Service level

Service level (SL) answers the question "what share of calls did we answer within how many seconds?". The target is usually written as two numbers: the widely used 80/20 target means answering 80 percent of calls within 20 seconds. That is a convention rather than a rule; the right target depends on your industry, what your callers expect and what it costs. Service level should not be confused with a service level agreement (SLA) on uptime, in which a provider commits to how long its infrastructure stays available: one measures how quickly a queue answers, the other how long a system stays up.

How it works

The basic formula is calls answered within the target time ÷ calls that entered the queue and reached an outcome (answered, abandoned, overflowed elsewhere). Choices such as whether to exclude calls abandoned within a few seconds, or how to count overflowed calls, change the result, so the definition should be written down alongside the report. Service level swings during the day, and a daily average can hide poor stretches at peak times.

Why it matters

Service level is the key input to staffing: methods such as Erlang C calculate how many agents are needed to hit a given target from it. A falling service level is often the first sign that the abandonment rate is about to rise. Read it together with average speed of answer and the longest wait.