Definition
Service level is expressed as a pair: a percentage and a threshold. An 80/20 target means 80% of contacts are answered within 20 seconds. Unlike ASA it is threshold-based, so it constrains the distribution of waits rather than just the mean.
Erlang C derives it from the probability of waiting: SL = 1 − P(wait) × e^(−(agents − erlangs) × threshold ÷ AHT). It is highly non-linear, near the point where agents barely cover workload, one extra agent can move service level by twenty points, while at comfortable staffing an extra agent adds a fraction of a point.
Service level should be measured and managed at interval level. A day that hits 80% overall can easily contain two intervals at 40%, and those intervals are where complaints come from.
Why it matters
- It is the most common contractual and SLA metric in outsourced contact centres, so money often depends on it directly.
- Its non-linearity explains why very high targets are disproportionately expensive and why brief understaffing causes an outsized customer impact.
- Judging service level alongside occupancy prevents the trap of hitting the target with no resilience to forecast error.
Get new WFM tools first
One short email when a new calculator, template or article goes live. No spam.