"The data is wrong" is the phrase that kills trust in a report. Once someone finds one bad number, they start doubting all of them. Data quality is not a technical detail — it is the foundation of any data-driven decision. And it is measured in six concrete dimensions.
Why quality is everything
The best dashboard over bad data is worse than no dashboard: it gives false confidence. Ensuring quality before analyzing is not perfectionism — it is the difference between deciding based on reality or on well-presented fiction.
The six dimensions of quality
- Accuracy: does the data reflect reality? A correct address, a correct value.
- Completeness: no essential fields or records are missing.
- Consistency: the same data matches across systems (total sales is the same everywhere).
- Timeliness: the data is recent enough for the decision it supports.
- Uniqueness: no duplicates inflating counts (the same customer only once).
- Validity: the data respects expected rules (a date is a date, an email has email format).
Measure to improve
You cannot manage what you do not measure. Defining simple rules per dimension — "0% invalid emails", "under 1% duplicates" — and monitoring them makes quality visible. When an indicator worsens, you attack the cause before it contaminates reports.
Quality starts at the source
Fixing data at the end of the pipeline is patching; the ideal is preventing the error at entry — form validations, rules in the source system, clear owners for each dataset. The earlier a problem is caught, the cheaper it is to fix.
It is also a matter of culture
Data quality is not only technology: it is everyone understanding that the data they enter feeds decisions. When each person treats data as a shared asset, quality rises naturally and sustainably.
In practice
Pick the dataset that most feeds your decisions and assess it against the six dimensions. Where does it fail? Start there. Trustworthy data is earned, not assumed. When did you last check the quality of the data behind your reports?
