Service Desk SLA: Setting Response and Resolution Targets

A service level agreement, or SLA, states the level of service a provider commits to deliver, usually including how quickly it will respond to and resolve requests. For an internal service desk, the SLA is an agreement with the rest of the organisation. For an external provider, it is part of a contract, and a miss may carry a penalty.

Three related terms

• SLA: the commitment made to the customer or user, often with consequences.
• SLO: a service level objective, an internal target set to meet the SLA with margin, or a target without a contractual promise.
• OLA: an operational level agreement between internal teams, such as the time the network team will take to respond to a request from the desk.

Response is not resolution

Define two clocks. Response time is how long until a person acknowledges and begins work. Resolution time is how long until the request is fixed. Promising resolution times for work whose duration is unpredictable is a trap, so for many request types the honest commitment is a response time and a target for resolution that is reviewed, not guaranteed.

Say when each clock starts and stops, and what pauses it. A request waiting for information from the user is waiting on them, not on you. State the hours the clock runs, such as business hours in a named time zone.

Set tiers by priority

Tie targets to the priority from the triage matrix. The values below are an illustration of structure, not a recommendation, and each organisation should set its own from its capacity and needs.

• Critical: immediate response, continuous work, frequent updates.
• High: response within the hour during service hours, resolution targeted within the working day.
• Medium: response within a few working hours, resolution within a few working days.
• Low: response within a working day, resolution scheduled.

Base the numbers on evidence

Measure first. Look at how long requests actually take, by type, and set targets you can meet most of the time, tightening as the process improves. A target that is missed constantly is ignored, and one that is never missed may be too loose.

Express targets in Atlas

Atlas does not have a request SLA timer that pauses, resumes and records breaches automatically. For requests handled as tasks, express a target through priority and due date, and review the overdue list as a routine. For incidents, severity levels carry response targets and the platform records time to acknowledge, mitigate and resolve from real instants, counting only the first acknowledgement. If you need contractual SLA timing with pauses and breach reports, use a dedicated ITSM suite.

Report honestly

• Report the percentage of requests that met the target, by priority, with the number of requests behind each percentage.
• Show the worst cases, not only the average.
• Explain misses with causes, and track the fix.
• Do not change the definition of the clock to improve the number.