You're navigating conflicting data quality standards. How do you ensure accuracy and consistency?

Conflicting Data Quality standards or standards that do not add value are very common in Data Quality Management. This is where a set of precise goals would help to bring the focus & efforts back on track and weed out the non-value adding or conflicting DQ standards. Definition of the goals should be simple and easy to understand, even for a completely non data professional. It should clearly tell about what metric is to be achieved and answer the following 'so what?'.

The key is to establish a common ground by bringing all stakeholders together to agree on what's non-negotiable. Start by identifying the most critical data points that impact decision-making and set a baseline for accuracy and consistency. From there, document these standards and make sure they're easy for everyone to follow. Regular audits and feedback loops are essential, so keep tweaking as you go to ensure the data stays reliable across the board.

There are two approaches to data quality. The business perspective and the technical perspective. The business perspective is usually based on business rules and is very business related and business and industry specific. The second one is rather technical and has much to do with data profiling based control environments.

Perform data profiling and build data quality metrics, leverage on these data profiling understanding and data quality metrics to establish consensus of data quality standardization across the organization with data reconciliation rules to deal with any future conflicting data quality issue. Perform regular data quality monitoring audit with master data management program.

Every action of yours in Data Quality Management should answer 'So What?'. Try answering the implications of conflicting standards on business, operations, cost, efforts, maintenance etc. If you have a solid answer to justify the standard's existence in maintaining a high quality data that your consumers demand then keep it. Remember that, more often than ever we only consider the data producer's perspective and not consumers in maintaining the data quality, and this fixated perspective on producers makes your DQ program biased.

While assessing the conflicting standards, it is useful to keep the organization's goals in mind and evaluate how best to achieve those goals. More often than not, the 'best' approach is also the most expensive in terms of resources. Maybe taking baby steps would be more useful, aligned with business needs. Come up with alternatives and let the stakeholders decide.

Nowadays the most used dimensions are the ones included in the DmBok2 of DAMA International, although this framework is functional, there are other more technical and specific dimensions for different compliance and industries.

?? ???????????? ?? ?????????????????????? ???????? ????????????????????: Consolidate data from various sources into a single, centralized repository. This can be a data warehouse, data lake, or a cloud-based data platform. ?? ???????? ?????????????????????? ?????? ??????????????????????????????: Ensure that data from different sources is integrated and synchronized with the centralized repository, so that updates are reflected consistently across the organization. ?? ?????????????? ?? ????????-???????????? ??????????????: Encourage a culture where data quality is everyone’s responsibility, and users are empowered to maintain and improve the quality of data.

The intensive use of rich metadata and new data platforms, many of them based on connectors and api's, allows an exhaustive control of data quality.

It is essential to use policies in accordance with current regulations. The development of a policy is a crucial moment for an organization that wants to be data driven. We are talking about the “rules of the game” of the data lifecycle.

Monitoring data quality is essential and can be approached in two ways with dashboards and sentinels. The use of the latter is more effective because it can trigger direct warnings to data stewards and data owners that something is wrong with their data.

Conflicting data quality standards can lead to significant problems in maintaining accuracy and consistency .. Standards: Ensure that all teams and stakeholders agree on common data quality standards. Hold workshops or meetings to agree on definitions, metrics and processes to avoid discrepancies and misunderstandings. Data ownership: Assign clear data ownership roles to manage and enforce data quality standards. This accountability helps maintain consistency and ensures there is a point of contact for resolving conflicts. Quality checks: Implement automated tools to regularly review data against established standards. Automation allows issues to be identified and resolved quickly, ensuring continued accuracy without manual intervention.

Conflicting data quality standards will weaken the data foundation for any organisation. We need enterprise data models which cater to various requirements before going down the data Governance path. For example, whether we want to follow an UN standard or a UPC standard needs to be clarified first. It also depends on the countries where you operate and how to adhere to the requirements there. There shouldn't be conflicting data Governance requirements but rather exceptions which should be catered to by the overall Governance.

When Priya faced conflicting data quality standards across departments, she knew that alignment was crucial. * She first gathered all stakeholders, facilitating open discussions to understand each team's priorities and challenges. * Priya then established a unified data quality framework, incorporating the most critical aspects from each department's standards. * She implemented regular cross-departmental audits and set up a centralized data governance committee to oversee adherence. By fostering collaboration and setting clear guidelines, Priya ensured that data accuracy and consistency were maintained across the organization.