You are a project manager or a line of business manager and you realize that data quality has a direct impact on daily processes and your ability to realize business goals. You have rolled up your sleeves and want to implement a data quality (DQ) management process. How do you ensure a successful DQ initiative? Here are 3 key characteristics that will increase the likelihood of success. Establish these as part of your approach, and looking for these capabilities in the data quality tools you are evaluating – whether buying off-the-shelf or building it in-house.
- Business-Driven, not IT-Driven
Ingredient 1: Make it a Business Initiative, not an IT Initiative
Most of the data within your enterprise are owned by the business. Consider master data like Customers in your CRM system, which is owned by Sales. Even transactional data ultimately is owned by the business, for example, Purchase Orders are owned by Procurement. Without having a clear business reason and value aligned with your project, you won’t be able to drive the buy-in required to succeed. Studies show a lack of user adoption by business users is the primary reason data initiatives fail.
Most DQ initiatives are driven by IT and most of the data quality tools in the market are exclusively built for IT. The challenge in that approach is the Lack of Ownership. The IT team may not possess the domain knowledge of the data, so they cannot effectively identify the data quality gaps and the business requirements that will drive value. And even if they do identify gaps, unless the business users are embedded in the DQ initiative, that data gap is not going to get addressed proactively. The result is a continuous cycle where actions are happening but the on-going quality of data is not improving.
First, identify the business reason and value associated with your data initiative. Then make the data quality monitoring and maintenance a part of the business process. Data Quality tools should have simple and engaging user interfaces for business user adoption. They should allow business rules modeling without deep technical knowledge of the underlying databases. They should allow for easy modeling of DQ rules, on-demand execution, and provide easily actionable insights for the business.
DvSum DataPARC, for example, enables Predictive modeling. This feature allows business users to type their rules in plain English and the data quality tool uses data catalog intelligence and text parsing to suggest the rule definition.