Should CRM Users Be Measured on Data Quality KPIs?

I would add explicit coverage and defect attribution to the role-based KPIs Hellen described. Each metric should show the eligible population, the assessed population and the origin of the defect. A failed import, an unassessed record and an incorrect value entered by a user require different owners and corrective actions.

I implemented this separation in PolicyWatcher: source retrieval and freshness are tracked separately from assessment coverage, and unknown values remain unassessed. A successful retrieval does not establish the accuracy of the subsequent interpretation. The same principle is useful in CRM: completing a field proves completeness, while accuracy needs a check against reliable evidence. The method is documented at https://policywatcher.online/methodology/confidence

For example, 80 valid records out of 100 assessed gives 80% validity within that subset. If 1,000 records were eligible, coverage is only 10%; the remaining 900 cannot be silently treated as valid or invalid. For opportunity updates, define the eligible stage, review period and responsible role before calculating the KPI. Track integration failures and approved exceptions alongside the result. I would start with team-level diagnosis and sample the underlying records before using the score in individual performance reviews.