Solution · Problem management
Fix the cause once, instead of fixing the symptom every week.
An incident is a fire; a problem is the faulty wiring that keeps starting them. Teams that only fight fires get very good at firefighting — and never find the time to fix the wiring.
Repeat incidents grouped under one problem, with one owner.
Root causes found with evidence, not guessed under pressure.
Fixes that stay fixed, because the cause was addressed.
The problem
Why the same incidents keep returning
Problem management is the practice teams postpone most, because its payoff is invisible: it is the incidents that stop happening. The cost of postponing it is visible every week.
Every incident treated as new
Ten incidents with the same cause are logged as ten separate incidents, each fixed from scratch. Nobody links them, so nobody sees the pattern — and the cause is never investigated.
Restarted, not resolved
The service is restarted and the ticket is closed, but the reason it failed is never asked. Restarting is fast and feels like progress; it is also a promise that the same failure will return.
No time to investigate
The queue is always full, so investigation is always tomorrow’s work. Tomorrow never comes, because the repeat incidents keep filling the queue that investigation would have emptied.
The fix is never shared
One person works out the cause and keeps it in their head. When they are away, the team starts from zero again — and the knowledge leaves with them when they move on.
The process
Five steps, and the itserva page for each
Five steps, from the pattern to the permanent fix.
- 01
Spot the pattern
Look at how often the same service fails. A short time between failures is the pattern announcing itself — the incidents are not bad luck, they are one problem repeating.
- 02
Log it as a problem
Open a problem record and link the incidents to it. The record gives the pattern a name, an owner, and a place to collect evidence — instead of ten tickets that never mention each other.
- 03
Find the root cause
Ask why, five times, and follow the evidence at each step. The first answer is usually a symptom; the cause sits further down, where the fix is cheaper and lasts longer.
- 04
Fix it for good
Turn the cause into actions with owners and dates: the patch, the monitor, the documented step, the replacement. A fix without an owner is a suggestion, and suggestions do not stop incidents.
- 05
Share the answer
Write the cause and the fix down where the next person will find it. The answer is only useful if the person on the next shift can use it without calling the person who found it.
Your team knowledge base
What’s included
Tools, templates and guides for problem management
Template
Root Cause Analysis Template
Work through the five whys and the contributing factors behind an incident.
Product
ITSM Toolkit
Eight free calculators for SLA uptime, downtime cost, MTTR, MTBF, staffing and more.
Measure it
The four numbers worth tracking
Four numbers show whether problems are being found and fixed, or only endured. Track them by service, and review them with the incident numbers.
Repeat incidents
Incidents with a cause you have seen before. This is the headline number for problem management: when it falls, the practice is working.
Open problems
Problems found but not yet fixed, and how long each has been open. A problem that sits open for months is a decision waiting to be made, not a mystery.
Time to root cause
How long it takes, from the first incident, to name the cause. Shorter means the team spots patterns earlier — and stops the repeats sooner.
Problems solved per quarter
Problems closed with a real fix in the period. It is the honest measure of progress: not how many problems were found, but how many stopped causing incidents.
Frequently asked questions
What is the difference between an incident and a problem?
An incident is an interruption to a service. A problem is the underlying cause of one or more incidents. Fixing the incident restores the service; fixing the problem stops the incidents from happening again.
When should we open a problem record?
When the same cause produces incidents more than once, or when a single serious incident clearly has a cause worth finding. One repeat is a hint; two is a pattern.
What is the five whys technique?
Ask why five times, following each answer. A server failed because a disk filled up; the disk filled up because logs were never rotated; the rotation was never set up because no one owned it. The fix sits at the end of the chain.
Do we need a separate problem process?
Not a heavy one. A problem record, an owner, and a regular look at repeat incidents is enough to start. The process grows when the number of problems justifies it.
Pick the incident that keeps coming back.
Open the root cause template and work through the five whys on the one incident your team has fixed most often. A free workspace saves the analysis, so the fix and its owner are on record.