Skip to main content

Optimization Thresholds

Optimization Thresholds configures the levers that drive the License Optimization engine. Each threshold maps to one class of recommendation — adjust them to match your organization's tolerance for waste vs noise.

Thresholds​

FieldDefaultDrives
Unused License Threshold (%)10A license is flagged Unused when utilization falls below this percent
Inactive User Days90A user with no activity for this many days is flagged Inactive User
Duplicate Overlap (%)50Two products are flagged Duplicate Tool when usage overlap exceeds this percent
Tier Utilization (%)20A user is flagged for Tier Mismatch (downgrade) when feature usage is below this percent
Redundancy Jaccard Score0.7Two tool catalogs are flagged redundant when their feature-set Jaccard similarity exceeds this score
Growth Anomaly Std Devs2A license-count growth above this many standard deviations from the rolling baseline is flagged anomalous

The Save Thresholds button persists the values; the change applies on the next Run Optimization pass.

Best Practices​

  • Start at defaults and tune per organization — a company with predictable seasonal seat fluctuations may need a higher Growth Anomaly Std Devs (e.g. 3) to suppress quarterly true-up noise.
  • Drop Inactive User Days to 60 for high-cost SaaS (e.g. Salesforce, Snowflake) and keep it at 90 for low-cost desktop tools — most reclaim value comes from the high-cost category.
  • A Duplicate Overlap % of 50 is a starting point; tune up to 70 if the engine produces too many Duplicate Tool recommendations that turn out to be intentional redundancy.
  • Re-run Run Optimization from the Optimization page after any threshold change so the queue reflects the new values.