Struggling with inconsistent data definitions in your projects?

I agree that defining definitions is a key need. However, as I spoke about recently at an event in London. small companies don't start as they think they are too small. Large companies see data activities as an insurmountable task. meaning no one starts early enough. There is no better time to start than now! Please don't go for an all-in let's create all our term definitions. It takes time to refine the definition and taking in all relevant parties' views to agree is not a quick task. Start small. Determine a process, who will be involved and communicate this widely. Then choose a key process or dashboard to define the key data definitions for. Set metrics. Report on progress. You are on your way, keep going!

Data Governance begins with establishing a company-wide unified data dictionary. This creates a common language and eliminates ambiguities that arise from varying definitions within the organization.

Centralizing control over data definitions is essential for consistency, but it's only one aspect of effective Data Governance. While having a dedicated team or Data Governance Officer oversee the data dictionary is important, success also requires a comprehensive strategy that includes clear policies, continuous monitoring, and regular audits. Moreover, executive management must actively support these initiatives to ensure their success. Their commitment underscores the importance of Data Governance and drives the cultural change needed for its effective implementation across the organization.

I am not sure that audits are the right way to go, implementing regular review cycles are certainly important. Understanding if a defined term has remained steady or there is some sort of update required needs to be established in regular intervals.

Often, we can incorporate this into any onboard process, and even offboarding process. Meanwhile, different means of training (i.e. 1-to-1, in group, online, offline) will raise the implementation efficiency along the DG program.

In my experience building a machine learning product, we ran into bottlenecks due to inconsistent data definitions. This inconsistency made it challenging to optimize the organization's data for model development and insights. Implementing comprehensive training programs proved to be an effective strategy to address this issue. By educating employees on standardized data definitions and governance policies, we ensured a consistent understanding and application across the organization. These training programs aligned teams with best practices, reduced errors, and improved data quality. Regular workshops, tutorials, and hands-on sessions empowered employees to manage and utilize data correctly, fostering a culture of accuracy and consistency.

The technology solutions will reflect how the organization will structure itself to manage their data from policies in several aspect like collection, retention, usage or deletion: evaluate how the current IT and data infrastructure give the control to build the data governance policies.