IN SEARCH OF GOOD DATA STEWARDS: A GUIDE TO FINDING THEM IN YOUR ORGANIZATION
By Laura Madsen
The elusive data steward. In the wild (also known as the business world), this creature is often referred to as an "applications analyst" or "systems analyst." But I suspect that those titles are just a ruse to protect a valuable resource from being commandeered for a large data warehouse project. Much like tracking an elusive mountain lion, finding someone who can effectively guide your data to complete documentation to ensure appropriate usage is not easy.
In reality, the best data steward is someone who resides in your organization as a systems analyst or an applications analyst and is already the "go to" person for questions about the data from the source system. Every company has those people - they just tend to be very much in demand, and tough for a department to relinquish. Many departments know it's the right thing to do, but it's a bit like giving up your morning coffee habit because you know it's better for you in the long run. In the beginning you feel tired and have a headache, but in a couple of weeks you won't miss it (coffee) or them (data stewards).
So how do you find a great data steward? First, you have to write a job description. Writing a job description will help you determine the primary skill sets you need and make it easier to communicate to others the type of person you are looking for in your organization. Odds are your company won't have a job description for a data steward just lying around. It's important to keep in mind as you start to do research for the job description that many companies confuse the role of data owner or lead data steward with data steward. The data owner is the person that is ultimately responsible for how the data is defined. They are the ones who approve changes to metrics; they are advised by data stewards. Data owners are typically higher up in the organization (directors, for example). Anything that the data owners can't resolve gets resolved by the data governance council. A lead data steward or chief data steward is responsible for ensuring that all the policies and procedures created by the data governance council are carried out. In many cases, the lead or chief data steward writes the policies and procedures that are then ratified by the data governance council. Many job descriptions for data steward available today combine these roles into one, but this is a mistake. They are distinct roles that need to be supported by different people in your organization at different leadership and skill levels. Take a look at this sample job description for a data steward:
The data steward's role is to strategically manage assigned data entities across the enterprise, while ensuring high levels of data quality, integrity, availability and privacy. This individual is responsible for standardizing data naming, establishing consistent data definitions, and monitoring/auditing the data quality. The data steward should focus on business benefits, foster the goal of data re-use and articulate the strategic significance of information to the organization. Other responsibilities include, but are not limited to:
- Serve as the single point of contact for assigned data entity.
- Part of a dedicated group of individuals, dedicated to maintaining and enhancing the overall quality of the data from each disparate source system.
- Perform data quality checks for each data source system.
- Resolve profile and affiliation issues on a case-by-case basis.
ACCOUNTABILITIES
- Work with the lead data steward to help uphold data governance policies and procedures to ensure standardized data naming, establish consistent data definitions and monitor overall data quality for assigned data entities.
- Ensure that data management methodologies include the steps, activities, and deliverables required to consistently achieve high quality data.
- Ensure that the applications adhere to defined data management practices, policies and procedures. Identify and manage the resolution of data quality issues, such as uniqueness, integrity, accuracy, consistency and completeness in a cost-effective and timely fashion.
JOB REQUIREMENTS
- BA/BS Accounting / Computer Science / Business / MIS (or equivalent work experience to substitute for education); Masters a plus.
- Strong understanding of relational database structures, theories, principles and practices.
- Experience with data processing flowcharting techniques preferred.
- Demonstrated leadership skills.
- Must be able to represent the data management team as a key subject-matter expert and resource for the assigned data entity.
- Exceptional analytical, conceptual and problem-solving abilities required.
- Strong written, technical documentation and oral communication skills.
- Strong presentation and interpersonal skills.
- Ability to present ideas in user-friendly language.
- Ability to prioritize and execute tasks in a high-pressure environment.
- Experience working in a team-oriented, collaborative environment.
Next you must decide how much time the data steward will dedicate to the task. This will help determine if you need to dedicate a full-time employee to fill the position or just need an employee to add this role to existing responsibilities. This depends on the scope and criticality of the work at hand. Answering these questions can help you determine what you need:
- How many source systems do you have that have data you need to manage?
- How many of those source systems are "key" to your business? In other words, if you are a healthcare company, your claims adjudication system would be a key system. If you are a marketing firm, your CRM is a key system.
- Do you currently manage data today?
- Are you completing your Business Intelligence/data warehousing project in phases?
If you have a big number for questions 1 and 2 and answered "no" to questions 3 and 4, you need full time data stewards. If you have a low number for questions 1 and 2 and answered "yes" to 3 and 4, then you may be able to have an employee add this to his or her current responsibilities. However, it may still be best to have one full-time data steward for each of your key systems. Those systems will likely be involved in all the phases of your Business Intelligence/data warehousing projects.
After you have a job description and know the time commitment that is required, your next step is to make a presentation to your data governance council. (If you do not have one, stop here and go back to the drawing board. No data steward will be successful without a governance council directing his or her work.) Present a formal request to your governance council to fill the data steward position(s). Your best data stewards are embedded in existing departments today. You have to convince the council that giving them up to become data stewards is better for the company than leaving them where they are. Some key pieces of information to convey in this presentation are:
- These targeted employees are already doing work that could be defined as data stewardship.
- Data stewards will reside in the business to stay close to the users of the source system.
- Good data quality can positively impact the bottom-line of the company.
- Bad data quality can put the company at risk.
- Stewardship will improve our ability to make decisions based on data (not on hunches).
- Supporting appropriate use of data is important.
It's the responsibility of the governance council to ensure that the organization's data is governed appropriately. That duty should encourage them to identify the data stewards for their departments. If a department can't name a steward, then HR may have to start searching for one.
As you look for a data steward, remember that data and source systems can be taught, but the ability to communicate effectively to a broad audience, foster the goal of data re-uses and articulate the strategic significance of information to the organization cannot be taught. I recently had a conversation with a lead data steward, and I asked him what he considered the most important attribute of a data steward. He said, "Communication." The ability to have a democratic approach and build trust and support was far more important to him than the knowledge of relational data structures, data modeling basics or data analysis experience. All of those things are helpful, but the ability to communicate was the foremost attribute in the data stewards that he selected. Good communicators make good data stewards.
Another critical attribute is the ability to consider opposing viewpoints and come to a reasonable resolution. If you want to see your analysts get fired up, put them all in a room with a couple of data stewards and tell them you are changing the definition of field. That is exactly what you should do, but it's a painful experience. The data steward's ability to manage those conversations is absolutely critical to the long-term success of the program.
When looking for a data steward, it's important to remember these things:
- Write a job description.
- Look at your existing staff to see who is already fulfilling this role.
- Rely on your governance council.
- Seek out the good communicators.
About Laura Madsen
Laura helps companies understand the value of their data. As the Health Care Practice Leader for Lancet Software, Madsen brings more than 10 years experience in Business Intelligence (BI). She served as Director of Business Intelligence at a pharmacy benefit management company where she led an enterprise-wide Business Intelligence project. She is also a veteran of United Health Group, where she managed a Business Intelligence tool suite. Through this work, she has started or supported more than a dozen companies with BI initiatives. You may reach her at lmadsen@lancetsoftware.com.