Quality
Quality data: consistent units

Two values can look comparable while describing different quantities. A measurement in millimetres and one in inches, or results from different test conditions, may enter the same summary without the error becoming obvious. The risk increases when a number moves between spreadsheets and loses its context.
I would check that the unit, method, specification revision and relevant conditions travel with the measurement. Conversion needs a defined, checked method, rather than untraceable manual corrections. Rounding and decimal separators also deserve attention when data comes from different suppliers or systems.
A useful check follows one measurement from collection to the acceptance decision. Where is it transformed? Who recognises an incompatible value? What happens to results already used if an error is discovered? A simple control at entry can prevent extensive rework, provided exceptions have a clear route. Which field in your system actually prevents people from comparing data that is not equivalent?