You need to distinguish the amount of unresolved work from the age, risk, and recoverability of that work.
The team separates new arrivals, old exceptions, blocked work, and SLA-risk cases instead of managing one blended count.
Celebrating a smaller backlog after closing, merging, or rushing work that returns as repeat demand.
Backlog(t) = Backlog(t-1) + new tickets - resolved ticketsBacklog is the standing pile of open work — everything created but not yet resolved. A little backlog is healthy buffer; a growing one means arrivals are outpacing resolutions, which is a capacity or process problem no amount of hustle fixes on its own.
The useful lens is flow: backlog grows whenever inflow exceeds throughput over a period.
What it hides: a single backlog count flattens urgency. Ten tickets an hour old and ten tickets three weeks old are not the same emergency, yet they sum identically. Track backlog by age and priority, and watch the trend (growing, flat, shrinking) more than the raw number.