Which repeats deserve a problem record

Repeats are a signal, not an automatic work order. I open a problem record when investigating the cause is likely to reduce future impact, not because the counter ticked over an arbitrary number.

Why it matters

If every duplicate incident becomes a problem, the backlog fills with noise and real systemic issues wait. If almost nothing becomes a problem, the organisation keeps paying the same restoration cost forever.

Problem management needs a filter. Volume matters. So do severity, trend direction, and whether a durable fix or workaround is realistic.

Criteria I use

I look for one or more of these:

1. The same failure has returned often enough that restoration cost is becoming routine debt.
2. A single occurrence had major or widespread impact and the cause is still unclear or unfixed.
3. A workaround exists but is fragile, manual, or unsafe to keep forever.
4. Monitoring or trend review shows a rising pattern before users escalate loudly.
5. A supplier, change, or known weak component is implicated and needs structured follow-up.

I also ask whether someone can own the investigation. A problem without an owner is a parking space.

What good looks like

Incident to problem handover states the symptoms, frequency, impact, and what already restored service.

Low-noise repeats with a solid known error stay managed through that known error until a fix is planned. They do not spawn duplicate problem records.

Major incidents usually create or link a problem unless the cause was already known and the fix path is clear.

The backlog is reviewed. Stale problems are closed, merged, or re-scoped instead of ageing quietly.

Common mistakes

Opening a problem for every incident “just in case.” That hides priorities.

Waiting for a perfect count threshold while a high-impact pattern grows.

Investigating without linking related incidents, so nobody can see the real volume.

Closing problems when the tickets stop, even though the weakness is still in the service.

Takeaway

Spend investigation effort where repeats, risk, and the chance of a lasting fix meet. A smaller, honest problem backlog beats a large decorative one.