The Anatomy of a Broken Dispatch Queue: 6 Failure Patterns We See in Every MSP
The dispatch queue is the single most important screen in your MSP. It's where SLAs get hit or missed, where utilization gets won or lost, and where customers form their opinion of you. It's also the most consistently broken artifact in the entire MSP stack.
We've audited dozens of service boards across ConnectWise Manage, Autotask, and Halo. The boards look different. The failure patterns are identical. Six of them, every time.
This isn't a dispatcher discipline problem. It's a structural problem with how PSAs render queues, how tickets get classified, and how decisions are forced onto a human staring at 400 rows of mixed-priority work.
Frequently asked questions
What are the most common reasons MSP dispatch queues break?
Six patterns show up in nearly every MSP service board: the traffic jam (unranked queue), the status drift (manual statuses that don't match reality), the reassignment tax (routing thrash), the priority inversion (per-ticket priority ignores customer tier), the after-hours black hole (off-hours work not captured), and the senior-tech crutch (decision intelligence trapped in one person's head).
Will switching from ConnectWise to HaloPSA or Autotask fix our dispatch problems?
Usually not. All three render the queue as a flat, manually-prioritized list. The field names and color schemes differ, but the architectural problem — list-vs-decision — is identical. Switching PSAs trades known friction for different friction, while costing $50K–$200K and 3–6 months of disruption.
Why does dispatch fall apart when our senior dispatcher takes PTO?
Because the decision logic for triage — customer context, tech specialization, agreement health, churn risk — lives in their head, not in the PSA. The board only shows a list. The senior dispatcher converts it into a decision using context the system doesn't capture. When they're out, that conversion stops happening and the queue collapses into chronological order.
What does an effective dispatch decision layer actually do?
It re-ranks the queue continuously by SLA risk, customer tier, agreement margin, and live tech capacity; derives status from activity instead of manual flags; proposes assignments based on real availability and skill match; captures after-hours work into the morning queue; and encodes institutional knowledge so dispatchers manage exceptions instead of triaging every ticket from scratch.
How many tickets does it take before a dispatch queue starts breaking?
Most MSP queues start showing 2–3 of the six failure patterns once daily open-ticket volume exceeds roughly 150–200, and all six patterns become structural around 400+. The patterns aren't volume-caused — they're volume-revealed. They exist at lower counts; they just don't hurt enough to fix.