Top Down vs. Bottom Up
Have you heard the terms “top-down” or “bottom-up” associated with approaches to Data Governance? If so, do you think top down is the only way to execute your Data Governance Program?
Well think again…
A top-down approach to Data Governance is usually driven by regulations, audits, compliance or it’s a program driven from an executive. These approaches are advantageous to organizations with large budgets and industries that are highly regulated or have a desire to put in policies and procedures for data driven activities.
What if your organization is not driven by regulations or doesn’t have the budget to support dedicated resources? Policies and procedures aren’t the driving force for your company? That’s OK! What if you don’t have an executive sponsor or funding for dedicated data stewards? TOTALLY FINE!
Getting Started with Bottom-Up
You are in luck!! The bottom-up approach to Data Governance is alive and well. So, where do you start?
Over the last year, many of my clients have been having issues with definitions, quality of data and ownership around metrics and reports. Is this a valid use case for Data Governance? ABSOLUTELY!
Does this sound familiar?
“My company has issues with reports/metrics having different numbers but similar names but different definitions.”
These are all reasons to begin a data governance initative. All you need is a use case to drive change within your organization. You take that use case and begin putting together templates, processes, ownership and definitions in a centralized location to showcase awareness.
If your Data Quality issues are plaguing your source data:
- Start profiling data where the issues reside.
- Showcase the profiling results to the stakeholders who care about the accurate/inaccurate data.
- Turn the prioritized profiling results into repeatable rules to show trending and tracking.
- Perform root cause analysis for quick wins.
- Continue this process.
When you present the results to the business and IT stakeholders, this is the start to your Data Governance Committee for a specific domain. You just started your data governance efforts.
If Metadata is needed to clarify definitions or to understand what reports reside on which data platform:
- Create an Excel document for those pieces of information required in a catalog.
- Not sure what data elements should be in a catalog? Get a demo of a few to see what is being captured in the tools.
- Document owners/stewards, definitions, data platforms, acronyms, data lineage etc. When this spreadsheet grows and becomes management, then you can present a use case for a tool to make this process easier and more streamlined.
If you are struggling with establishing Ownership for Metrics/Reports or you don’t know who to contact when technical issues arise in reports:
- Owners of Metrics or Reports are usually the people who originally requested for these metrics/reports to be created. We aren’t talking about the business or data analyst that wrote the requirements. Who is the VP or executive that funded to have the metric created?
- Still can’t find an owner? Ask if your BI team has a data usage report to see who is viewing or using the report most often.
- Another approach is not always the most desirable, but it can work. If you take away the metric/report and wait to see who reaches out to bring it back online, then that person may be an owner.
When establishing a use case for Data Governance, the best way is to just START! The additional use cases will begin to arise and evolve over time. You may be surprised at how it organically builds itself over time.
What About the Data Governance and Data Quality Tools?
Has your organization ever bought a data governance or data quality tool that sat there collecting dust with minimal user involvement? Why is that?
Did you buy a tool before a process was established for data governance or data quality?
It might be beneficial to understand the process of capturing metadata, profiling and remediating data, and completing metrics needed for the initiatives, before having a vendor come in and make your process more efficient and automated. Understanding who will be doing what and when in the process is critical to establishing consistent and accurate processes.
Can IT Drive Data Governance?
Why not? If in organizations where IT is known to “own the data” or IT fixes all the data issues, then why can’t you start executing data governance initiatives from where you are in your data journey?
This is where things get interesting because you have to SELL the Data Governance and Data Quality concepts back to the business. Why should they care if IT owns the data, right?!?!
An approach to overcome this situation is by establishing ownership to data from the business side. IT can facilitate the ownership by selling the value to the business, process owners etc. After your program starts to grow, a need may be established to have someone from the business begin to facilitate the Data Governance initiatives as the hype grows around your organization.
If you have use cases or problems which could use the data governance/quality principles to help solve them, then you have a platform and reason for a data governance program. Remember not all programs are the same or need the same tools. Start with where you are in your organization and expand to see how things grow over time.