Effective Data Governance: Focus on Impact

While data governance metrics like data accuracy are crucial for maintaining order, the question isn't just, "Are the data governance metrics robust?" but more critically, "Is the data useful?" Data should be actionable, drive decisions, and lead to tangible outcomes. It’s the usefulness of data that turns it into a strategic asset.

Ironically, as long as your data is consistent, in some cases, it doesn't actually matter if it’s accurate. For example, imagine a dataset where the recorded gender of customers is consistently incorrect, 100% wrong but always in the same way. If you use this data to create a traditional marketing model, as long as the gender inaccuracies are consistent, your model should still effectively identify high-potential customer segments. The model's performance hinges on the consistency of the data it was trained on and the data it scores, not the accuracy per se (I’m not suggesting this as a best practice, of course, since your customer profiles would still be off).

Moreover, if you have multiple models in operation and discover a data error, you should not just correct it without proper change management procedures. Sudden corrections without careful planning can negatively impact model performance and disrupt operations.

Prioritize delivering and maintaining data that drives decisions and generates tangible outcomes.