convention that will work across different systems and data types. Here’s one small example. It starts out simple enough and gets increasingly complex as you get more granular with the data types. 

1. System – object. Say you’re looking to bring in Salesforce data. Salesforce would be the system and the object would be the data type or field you’re going to bring in. The name would be Salesforce – contact. 
2. System – object ongoing. This lets you know what you’re doing with this dataset. Are you grabbing a whole history with this ETL (extract, transform, and load) tool? Or only yesterday’s data and appending it to the dataset? If it’s the latter, the name would be Salesforce – contact ongoing. 
3. System/System – object/object. If you’ve integrated systems outside of your BI tool but want to bring in that integrated data, you’ll need naming conventions to reflect that. For example, you could be using Pardot to track prospects and have it integrated into Salesforce. The name would be Salesforce/Pardot – Contact/Prospect. 

This is a simple demonstration of what a naming convention could be. Find what works for your team. Understand that the more systems and granular data you bring in, the more you’ll need descriptive and consistent names for all systems and ETLs.  

Get It Right on Implementation

You love having your data integrated. You want all the data integrated. You want to bring in every system so you can start doing analysis tomorrow. But, if you don’t do it right, you can end up causing more headaches and less business intelligence down the road. Consider the following steps when you’re at the beginning of integrating a new system. 

Determine the best way to bring in data from the system. Ask questions like:

• What data am I bringing in?
• How long will it take to bring this data in?
• Do I need all of the data or just pieces of it?

Map out how to connect the data to other systems. 

• What value will this provide to other data?
• What will the naming convention look like?

Determine if you need historical data.

• How often do I need to bring this data in?
• Will I be able to get a one-time historical load or will I be bringing in new data monthly or daily?

QA the data against the original system. (This is the most important step and should take the greatest portion of time during implementation of the new system).

• Is the connector bringing in the data I expected?
• Do I understand how the connector is working?

Communicate with leadership. 

• Share when the system is launched.
• Talk about what to do with the data and how they can use it. 
• Describe the potential benefits of this newly integrated system. 

This process can also be helpful when establishing a new ETL pipeline to bring in just a new piece of data from an already incorporated system. 

This entire process should take as long as it needs to, even months, if necessary. While you may want to move fast, taking your time to do it right will only increase your business intelligence opportunities. It will also ensure you don’t run into headaches down the road with mismatched data or a lack of transparency into where data is coming from. 

Get Leadership On Board

It will be critical to get leadership on the same page with data governance. This will ensure a smooth rollout and support for what can seem like unimportant details and steps on the BI journey. Here are some tips to help them catch the data governance vision:

1. Explain to leadership that data governance doesn’t cost anything. 
2. Identify the person who defines what data governance is for your organization. 
3. Outline naming conventions and strategy for a rollout. 
4. Get data governance integrated into your team’s goals so every person will have ownership and buy into the concept of data governance.

Common Roadblocks

Data governance is critical, but it won’t come without its challenges. You may not immediately see things moving as quickly as you’d like. While all companies want to be agile and responsive, in the long run following your process for data governance will ensure your BI agility won’t be hampered by a haphazard data integration.

Using naming conventions, standard implementation processes, and, finally, getting leadership buy-in will result in a cohesive data governance strategy moving forward that will pay dividends in your BI strategy. These steps help build credibility for your team and provide an accurate foundation to build on.