When data quality is worth the effort

It is rarely the thing that gets a project approved, and often the thing that decides how it goes. Data quality is not free, and pretending otherwise leads to bad decisions.

Data outlives the applications built on top of it, which is why the model deserves more thought than the screens. It rarely shows up as a line item, which is exactly why it slips.

When it is worth it

Bad data spreads faster than anyone corrects it. That sounds obvious written down. It is still the thing most often skipped. Write the reasoning down alongside the decision, because the reasoning is what changes first.

When it is not

If nothing downstream depends on it and nobody is complaining, it can wait. There is a version of this that is over-engineered, and it is worth avoiding.

How to decide

Someone must own each dataset or nobody does. In practice this is a scheduling problem more than a technical one. Budget a little time for it every quarter and it never becomes a project of its own.

In practice

Numbers get quoted in meetings long after anyone remembers how they were calculated. Three things worth confirming about data quality before you move on:

• Someone can say what the current setup is without going to look
• Automated checks catch drift that eyeballs miss — and you know whether that is true here
• There is a way to tell whether the last change to this helped

If any of that sounds like a description of your current setup, it is fixable.