Collaboration Best Practices for Data Management/Business Process Reengineering
Is it time to re-examine and reengineer how you create, read, update, and delete data in your enterprise? These tips will help you get off to a great start.
By Theresa Kushner and Maria Villar
A typical data management program evolves in four phases: discovery, detection, correction, and prevention. Manually detecting and fixing data defects -- although a great starting place for showing immediate progress -- is an expensive endeavor that does not scale.
With today's ever-increasing data volumes, real-time business intelligence, and cost controls, we must prevent data defects -- not just fix them. Yet, prevention is often difficult for most large enterprises because it requires getting to the source of the defect and changing the business processes that cause the data quality issue.
The most frequently pursued solution is to implement master data management (MDM) to consolidate processes for the CRUD (short for create, read, update, and delete) of master data. MDM separates CRUD from the originating business process and system and allows a clear investigation of data quality issues. This works well in many organizations.
If a separate MDM solution is not feasible and true business process reengineering is required, the task is much more daunting. The processes that create, read, update, and delete data are integrated into many business and IT processes. For example, customer data CRUD processes in marketing may be embedded in Web registration and campaign management business and IT processes. In sales, processes for telesales call centers and lead management often include CRUD steps. Likewise, order management and service processes embed CRUD processes.
Reengineering all systems that could have CRUD processes is challenging because the owners of the business processes are all different, and the requirements for the data they need or want are different. These processes are scattered across the company, and each owner has implemented CRUD data processes differently using different data standards and data quality checks. However, these are the very processes that have to be changed to prevent further data quality defects.
The objective of CRUD process reengineering is to create new CRUD business processes that are:
Automated. Often the tendency when automating a CRUD process is to make data entry tedious with data lookups and required fields. Engaging an expert user interface team to help minimize the application requirements for automation is highly useful.
Simple to use. Testing CRUD with the user community during design and before implementation helps ensure ease of use.
Compliant with data standards. This step ensures that data created by the simple, easy-to-use automation techniques meets the standards established by the data governance work.
Use common business rules. Regardless of the business process, common business rules for CRUD should be defined. The data management team should lead this effort, working with the business process owners to understand business process requirements for CRUD. One of the best places to enforce common business rules is at the point of data entry. Take advantage of this reengineering effort to reexamine business rules and ensure common usage.
Monitor for data quality performance. Employing data monitoring to ensure that the CRUD process is meeting expectations can help you avoid problems with downstream systems and processes.
Assign accountable parties for root-cause analyses and ongoing data maintenance. Problems will arise with CRUD process performance over time, and adjustments will need to be made based on business changes. This step ensures that available and accountable resources are assigned to review CRUD performance and metrics after initial deployment and throughout the data life cycle to ensure CRUD processes are meeting the business needs.
Because this reengineering effort can greatly affect organizations not directly involved in CRUD, data quality managers may be challenged by organizational barriers. How do data management teams convince business process owners to accept these data management responsibilities and change their processes? How should the data management team participate in the CRUD process design and collaborate with the business process teams? What governance is required to continually manage the resulting process?
As with any reengineering effort, obtaining business executive sponsorship is critical to establish a partnership between data management and business processes. If the company has appointed a chief process officer (CPO), this person needs to reengineer the CRUD business processes, along with other business processes, within the company. This leadership is crucial to the success of the project.
Leadership from the top should support the collaboration of teams across the organization. This takes unique skills. To collaborate effectively with business process owners, data management professionals should learn to speak in business process terminology and understand the basics of business process levels for classifying processes. Become familiar with an enterprise business process description if it exists. You should know the CRUD data processes that are usually level 3 or level 4 of other business processes.
Furthermore, business process maturity assessments are regularly conducted to evaluate the current level of business process development. Such assessments are compatible with other appraisals, such as the capability maturity model (CMM) process model initially used for software development process assessments. Much information is available online for these two topics.
The best time to undertake the reengineering of CRUD is when your company is undertaking a large-scale project. Starting a CRUD project as part of this effort gives immediacy to the task and helps the team establish processes that can be shared with other master data CRUD processes and owners. This makes it possible for the teams to establish an important link between process and data quality using real-world examples.
With the CPO committed to the task, the role of the global data team, including the data stewards and data quality leads, is to enable this change by performing the following tasks in collaboration with the business process teams:
Find all the CRUD processes and identify business owners for each. This task lays the groundwork for the reengineering of a CRUD process as well as the ongoing support and maintenance of the data.
Create a baseline of the data quality of these processes. Taking measurements for data quality at this point enables the teams to report progress against the baseline.
Prioritize the processes against established criteria. Business importance, severity of errors, and number of defects produced are all viable criteria options. These criteria are best discovered and recommended by the data stewards who know the data best.
Identify the master data quality requirements of these critical processes. Clarify what is expected from data elements and from the processes where they are used. Implement a process for identifying requirements and changing them; having such a process in place prior to process reengineering is a must.
Establish global business rules to be used by all business owners. Rules for creation, deletion, archiving, and data quality are important starting points for agreement among business owners. Although creating a customer record by the sales department may require a different master data requirement than that proposed by service or support teams within your enterprise, find common ground so that common rules can be applied and only a small percentage of customized rules are necessary.
Establish global CRUD best practices to be used by all business process owners. Educate, educate, educate.
Automate the rules and best practices with IT solutions. It helps at this phase to ensure that the IT team, which should have been involved with the process from the beginning, incorporates the business rules into the CRUD IT solutions that support the business processes.
Monitor CRUD process execution with data management and business process improvement key performance indicators (KPIs). Pick a few meaningful ones such as data completeness, data consistency, or data accuracy. Also include process-efficiency KPIs, such as process time improvements, manual steps removed, or improved time to complete a CRUD process.
Collaboration between business process owners, data stewards, and IT is imperative for the success of this endeavor. An effective collaboration project jointly walks through the existing CRUD process with both the data and business process teams gathering key metrics such as: the number of steps required to complete the CRUD process, the length of time a CRUD process takes, the number of quality checks to take (and how to know they're complete), the number of people involved (and their skill level) in the CRUD process, and the manual nature of the process. You will also want to establish if the quality level of the incoming data is sufficient, and if not, what reengineering may need to start earlier in the process. The team should understand how many data quality defects are created, and if a process is changed, what change management will be required.
These exercises can open the eyes of business process owners. Document the "aha" moments by videotaping sessions. The facts speak for themselves and become the basis for KPIs in the "to be" CRUD process design. Establishing KPIs that show a direct benefit between the CRUD process improvements and the business process improvements keeps the business process owner involved and interested.
CRUD process responsibilities will be a new learning experience for business process owners. Most did not learn this in business school. Data quality managers and data stewards become the teachers. Take the time to educate them, as their success is your success.
Theresa Kushner and Maria C. Villar are co-authors of Managing Your Business Data From Chaos to Confidence (Racom Books, 2008).
Theresa Kushner is the senior director of customer/influencer intelligence at Cisco Systems where she manages marketing data to drive fact-based decision making. She can be contacted at email@example.com.
Maria C, Villar is managing partner at Business Data Leadership, a company dedicated to advancing the effective management of critical business data. She has over 15 years of executive experience in building data governance and data management programs at Fortune 500 companies. She can be contacted at firstname.lastname@example.org.