Some thoughts on building a successful data governance organization
Now that we’ve talked about the role of organizational change and education, readiness assessment, developing a strategic roadmap, creating a business case and designing the Data Governance initiative before trying to execute on it, it’s time to talk about the implementation process itself.
At this point, you’ve gotten everyone up to speed on what master data management is, and how critical Data Governance will be to successfully managing master data. You’ve figured out where the weak spots are, in the organization, in the current MDM and data governance processes, in the technology landscape, and in current information management efforts. You’ve accurately described the “As Is” state, and worked with the organization to envision the “To Be” or future state. And you’ve spent the up front time to design a new Data Governance framework for the enterprise going forward.
So it’s time to execute – to deliver on the vision. The client company has a lot of responsibility here. We’re all familiar with the saying “you can bring a horse to water, but you can’t make him drink.” If the client has problems making decisions and carrying them out, the Data Governance practitioner is going to struggle, no question about it.
I usually start with the organizational side of things, and I start at the top. If the senior executives aren’t bought in yet, back up and find out what it will take to get them there. These initiatives truly cannot be driven “from the bottom up”. It’s true that you’ve got to get the various lines of business on board, and to have the active support of IT. But if the “top of the house” isn’t there yet, you won’t get too far with your Data Governance implementation.
If your executive sponsorship is solid though, and you’ve got the makings of a Data Governance Executive Steering Committee, you’re ready to start worrying about building out the Data Governance Office. This is like a Program Management Office for IT projects, or a “Finance Department for Data”, as I’ve sometimes described it.
You’ll probably want to start with an internal person who already knows the business and who brings some existing political capital and relationships to the role. Knowledge of the business and the right leadership traits and personal characteristics are critical. Functional experience in Data Governance is certainly important, but it’s very hard to find internal candidates that know the business, have the right leadership skills, and have DG experience to boot. So go for the first two, and use your Data Governance consulting partner to bring the new Data Governance Office leader up to speed.
Of course, you may have to go outside the company, but try to avoid that if you can.
Then you can start filling the other positions in the DG Office. Look for people with internal experience managing data, and make sure you’ve involved them in the earlier steps of education, readiness assessment, strategic road mapping, and data governance framework design. That way, they should be bought into your strategy, since they had a hand in shaping it.
After settling the Executive Steering Committee into their role, and filling the necessary roles in the Data Governance Office, you’re ready to get the Data Stewardship community in the business (and the corresponding IT stewards) lined up. These folks should have been involved from the beginning too, taking part in the educational workshops, in defining the Current State and the Future State, and in buying into the Strategic Roadmap and Data Governance Framework / design.
You’ll need to spend a fair amount of time socializing the strategy and the design with these folks, so that they see being involved as a data steward as just the next logical step in their involvement in the Data Governance program.
Set up individual data stewards in the various functional areas, business units and geographies. Almost all of the time, these are going to be existing people in existing roles, but the Data Governance program is going to be formalizing their responsibilities around creating and maintaining corporate master data and reference data.
You’ll need to work with Human Resources in most cases, because job descriptions will have to be revised a bit, and existing compensation structures revisited. This is because the data stewards naturally will tend to do what they’re financially incentivized to do, like most people. So be sure to allow plenty of time for this reshaping with HR, and make sure your Executive Sponsor and Steering Committee will back you up when making these changes.
From a process perspective, a periodic meeting between the Data Governance Office and the Steering Committee is a good idea. And forming a Data Governance Council, consisting of the DG Office plus some or all of the data stewards, will also be useful. This group may need to meet weekly at first, until you reach some type of “steady state”, and can shift to bi-weekly or monthly meetings. Of course, if some big projects like an ERP upgrade or a CRM implementation come up, you may need to go back to weekly meetings.
It’s also a good idea to set up a “data help desk”, which allows anyone in the company to notify the Data Governance Office of a new data related issue. This type of case management capability will help ensure that issues aren’t falling through the cracks.
In terms of technology, make sure you’ve got the Five Essential Elements covered:
(1) an MDM hub (either already implemented, in the implementation process, or at least planned for the future via the strategic roadmap),
(2) adequate data integration technology (an ETL tool by itself is probably not enough, you’ll want a service-oriented architecture or service bus based approach to handle real-time or near real-time integration)
(3) a data quality tool of some type, with strong profiling, standardization, matching, de-duplication, and “golden record” creation functionality
(4) third party enrichment capability, particularly if you’re dealing with the Customer domain – don’t put this off to a later phase, the information you don’t have which is available from a third party provider like Dun & Bradstreet and Equifax (for information on businesses) or Acxiom (for information on consumers) can be invaluable in answering the questions the business has
(5) technology to facilitate data governance, including tools for metadata management, data modeling, information access and security management, workflow and business process management, and policy management. You may not need to buy all of this right away, but Data Governance for a large enterprise is a big task, and anything you can do to automate routine tasks will usually have a positive ROI. Don’t try to do everything using Excel spreadsheets, Word documents, and e-mail. That type of “least common denominator” approach is a false economy. Technology won’t ever be a silver bullet for Data Governance (about 80% of the work will be with organizational, cultural, political and business process matters) but don’t ignore it altogether; it can be a powerful “force multiplier”.
I’m alway amazed how many companies try to do master data management and Data Governance without one or more of the above capabilities.
If you’ve got your organization built out, your processes designed and in place, some new technology fitting in to automate labor intensive tasks, and your information is starting to get some dedicated scrutiny from the centralized Data Governance Office and the distributed data stewardship community, you’re going to be well on the way to a solid Data Governance implementation.
Just remember that there are five or six levels of maturity in most Data Governance Maturity models (the National Association of State CIOs has a great review of them). So after everyone agreed on the current level and desired new level during the readiness assessment and strategic roadmap phases, make sure you manage expectations that the company isn’t going to magically go from a Level 1 to a Level 5 in one year.
One company I’ve worked with and stayed in touch with has been at this for ten years, and at a recent conference, I heard their CIO say they only graded themselves a “4” on a 5-point maturity scale for how they manage their master data.
So be realistic with everyone on how long things are going to take, and the fact that the company is usually embarking on the first year of a continuous improvement process that requires organizational change management, cultural evolution, process redesign, technology adoption, and a new dedication to managing information.
To be a “data driven company” – which I believe is required to succeed in the twenty-first century – is a lot of work. And one of the most important tasks for the Data Governance Office during Year 1 is to be planning and working towards Year 2 and beyond.