Waiting for Clean Data
TBM Therapy - Issue #1
The Question: How do you handle bad data quality?
What’s Going On: I know our IT and Finance data isn't very good. There are several systems to find apps, infrastructure and cost in IT and often they don't align with Finance's files. It seems like a lot of work is needed to go clean up the data before we can even start.
What I’m Afraid Of: I'm worried our leaders won't support any effort unless the data is +95% validated
Every company says its data is bad. In my experience it's rarely as bad as people think. The most common problem is that nobody with authority is ever forced to see it. Everyone builds their own offline mappings and workarounds, and those workarounds hide the low quality from the systems where it needs to get fixed. Meanwhile cleanup is expensive, slow, and never wins against whatever else is on the roadmap. It gets funded when the mess appears in something leadership actually needs.
So don't wait. Build the view your leadership wants and put a validation grade in the corner. Make them ask why it's only 60% validated, and have a specific plan ready for getting to 80%. Don’t boil the ocean. Just the data you need cleaned for information they can act on.
Two things worth considering:
If this is leadership’s first real look at IT cost, make the grade more prominent than feels comfortable, because whatever number you show becomes the baseline you get measured against next cycle.
Review the report with data owners before it’s shared with leadership. A data owner who finds out you scored their data poorly on a slide to the CIO can become an obstacle long-term.
Data quality is a technical problem with a relationship solution: find the root cause in the source system, then go work with the person who owns that system. As they say in therapy, “you heal what you reveal”, and people only fix what they've been made to look at. So the grade tactic is less of a disclaimer and more of a mechanism for change.
Have a question of your own? Submit here.
Want next week's answer in your inbox? Subscribe here.