Ticket Management Process

The one rule: every channel terminates in a ticket

If it's not in a ticket, it didn't happen. Every intake channel – email, phone, portal, walk-up, a text to a tech's cell – must end in a ticket in your PSA. The standard intake set is email-to-ticket, a client portal, and phone. Progressive MSPs push portal or agent-tray submission because it captures triage-quality data (user, machine, category) at intake – but never remove email and phone; you meet clients where they are and convert the request into a ticket yourself.

Everything else in this process – SLA measurement, profitability analysis, staffing decisions – is built on this rule. A ticketing system that captures 80% of the work reports on a business that doesn't exist.

Roles

• Dispatcher / service coordinator – once you're past roughly 3–4 technicians, appoint one. The dispatcher owns priority assignment, scheduling, and the SLA clocks, so engineers execute instead of self-selecting the easy tickets from the queue. Two working models: a dedicated dispatcher, or a rotating intake/triage tech. Before that size, the founder or service lead owns triage at fixed times each day.
• Engineers (T1–T3) – work what's assigned, log time as they go, escalate on the time-box (below).

The ticket lifecycle

1. Intake. The request arrives via any channel and becomes a ticket before work starts.
2. Triage. The dispatcher sets priority by impact × urgency per your published SLA matrix, and sets type/subtype from a standardized taxonomy. Inconsistent categorization makes ticket-count and trend reports useless – the taxonomy is small, fixed, and enforced.
3. Dispatch. Assign by tier and schedule. The dispatcher, not the engineer, owns the SLA clock.
4. Work. Notes and time entries go in as the work happens. Known issues should have a runbook linked to the ticket type, per your documentation standards.
5. Wait honestly. "Waiting on Customer" and "Waiting on Vendor" pause the SLA clock – essential for honest metrics, and the difference between measuring yourself and measuring your clients' reply speed.
6. Resolve, then close. Resolved means the fix is delivered with a resolution note; Closed follows after client confirmation or a defined aging window.

Keep statuses minimal:

Every extra status you invent is a place for tickets to hide.

Escalation tiers

Define escalation criteria by time-boxing, not by feel: for example, T1 escalates after 30 minutes stuck. Time-boxes stop ticket-hogging – the tech who sits on a ticket for half a day rather than admit they're stuck. Escalating on the time-box is following the process, not failing it, and it's the dispatcher's job to enforce that framing.

Time entry: every minute logged

Time logged against tickets is the only way to know cost-per-ticket, per-client profitability, and true agreement margins – it is literally how you answer "which client is eating my payroll?" The consensus antipattern across the industry is letting untracked work become the norm: a one-minute favor occasionally is fine, but systemic untracked work makes gross-margin analysis fiction.

Enter time in real time, as the work happens – not reconstructed on Friday afternoon, when memory invents a flattering week.

Kill the shoulder tap

Undocumented walk-up and DM work destroys metrics, hides toxic clients, causes double work (the next tech has no history), and burns out your best technicians – the helpful ones absorb an invisible workload nobody staffs for.

The standard script: help happily, but first – "let me open a ticket so this gets tracked." Clients train to the process within weeks. Apply the same rule internally: founders are the worst shoulder-tap offenders in most shops.

Daily queue hygiene

• Morning (dispatcher): triage everything new, assign owners, work the queue SLA-risk first.
• During the day: confirm P1/P2 tickets are actually moving; chase stale "Waiting on Customer" items.
• End of day: no untriaged tickets, no unassigned tickets, every tech's time entries complete.
• Weekly: review aged tickets oldest-first, close out confirmed resolutions, spot-check taxonomy consistency.

Metrics to watch

• Tickets per endpoint per month – mature MSPs target below 1. Sustained high noise signals misconfigured environments, weak onboarding, or wrong-fit clients – not a need for more techs.
• First response time – commonly targeted under one hour during business hours as of 2026; report attainment against SLA per priority, never a single blended average.
• SLA attainment – honest only if waiting states pause the clock.
• CSAT and NPS – Acronis PSA can send customer surveys and report NPS through its service desk; Simplesat is a dedicated option for per-ticket CSAT. Track the metrics separately because they measure different questions. ScalePad's 2025 trends research links high CSAT with retention.
• Time logged vs. hours worked – the discipline metric that makes every other metric trustworthy.

Benchmarks and targets for these live in MSP KPIs and benchmarks.

Exit criteria

You know the process is working when:

• No work reaches a technician without a ticket, from any channel, including the founder's.
• Every ticket has a priority, type, and owner by the end of morning triage.
• Escalations happen on the time-box, not after a day of quiet struggling.
• Time entries are same-day, and your agreement-margin reports are believable enough to price from.
• Queue reviews are boring – which is exactly what you want.