Here we are at the final installment in the series - essential number eight. I hope that you have gained as much insight from reading the series as I have from writing it. While the earlier essentials focused on analysis and design, essential number eight pulls everything into a plan for the future.
Essential #8: Figure out the path forward
It's interesting to get someone else's perspective on your work. When I put together the slide for this last essential (I originally presented this idea at Enterprise Data World), I thought the picture was of a nice, gently winding road. It was intended to provide hope that the path toward MDM might not be straight, but it wasn't terrifying. At least my interpretation was that it wasn't terrifying. I had no idea that it could be viewed in a different way. But apparently that nice, gently winding road may have been quiet, but it was also desolate. The grass was green, but there were no trees, people, or buildings. Nothing. Did I mention that it didn't go anywhere either? It just sort of ended. That was something I hadn't noticed until it was pointed out to me. I suppose that just serves to reinforce the need to get some outside perspectives and think about how your initiative may benefit from someone else's advice. (and, in this case, I am hoping that the picture above is at least slightly more hopeful than what I had before!)
Setting the Path
So how do you figure out the path forward? Where do you start? After setting an end goal, assessing what you have, and examining the gaps, you already have a great start in the planning process. Now, it's just a matter of tying the pieces together and prioritizing.
1. The business case is key. Everything in the plan should tie back to the pain points within the business case. Map the paint points and user stories to requirements, then requirements to the design components. You can then take your current state analysis and lay that over the top to identify gaps. Depending on the complexity of the project, you might consider using a tool that has sophisticated mapping and dependency capabilities.
2. Start small to prove value. MDM has a reputation of being a long, drawn-out process that takes years to deliver value. Saying that your first delivery is 3 years out is probably not the best strategy to disprove the myths and keep the program funded. For example, you might start with enough functionality to create identities for a single master data element (and limit the scope of the initial data set to only the most important entities), then, in subsequent releases, add additional data or relationships. Deliver enough to solve at least part of the business problem and prove that MDM is more than just an academic exercise.
3. Prioritize the subject areas of interest. Look for the subject areas of most value that also have a good probability of successful implementation. A data element that sounds critical may also have business process or definition barriers that must be dealt with before a clear path can be set for that element.
4. Be sure to look at dependencies. There may be dependencies in both processes and data. Mapping applications and data elements to a process flow can help identify existing (and potential) issues with governance and data dependencies. Are people within the group that "owns" Customer data, really the first to create identities for that data? Is the success of the implementation dependent on a source system that stores data in a proprietary format and doesn't have existing mechanisms to access the data?
5. Organize steps and communicate with a t-map or roadmap. I find that it can be helpful to take the analysis and design pieces from the previous essentials and group them together into categories that form the streams of a t-map. (more on that below)
I love t-maps almost as much as self-adhesive notes. Almost. Maybe it's those hopeful statements in the upper-right corner?
The key is to align those statements with the end goal that you set with essential number one. Focus on the value to the business, such as providing more accurate financial information or increasing sales.
The streams align with key focus areas for delivery. Governance is an important stream that can be an indicator of success in the other streams. If data quality is an important factor, include that. You might also have a stream for metadata, architecture, and one or more streams for data elements.
Wrapping it Up
As we traveled through the 8 essentials, we set the end-state goals, analyzed our current situation, designed the necessary pieces for the desired end state, and set a course on the MDM road. It may not be a smooth road, but if we take one mile at a time, reassessing as we go, we can reach our destination.
NOTE: Thoughts expressed are those of the author and not her employer.