What if Service Hours and Support Hours don't match?

This post is another in the series I’ve been writing about Service Level Management. The last post on this topic is here.

My post for today is on a subject that seems to confuse and vex quite a few people –  service times and support times. What do we do if they are different?

For example:

A service desk within a call centre services company provides a sales order processing system that is used 7 days per week from 08:00 until 20:00. The service desk is open from 09:00 until 17:00 Mon-Sat.

Of course, it would be nice if this sort of thing didn’t come up at all – but life has a habit of being messy. Since Service Level Agreements are agreements, the answer is we handle this by agreement with our key customer signatories.

Typically, I’d expect the SLA to state that the service availability is 08:00 until 20:00 Mon-Sun, and I’d also expect the SLA to record the hours of availability of support (Serio Administrators please note, when you set-up and SLA in Serio that the start/end pairs refer to service availability, not necessarily support availability hours). This means our statistics on availability (something I’d expect to see in the SLA, as per my last post) would be produced on the longer seven days figure. If you are tempted to do otherwise remember that Incidents are probably still affecting the business and the weekends – you are just airbrushing it out of the statistics.

The objection to this usually comes from the service desk itself – Incidents are going to occur when there is no service devoted to their resolution, and will be impacted as a result. The answer is that you factor this into the SLA you negotiate your targets – which might have lower availability service levels specified than might otherwise be the case.

One of the benefits of this kind of discussion is there is a formal focus on IT service for the enterprise. One of the things that might be considered is extending service desk hours and providing a costed proposal for this,  or examining the costs of having a suitable support engineer ‘on call’ for the extended periods. Try to compare the costs of unavailability of our example sales order processing system with that of the costs of an extended support service (see our availability white paper for more info on computing the costs of unavailability).