One Version of the Truth is Not Attainable
I recently met with a financial executive of a large distribution company and he expressed to me his frustration. The company had invested in a Data Warehouse and deployed a leading BI solution and yet, they still had issues with “one version of the truth” (OVOT). They spent significant hours trying to validate which numbers were correct.
With a slight smile I said, “One Version of the Truth is Not Attainable.” He looked at me in shock, because many highly paid consultants had told him he just needed to invest in a data warehouse and master data management (MDM) and then put a BI tool on top. I continued to explain that for most organizations; especially those growing through acquisitions, one version of the truth would not be a reality. However, there are definite steps and considerations that can get you closer to “One Version of the Truth”.
1. Leadership needs to come from the Business Executives. Guiding your organization to realize the benefits of OVOT requires strict discipline. The Executive Sponsor needs to plan out their strategy for implementing the solutions, while looking at both the short term and long term goals.
a. Putting the right resource in place. Including knowledgeable subject matter experts in the core team may require pulling them away from their regular jobs and having others backfill, but the benefits will be tremendous.
b. Providing appropriate funding. By setting short and long term goals, you can prove incremental ROI’s and justify future expenditures.
c. Setting priorities. The leadership team needs to set the priorities and communicate them to the entire organization. This should include establishing incremental steps to measure the ROI along the way. Often times, other organizational priorities will be in conflict with the OVOT initiative, so executives must set the priorities.
d. Identifying and communicating risks of bad data. The business is really the owner and creator of data. IT protects it. The executives need to communicate that. To be successful, the business must take ownership and understand the risks of bad or incomplete data. When this message comes from the top, people listen.
2. Master Data Management (MDM) is not an IT project. The business must be involved in MDM or cleaning and managing the data. We recommend that someone with in-depth knowledge of the business be assigned to the team, as they can understand and articulate the impact of poor data quality on the business.
3. Strict change management. Once the data warehouse and MDM foundation have been established, strict change management rules must be established and enforced. Various subsystems and business units should not be operating in silos where their changes can impact other areas of the business.
4. Silos of data in an acquisition. The reality is that if you are in acquisition mode, you will always have data sources that are not fully migrated to the corporate structures and systems. You must plan to have short term and long term solutions to deliver reporting and analysis that includes data that has not come on board. I refer to this as the MacGyver approach, as you have to make it work with what you have on hand.
5. Eliminate “duct tape and super glue” reporting. Many talented BI report writers have been instructed to just make the reports work, so despite challenges with the data, they write queries and code into the reports to make them work. I call this the “duct tape and super glue” method. The report writers need to work with and communicate the data challenges to the data team so they can correct issues at the source. This issue is magnified as the organization pushes out self-service to the end users. Duct tape and super glue will not withstand the pounding of end users.
If you would like to learn more about how to navigate data and realize an increased ROI on your Business Intelligence investment, contact Lodestar Solutions about our new Executive Coaching for Business Analytics at 813-254-2040 or email us at Sales@LodestarSolutions.com.