Creating a Problem
Not every incident becomes a problem. Criteria: repeated pattern, high-impact one-offs, unresolved incident. Without criteria, problem count explodes and RCA loses focus.
RCA Frameworks
Pick one: 5 Whys, Ishikawa, Kepner-Tregoe. Don’t switch mid-investigation. Record the RCA on the problem record with the rationale at each step.
Known Errors
A known error is a problem with documented workaround and pending fix. Publish known errors to the KB so service desk can resolve repeated tickets without re-investigating.
Linking Incidents
Every new incident should search for a matching problem. Configure the incident form to suggest matches. Resolved via problem — means the fix is in the pipeline, not unresolved.
Measuring Value
Problem management’s ROI is prevention. Track: incidents avoided (linked to a closed problem’s fix), mean time to known error, open problem age. Publish quarterly.