Service desk best practices for midsize organisations

Staff judge IT almost entirely by the service desk. With 50 to 500 employees, you are too big for ad hoc support and too small for textbook enterprise ITIL. These practices work at that size.

The ticketing system

A ticketing system needs these capabilities.

• Email that turns into tickets, so users learn no new tool.
• SLA tracking that escalates as deadlines near.
• Asset data, so technicians see each user’s devices and software.
• A built-in knowledge base for technicians and users.
• Reports that need no data scientist.
• Automatic routing, priority and status updates.

Platform	Fit at midsize
Freshservice	A good balance of structure and ease, at a fair price
Jira Service Management	Strong if you already use Atlassian, heavy for a simple help desk
HaloPSA or Halo ITSM	Popular with MSPs and IT teams, with good automation
ConnectWise or Autotask	Common at MSPs, so learn what yours reports

Avoid a shared support inbox with no ticketing, where nothing is tracked. Avoid ServiceNow or BMC Remedy too, which need a dedicated admin and six months of setup.

SLAs that matter

Response time is how fast IT picks up a ticket, which reassures the user. Resolution time is how long the fix takes, which measures effectiveness.

Priority	Meaning	Response	Resolution target
P1, critical	Work stopped for several users or a critical system	15 minutes	4 hours
P2, high	Serious impact on a person or department, no workaround	30 minutes	8 business hours
P3, medium	Moderate impact, with a workaround	2 hours	24 business hours
P4, low	A minor issue, request or question	4 hours	48 business hours

• Treat resolution targets as goals. Punishing honest complexity discourages careful fixes.
• Define business hours. Most midsize firms use business hours for P2 to P4 and 24/7 response for P1 alone.
• Report SLA adherence monthly. Constant misses mean the targets are wrong or the team is short.

Escalation

Without clear escalation, tickets bounce and users repeat themselves.

• Functional escalation moves a ticket to deeper skills, such as a network engineer. It should be routine.
• Hierarchical escalation brings in managers when a ticket stalls, an SLA nears breach or a user is unhappy. Trigger it by SLA timer, and allow it by hand.
• External escalation goes to vendors. Keep their contacts, contract details and case steps ready.

1. Send every P1 to a senior technician or team lead at once.
2. Escalate any ticket near its SLA deadline, and tell the team lead.
3. Write escalation paths for common issue types.
4. Require notes on what was tried, ruled out and suspected. “I cannot fix it” is not enough.
5. Escalate to management any ticket reassigned three or more times.

Self-service

A good self-service portal cuts ticket volume by 20% to 40% and fixes common requests faster.

• Password resets and unlocks. These are the biggest single task. Azure AD self-service password reset or Okta removes 15% to 25% of tickets.
• Standard requests. Software, hardware, access and onboarding forms with required fields get IT complete details first time.
• Ticket status. Users check progress without calling.
• Fix-it guides. Clearing the browser cache, rejoining Wi-Fi, restarting the VPN and updating software.

Pin the portal in Teams or Slack and every browser, and keep it simpler than sending an email. Automate standard requests, such as installs through Intune, so a day-long ticket becomes five minutes. Always keep a phone number and email for people who prefer a person.

A knowledge base people use

Knowledge bases fail when nobody maintains them. Outdated articles destroy trust.

For technicians, keep procedures for new users, offboarding, printers and VPN. Add known issues, architecture documents, where credentials live, and incident runbooks. When a fix needed research, the technician writes or updates an article before closing the ticket.

For users, keep how-to guides for Wi-Fi, phone email and VPN. Add the questions the help desk hears most and plain policy summaries.

Each month, check the top 20 ticket categories. Write an article for any category without one, and improve any article that still draws tickets.

Metrics

Metric	Healthy level
Mean time to resolution, by priority	Falling over months, judged by trend
First-contact resolution	65% to 75%, with under 50% a warning
SLA adherence	90% or more for P2 to P4, 95% or more for P1
Satisfaction survey, scored 1 to 5	Over 20% respond, and the comments matter most

Also watch ticket volume by category, hour and day, which shapes staffing and shows recurring causes. Watch reopened tickets, which mean incomplete fixes. Track self-service adoption and the age of the backlog.

Do not obsess over total ticket count, since more tickets can mean more trust. Do not chase average handle time either, since rushing raises reopens.

Staffing

A common benchmark is one help desk technician per 70 to 100 users. That assumes business hours, a stable estate, self-service and a knowledge base. With longer hours, a complex estate or no self-service, plan one per 50 to 70 users.

Tier	Handles	Staffed by
Tier 1, help desk	Passwords, basic fixes, triage and common requests	Junior to mid-level technicians
Tier 2, systems	Escalations, servers, applications and networks	Experienced technicians and admins
Tier 3, engineering	Complex infrastructure, security incidents, projects and planning	Senior engineers and architects
Service desk manager	Operations, escalations, vendors, metrics and improvement	Needed above 75 to 100 employees

Midsize firms often outsource Tier 1, and sometimes Tier 2, to a managed IT provider. They keep Tier 3 in house for knowledge and strategy. That gives affordable coverage, 24/7 monitoring and a wider team.

Improving each quarter

1. Review the metrics against targets and past quarters.
2. Fix the root causes of the top five ticket categories.
3. Audit the knowledge base.
4. Survey users and act on what they say.
5. Look for new tools and automation.
6. Adjust SLAs and staffing to business changes.

Service desks run as improving operations consistently beat those run as costs to cut.

Questions people ask

What is a good first-contact resolution rate?

65% to 75% for a midsize firm, and 80% or more at the best. Below 50%, improve Tier 1 training, the knowledge base and ticket routing.

How do we measure service desk performance?

Track mean time to resolution by priority, first-contact resolution, SLA adherence and satisfaction. Together they show speed, quality, reliability and experience. Review them monthly, and watch trends.

Should a midsize company use ITIL?

Use its principles, not the whole framework. Light versions of incident, change, request and knowledge management cover about 80% of what you need.

How many IT staff does 100 people need?

One or two Tier 1 technicians, one Tier 2 administrator and part-time access to a senior engineer. That is about 2.5 to 3.5 full-time staff. Or outsource Tiers 1 and 2 and keep one IT manager.

Which ticketing system is best?

Freshservice, Jira Service Management and HaloPSA rank well at midsize. The best is the one your team will use. Look for ease, SLA tracking, a knowledge base and reports, and skip ServiceNow without a dedicated admin.