Data Problems Are People Problems...
This means data solutions are people solutions, too!
Data governance can be divided into three components: people, processes, and technology. In this post, we’ll talk about people. They’re the non-technical, ever-present, and most important component of anything you do with data. As usual.
Relying on people is as risky as relying on technology, and just as much a safeguard. If everyone in your organization is clear on your procedures and policies, and in their role in carrying those out, they will be the first line of defense in a potential data breach, data overshare, or even a data quality debacle.
Building a Culture of Data Governance
Data problems ARE people problems. Overlook this at your peril. Just like most IT projects fail because technicians overlook the human component, the same can be said of data governance projects.
Data culture is the set of values, goals, attitudes, and practices around data, data collection, and data handling. Data culture doesn’t happen by accident.
For data governance to be successful, people need to be included at every step as a forethought, not an afterthought.
Include people in the planning phases by:
- Talking to stakeholders to analyze their needs and understand where a lack of data governance is making their life more difficult.
- Having conversations throughout the organization about the benefits of data governance, and what the risks are of choosing to do nothing, whether that means not implementing any process or failing to improve your organization’s data governance, if you already have some processes in place.
- Make sure everyone understands the goals of improving your data governance and how it will benefit them and the organization.
It’s also important at this stage to decide what the relationship will be between your program folks and your IT folks. Who is leading? What are the roles? More on roles later…
Finally, make sure that your data governance work and the accountability and expectations around it are visible, and that people are held accountable for their part in it.
Improving your data governance will involve effort for everyone. So it’s important to communicate the advantages. For example, data governance results in:
- Better analytics, (analytics based on higher quality data, and synthesized from multiple sources) which enables better decisions
- A more intentional approach to your data. You get to decide what’s most important and prioritize that. Is it privacy? Data quality? Internal use? Compliance?
- Job satisfaction - people know who is doing what, and how they will do it. And you can then ensure that the right people get the training they need so they have the skills and support to do the data governance work that their role entails.
Finally, support from the very top of the organization goes a long way to creating a data culture and sustainable data governance practices.
Start with your People: Not Your Tools or Technology
Data governance is the formalization of behavior around the definition, production, and use of data. It manages risk related to data and improves the quality and usability of data for an organization. So let’s talk about who is doing that behavior.
Role Definition and Accountability
Some stages of data governance are visible to all your data users; some are carried out behind the scenes. Before we can tell someone what steps they must carry out, your organization must determine what roles and people are best suited to take on specific responsibilities.
Role definition and accountability go hand-in-hand. When you identify roles related to managing data, whether those are roles you read about here or roles that fit within your existing org chart, it becomes possible to identify who is responsible for what activities and to hold them accountable for those tasks.
Why is it important to do the (sometimes tedious) work of defining and assigning roles in a data governance context?
- Role definition makes sure that all the most important data functions are being covered and identifies the person or job title that is responsible. This helps you to make sure all the important tasks are being taken care of.
- Similarly, it helps you address gaps in performing essential functions (such as quarantining secure data or checking and addressing data quality problems).
- Without clear role definition you cannot hold people accountable for entering, cleaning, tagging, or doing QA on the data.
- Role definition and assignment reduce confusion and duplication of efforts.
The authors of Data Governance - The Definitive Guide outline no less than nine key “hats”, which fall into four roles: ancillary, governor, approver, and user. A common breakdown of categories for data governance is:
- Knows and communicates legal compliance requirements
- May be an attorney
Privacy Tsar (Governor)
- AKA Director of Data Governance
- Ensures compliance and oversees policies
- Defines policies and processes and ensures they are followed
- May or may not be technical
Data Owner (Approver/Governor)
- Physically implements governance strategy
- Architecture, tools, pipeline
- Technical Person
Data Steward (Governor)
- Categorizes and classifies data
- This is a lot of manual work
- Often done as part of other work
Data Analyst/Scientist (User)
- Primary users
- Run complex queries
Business Analyst (User)
- Runs simple data analysis
- Sometimes these are the self-service users
Customer Support Specialist (User)
- Views but does not analyze the data
- Don’t usually engage with the data, but hold the purse strings
External Auditor (Ancillary)
Role definition is mentioned in every major text on data governance, so we know it’s important. At the same time we would encourage you to follow Einstein’s advice and make your role as simple as possible…but no simpler.
For example, you might consider these three roles:
- Individuals responsible for creating policies and processes related to data, and ensuring they are carried out.
- Individuals who typically implement the data strategy to guide the creation of the tools and pipelines.
- Individuals who are involved in the classification and categorization of data.
Roles and responsibilities for data governance is an obvious place to think about your people and your data governance. Who is responsible for creating policies? For enforcing them? How will decisions be made, especially across organizational units? What processes will ensure that people can be held accountable?
Effective data governance involves identifying key roles, being clear about what responsibilities those roles involve, clearly communicating those roles and providing training when needed, and knowing how people are held accountable for managing and using data. One person might have multiple roles, and roles may not align with job titles. The important thing is to be clear on who makes decisions (or is the holder of those decisions), who describes and defines the data, and who actually carries out the building of the systems that align with those decisions and policies.
Essentially you are formalizing behavior around data. This is the core of data governance. If the many hats and roles above are overwhelming to you, you might think of it in three buckets, we often do.
- Who can see that data? Consumers of data might be anyone from the CEO or ED, to an administrative assistant. People who need to see data (either reports or records) could constitute a role. You could call that person or role a user, a consumer, or whatever name fits with your organization and its use of language.
- Who can make changes to the data? There are many levels to this, and you might have roles within it. Who can enter data? Who can update that data? What about deleting or moving data? In your organization that might be one role or multiple roles.
- Who decides about your organization’s data, in general? Who makes decisions about data standards and data quality?
Ownership of Data
Who has ownership of the data? Depending on the size of an organization, a single person or a cross-functional team may be appropriate to act as data owners and stewards. In this role, they periodically review data quality and practices, set goals for organizational data quality or usage, and provide guidance and leadership to other organization members around all aspects of data governance. If you have a small organization with a single person responsible, then it’s helpful if they are in leadership or at least have a direct line to leadership. If you have a medium or larger organization and share the data ownership across a team, it’s helpful to have a cross-functional team with at least one member of senior leadership involved. This diversity helps to ensure that the policies and practices developed by the team are both practical and strategic.
Ownership (sometimes referred to as being a “data steward”) might mean making sure that:
- Data is produced following certain business rules.
- Data is entered in a timely manner.
- The right people are notified when data are updated.
- The right people are notified when data hasn’t been received or is low quality.
None of this matters if people don’t know or understand their role, or are not held accountable for the roles that they are given. Accountability makes sure the work gets done. Adding data governance tasks to someone’s job description or performance review can be a great way to hold them accountable. It ALSO can be a way of formalizing work that people are already doing, which for some is a welcome acknowledgment.
Another thing to consider in terms of roles is that you may have a person who really understands the content area for a particular area of data in your organization. This person may not fit neatly into any of the roles we’ve defined above, but may be valuable in their understanding of the domain area. For example, someone who understands the finance data, where it comes from, and how it needs to be used. Or someone in fundraising who has content knowledge about what the data sources are, how they are gathered, and what’s needed to make a successful ask of a potential donor.
In this case, some organizations identify “data domain stewards” (choose your own role title). These people can be responsible for all the data in a particular domain, making sure that data governance practices are appropriate for the content area or department, and are carried out properly.
We would encourage you to also identify data domain stewards as consultants and support for your work. A data domain steward may not know the first thing about managing data, but might still be a valuable resource on what makes sense, and how to work with the people in their department.
Communication and Training
Communication and training is not the same as data governance, but it’s the oil that makes the engine of data governance run. It is important in working with people and keeping them from being disgruntled, confused, and frustrated. So we wanted to address communication and training in the context of data governance.
Communication and training facilitate the creation of an organizational culture that values data. Like any well-run nonprofit, you will want to think hard about how you communicate to your staff (and possibly vendors) about data governance. Make sure everyone understands WHY you are doing this work. And also that everyone understands WHAT you want them to do. When data governance policies are created and decisions are made, it can be like a tree that falls in the forest. Ensure that everyone knows what the new policies and procedures are and why it’s important to follow them. You might communicate this at a high level in a staff meeting, and at a more granular level in written documents. We can’t state strongly enough that if you make a decision, you must write it down, and then share it. Writing it down ensures that new people, and people in new roles, have clear guidelines about what’s expected. It’s also the key to ensuring accountability. It’s hard to hold people accountable if there are no written standards to follow.
How do you train those who interact with data to ensure that people are consistently carrying out the decisions you’ve made? This might involve anything from teaching people how to enter data properly, to teaching people how to apply metadata in a data warehouse. You may already know where people will need support, but asking them goes a long way. Once you ask, be prepared to follow up with training, documentation, and support. There is nothing worse than being asked what help you need and then it is never forthcoming. Written training is like written documentation. It helps manage natural staff attrition and migration by reducing the risk of misunderstandings or lack of compliance by shortening the learning curve as employees learn new data governance responsibilities. It can also be used to ensure consistency in how you support staff across the organization, and in how you train new staff or existing staff in new roles.
A Few Issues with Data Governance and the People that Carry it Out
Here are just a few things we see that you might want to keep an eye out for:
Job Titles vs. Roles
- You might be the Development Director, but serve as a data domain expert. Is that part of your job, or a role that you are taking on in addition to your regular job duties? A Program Director might naturally serve the role of the person who ensures data quality is up to snuff, but not all people can serve as data domain experts. When you transfer to a new position, the new Program Director may not have those skills. So you hold a role that doesn’t transfer with the job title. The important thing is to know that the role of ensuring data quality (or HIPAA compliance, or making decisions about data security) can move from person to person, and from one job title to another. It’s important to be clear about what tasks go with what role, vs. a job description.
Informal Knowledge and Subject Matter Experts
- Sometimes data analysts decide what data sets are high quality or what data requires security protocols, and communicate that by word-of-mouth. This is a fragile process that can easily break down when someone leaves. I know we say this a lot, but WRITE IT DOWN.
- Similarly, if the Development Director knows what data source is best for prospecting, don’t assume anyone else knows, and don’t accept a process of verbal transfer of knowledge. Document it.
Definition of Data
- People sometimes purchase software to help with defining or tagging data. While these tools can be very helpful, it still requires humans to make decisions about meaning. Make sure that your tools always have the right flesh and blood support to accompany them. Maybe someday AI will do it all, but for now being able to make meaning out of your data, and use it to then make decisions requires trustworthy data, and human brains. Take the time and include the people to have those conversations and make those decisions.
We hope we have convinced you that the people part of data governance is as important (more important really) as the technical tools that we use to manage data. Next month, tune in for a conversation about another non-sexy topic….processes! Make sure to sign-up for our newsletter to receive our upcoming posts!
At Inciter, we work with organizations to tackle tough questions related to data governance, as well as building data systems. We make sure that you’re aware of the organizational data management issues that need to be in place before we throw you into a high-tech data management system. If you need warehousing, pipelines, data cleaning or reporting help, but you aren’t sure your organization has the policies and practices in place to take advantage of the new tools that are available, we can do a data assessment for your organization at a very reasonable cost. With an accurate data assessment you can put in place the right technology for your organization, no matter what system you choose, and know that you can get the most out of it.
To see how Inciter can help your organization get the most out of its data, contact us at email@example.com.