Benefit Quantification for Improving Risk Numbers

Hi all
who has a solid practical approach to the quantification of improving risk numbers by improving the underlying data quality? Some cases: Case 1:  improving instrument to issuer mapping, I have a have better transparency where ratings are used to calculate potential losses. Case 2: improving rating coverage for the instruments traded will lead to a better risk representation Case 3: You improve the sector definition and your risk model will act more preciously. If any of this is used for capital allocation, the business impact can be quantified. I am rather interested in those items where you just make a better decision; you know your data and your risk better.