SLA's form an Integral part of Incident Management and are a standard part of the ITIL Framework
This overview is taken from the viewpoint of an ICT helpdesk
What is an SLA
An SLA is a Service Level Agreement and it is prevalent in many industries in the world today particularly the service industry where response times for support organisations are the backbone of the service that is purchased.
This article explains what a Service Level Agreement is in relation to ICT support and how this can assist you when choosing a provider or if you are a provider how this can define your operational model.
When looking at a Support Desk an SLA usually is made up of the following:
- Time to Answer
- Time To Respond
- Fix Time
- Visit Time
Many organisations set out their target metrics and publish this when a new external/internal customer signs a contract for the service. The organisation will then report on this to their customers at regular intervals as a way of benchmarking their performance.
In more stringent contractual environments SLA's can be used to determine if penalty payments need to be made during periods of service degradation.
SLA's are often found in commercial business contracts but they can also be found in consumer based services. An example of an SLA is the response time offered when your boiler is broken or the time it takes for a company to respond to you when you lodge a complaint.
More mature service organisations should be able to give you a defined SLA that they offer in regards to certain issues. In commercial environments these are generally designed by the impact that this will have on the customer.
A generic SLA for a IT call centre will contain:
Average Time to Answer: 30 seconds
This means the average call answer time throughout the a defined will be 30 seconds. This is sometimes a percentage rather than an average but it is often used to guide the customer how quickly you will respond to telephone calls. I am sure many readers will have been in the situation where they have had to wait a number of minutes for the phone to be answered by a real person. By having this in place the company has to commit to responding to customers in a timely manner.
Time to Respond
Time to Respond is very similar to the Average Time to Answer above. The Time to Respond metric is often used where customers may not be calling in but may use other means like e-mail or web-logging to log faults. The contract will define how quickly the customer will receive a response after logging the call. When reading this in a contract be wary of what a response is determined to be. Some SLA's are written so that a response can be an automated e-mail thus meaning the supplier can always meet the SLA target.
This type of SLA is usually structured by priority on impact to the customer an example is below:
Priority 1 - 4 hours
Priority 2 - 8 hours
Priority 3 - 2 days
Priority 4 - 5 days
Priority 5 - As Agreed
Fix time can also be referred to as Closure Time or Time to Resolve. This is normally the most important of the metric to the customer. If a supplier can define an expected closure time it will immediately set the customer expectations and the service will be transparent. One metric that links to fix time is Reopen times and Recurring issues. If a supplier is likely to breach the Fix Time the more unscrupukous supplier may fix the issue with a "sticky plaster" and not a long term fix. By having Reopen and Recurring SLA's in place this get out can be avoided.
Fix time is normally also prioritsed in the same way as response times.
This is pretty much what it says on the tin. In some services it is important to get someone to site to fix an issue and this metric determines the time for an engineer to visit. The boiler example is one where the customer would need some defined times of when someone will visit the site.
This overview is provided as a quick introduction to SLA's in the service environment. Please check back for further updates on implementating an SLA and SLA design.