Suppose an analytics dataset has one row per workspace, but a customer organization can own several workspaces. Counting distinct workspace IDs overstates the number of customer organizations, while grouping by company name can merge unrelated businesses with similar names.
What identifier and mapping approach would you use for an organization-level customer metric? How would you treat workspace transfers, mergers, and records that have not yet been matched? I want a method that is understandable to report readers and can be checked against the source systems.