SLA stands for, of course, Service Level Agreement. You could say at its most generic, an SLA is almost like a promise that a service provider makes to a customer about the quality of service they will provide. Contact center SLAs can be external, for example in a contract with a customer for an Outsourcer (aka BPO), or internal, to measure, document and communicate your contact center’s performance for stakeholders within your organization, or to compare with other contact centers. Basically, the SLA helps you manage customer expectations, both external and internal.
How is a Service Level Calculated?
The metric called Service Level that is the basis of the Service Level Agreement within a contact center operation, is based on the assumption that my contact center agents will be able to (in the case of service calls, for example) answer 85% of the incoming calls within 20 seconds. As long as that is the case, I am “within SLA”. If my agents cannot do that (for example because there’s a higher volume of incoming calls than I expected), then I’m “outside of SLA” and have broken the threshold. Here’s the formula:
Service Level = [In SLA] / ([In SLA] + [Out SLA])
As mentioned above, this formula uses the following definitions for the data used:
- In SLA - The number of contacts the agents handled within the defined service level threshold.
- Out SLA - The number of contacts agents handled outside the specified service level threshold.
There are some decisions you need to make to get consistent, usable data, which are the following:
- What is your Service Level target?
- How do you treat Abandoned Calls?
- How do you define Time Intervals?
Let’s take a quick look at each of those.
- What is your Service Level Target? As mentioned earlier, you will need to define which Service Level you will want to manage against. This decision depends on many different factors, to name just a few: Do you feel you want to be on par with the competition, or do you want to do better? How do different media impact the service level? How long are customers willing to wait for an agent – that may be 20 seconds for a phone call, but 60 seconds for a chat, and an hour for an email? What volume of interactions do you expect? How many agents do you have? Can you hire? Can you invest?
- How do you treat Abandoned Calls? There are a number of settings concerning abandoned calls that you need to make decisions on, as they will impact your Service Level. First, you will need to determine your Short Abandon Threshold. For a phone call, an abandoned call is a call that was ended before it reached an agent. The Short Abandon Threshold is a value that allows you to define, in seconds, whether or not an abandoned call is considered a Short Abandon. For example, if your Short Abandon Threshold is 10 seconds, then a call where the caller hung up after 7 seconds in queue would be reported on as a Short Abandon. You may want to exclude those from your service level calculation – such a call may have been someone dialing the wrong number, in which case you would not want it to impact your Service Level.
- How do you define Time Intervals? With CXone, you have two options of defining the way how contacts are reported on with regards to Time Intervals – the most important guideline here is to be consistent within a calculation. The two available options you can choose from are True to Interval, and Contact Start. The difference between the two is the following: when you collect data based on the True to Interval data model, then a contact that spans more than one reporting interval (e.g. 15 minutes), will be counted in all intervals it is active in. Contact Start, however, would only report this interaction in the interval it started in, which may lead to lower overall numbers.
So why do you care?
The most important benefit you gain from measuring Service Level is that you can report on how many contacts were handled in your contact center as quickly as you’d want. Repeat regularly to understand how things evolve in your operation. However, keep in mind that while speed is an important factor, the Service Level will not provide insight into the quality of interactions. It does not enable you to understand, whether customer issues were resolved, or follow-up interactions were needed. Definitely use Service Level - as one of your KPIs
