Why your data quality issues keep coming back

Organisations have a data quality log. They may have a dashboard for tracks issues. However many see the same issues closed one quarter and reopened the next. When this happens, the usual response is more effort: another clean-up exercise, another tool or another working group. The log gets shorter for a while, then grows again. The problem isn’t effort but has the root cause been address and tested and agreed with those impacted, it is addressed.

Three causes behind most repeat issues

  1. Owners have the title but not the time. Many organisations name a data owner for every domain. Some put responsibility into anyone’s objectives, or give owners the authority to change the processes that create the errors. An owner without time or authority can’t stop an issue recurring. All they can do is report it.

  2. “Closed” means fixed once, not fixed for good. An issue is often marked closed when the bad records are corrected. But correcting records treats the symptom. Unless the root cause is addressed and a check is in place to catch it next time, the same data will drift back. Closure needs evidence: what was fixed, why it happened, and how you’ll know if it happens again.

  3. Nobody agrees what the number should be. If Finance, Operations and Commercial each define a key measure differently, every report is “wrong” to someone. Teams then fix data to match their own definition, which breaks it for everyone else. Until the critical measures have one agreed definition and one accountable owner, quality work goes in circles.

Why this matters more now

Regulators and boards expect numbers they can rely on. In water and energy, regulatory returns need clear assurance. In financial services, risk data aggregation rules set out expectations on accuracy and lineage. Across all sectors, AI raises the stakes: a model trained on data with unresolved quality problems repeats them at scale, and more convincingly.

Organisations investing in ERP, reporting or AI on top of reopening issues are building on a weak foundation.

What to fix first

You don’t need a large programme to start. Three steps make a difference within a quarter:

Pick five to ten critical measures, the ones the board, regulator or customers rely on, and agree one definition for each. Give each one an owner with time and authority, written into their objectives, not just a name on a slide. Change what “closed” means. Require evidence of the root cause and a monitoring check before an issue can be closed.

Then watch which issues come back. Those are the ones that point to the real problem.

Find out what’s driving your repeat issues

Every organisation’s mix of causes is different. Let’s explore how we can help. Book a consultation today.

Learn more about our services on the What We Do page.

Next
Next

A Holistic Approach to Data, Data Science, Machine Learning, and AI: Strategies for Success