Policies and Process During the Sustain Phase of Data Governance
Consistent Right Action Leads to Consistent Right Results
If you’ve ever had a sewer backup in your basement, you have some idea what the result is of not maintaining all the hard work you did planning and implementing your policies. Sure, you’ve got plumbing, but if you don’t do the maintenance…bad things can happen.
Actually, it’s not that dramatic. As with any initiative, not maintaining the work you’ve done is more likely to result in reverting to the status quo. In a data governance project, this means a return to poor data quality, siloed definitions and systems, and reports you don’t trust. But more than the sadness of having poor quality data (or data you can’t even locate) is the impact on the data culture and the morale of the people in your organization.
Once you get people on board, and they do the hard work of deciding on, creating, and communicating roles and policies, there is nothing more dispiriting than watching it all go away because everyone stopped paying attention to the policies once they were finished .
We can’t say this enough, we see data projects fail FAR more often because there is not enough attention paid to business processes and people, than because of technology problems. Sure, that 20 year old database in your server room that was built with dBase makes us want to cry, but the success of the fancy new system that you implement to replace it is totally dependent on your people.
So where should you focus your efforts on sustaining the processes and policies you put together during the sustain phase? We suggest you focus on two areas.
Create a mechanism to monitor data governance policy implementation and to address lapses:
It’s one thing to create policies, it’s another to know if people are actually following them. You won't know how to sustain and support the processes you’ve created unless you have some way to monitor whether they are being implemented. This may be as easy as creating a spreadsheet where analysts or other data users track issues and ask questions. You could also use a simple spreadsheet to track whether users have reviewed and signed off on policies.
For processes, let’s look at data quality and how you might track that. Some organizations use a data quality scorecard. A scorecard might measure how current data is, how complete it is, and how accurate it is, for example. You might also create a measure related to duplicate data. It depends on your use case, but once you’ve decided what the key measures are, tracking them over time will keep your data quality from slipping too far.
If you are just starting out, you might decide to prioritize your data quality measures. If missing data is the biggest thorn in your side (or clog in your pipe), consider starting there.
Communicate and Train People - Not Just Once, but Ongoing:
We will repeat ourselves on this a lot. Organizations are made up of people. People handle all the data. So you can’t have good data without supporting your people. Managing people and managing data really are the same thing. You want to create clear standards around what people’s roles and responsibilities are regarding how they handle and support trustworthy data. And then you need to support them, which means training them.
You will of course train everyone initially, and train new employees, but having regular professional development around data governance, tailored to different roles, will help support the data culture around trustworthy data. There is nothing more frustrating (or ineffective) than being trained in a new process or skill, and then not using it immediately and regularly. You forget and then have to learn it all over again. It’s also bad for morale and staff will start to roll their eyes every time there is a new data governance training. Also keep in mind that turnover happens, and be prepared to train new people, and train people when they change roles.
While you are at it, in the spirit of tracking and accountability, decide which trainings are mandatory and for whom, and keep track of who has taken the trainings and when. Whoever is conducting the training will have put a lot of time and effort into it, and you want to be able to see who took advantage of their hard work. Training is sort of like data quality, if you don’t track it, you don’t know if it’s happening, and you can’t address shortfalls.
In a smaller or more informal organization, you may also be able to create a peer-to-peer learning environment, where staff support and train each other, as long as accountability is clear. As with your policies, you will want to have key metrics around your training, to make sure staff were trained on time, with the appropriate information, and had a chance to ask questions and get support.
It’s about trust
Finally, while data governance is about having trustworthy data, you want your staff to trust the leadership of your organization around this as well. You want them to trust…
- That this is not just another useless exercise that will be forgotten
- That others in the organization are doing their part
- That when leadership said data governance is important, they meant it, and will dedicate resources to train and support those doing it.
The bottom line is – data governance is an ongoing process, and it’s never over. And that’s partially because technology changes, data models change, and systems change. But it’s also because people change, grow, move into new roles, and also forget. Make sure to build in the support, training, and communication to your processes and they will yield far more in terms of stable, consistent improvement in your data governance processes.
The next part of our Data Governance series will take a deeper dive into Technology. Sign-up for our newsletter to receive the next post (and other helpful resources) in your inbox!