In this guide
A useful SLA describes, at a minimum, coverage hours, contact channels, priority levels, initial response time, the restoration target and process, escalation, on-site terms, third-party dependencies and measurement. Response time and resolution time are different.
Response time ≠ resolution time
Response time is the period until a request is acknowledged and taken on according to its priority. Resolution time is different: it may depend on diagnosis, a replacement part, access to a third-party vendor or approval of a change.
An SLA that promises only a “solution within X hours” without describing exceptions and dependencies usually creates unrealistic expectations.
Priority should reflect business impact
One user unable to print and a company that has lost ERP access are different incidents. Priority should consider how many users are affected, which service has stopped and whether a temporary workaround exists.
This helps the Helpdesk work to shared criteria and prevents every request from being labelled “urgent”.
What the agreement should specify
- Coverage hours and days, including the process for out-of-hours incidents.
- Submission channels: portal, email and phone, and when each is used.
- Definitions of P1/P2/P3/P4 or equivalent levels, with examples.
- Response targets and the precise meaning of “response”.
- Escalation: when the issue is passed to a senior engineer, vendor or management.
- On-site support: when it is activated, the areas covered and how visits are scheduled.
- Exceptions and dependencies involving ISPs, hardware vendors, Microsoft or other third parties.
- Reporting: open tickets, recurring incidents, trends and proposed improvements.
The SLA must match the actual scope
If the MSP does not manage an application, it cannot commit to how long its developer will take to fix a bug. It can commit to when it will take on the issue, how it will diagnose it, when it will escalate and how it will provide updates.
A clear distinction between managed scope and third parties is essential to a mature service relationship.
Frequently asked questions
What is a P1 incident?
It usually means a serious incident with major business impact, but the precise definition must be in the contract. There is no universal definition that applies to everyone.
Does an SLA guarantee that every problem will be resolved within a specific time?
Not necessarily. It should distinguish commitments within the provider’s control from third-party dependencies and the time required for the technical work itself.
Does a small business need an SLA?
Yes. Even a simple framework clarifies when and how support is provided, especially when IT services are critical to daily operations.
Related services and guides
Discuss your next step with UniLoud
Discuss which services are critical and the hours they require with us. We will define a scope and SLA that reflect the business impact.
