The "2% of our users" is indeed misleading since most users don't run organizations, and it's mainly orgs that were affected. A better metric would be what % of orgs were impacted.