Contact centre service level management
Not all service level failures have the same cause — and they do not have the same fix. Diagnosing the cause before acting is the difference between solving the problem and making it worse.
Five causes of SL failure and their distinct responses
Volume above forecast
Diagnosis
Current call arrival rate is above forecast for the interval. WAPE for this interval shows actual volume 15%+ above plan.
Intraday response
Deploy flexible resource (held breaks, suspended training, back-office agents with voice skills). Open overflow routing if available. Alert team leaders.
Medium-term fix
Investigate whether the volume spike was a random event or a forecast model error. If systematic, update the forecasting model assumption (see forecasting review guide).
Below-target adherence
Diagnosis
Volume is at or below forecast but occupancy is low and agents in available state is below scheduled requirement. RTA system shows agents in non-adherent states.
Intraday response
Intraday management intervention: TL contact for agents in extended breaks or wrap. Check for system-driven states (logged out due to IT issue, system freeze). Do not instruct agents to reduce AHT to compensate.
Medium-term fix
Review adherence management — if below-target adherence is systemic (not isolated), investigate root causes: are agents consistently overrunning breaks? Are there unmanaged long calls at shift end? Is occupancy too high, causing extended recovery?
AHT above assumption
Diagnosis
Volume is at or below forecast, adherence is normal, but agents are handling fewer contacts per hour than planned because average handle time has increased.
Intraday response
Investigate root cause immediately. Common causes: new product or promotion generating longer queries; system slowness (check IT for known issues); specific agent(s) with outlier AHT pulling the group average up.
Medium-term fix
If AHT increase is sustained (3+ days), update the AHT assumption in the staffing model for future scheduling. Investigate whether knowledge base updates would reduce handle time. Distinguish talk-time causes (complexity or inefficiency) from ACW causes (wrap code selection or admin).
Skill imbalance
Diagnosis
Overall volume is at forecast but a specific skill group is in crisis while others are understaffed. Contacts are waiting in one queue while agents in adjacent queues are at low occupancy.
Intraday response
Temporarily reassign proficient agents from light queues to the heavy queue. Adjust routing priority thresholds if the WFM system allows real-time routing changes. Brief TLs in both the gaining and releasing team.
Medium-term fix
Review whether the skill balance in the weekly schedule reflects current contact mix accurately. If a specific skill consistently overloads relative to others, the multi-skill proficiency model or routing design needs review.
Understaffing in the schedule
Diagnosis
Volume and AHT are both on forecast, adherence is normal — but there are simply not enough agents scheduled. Occupancy is at or near 90% even with good adherence.
Intraday response
Request voluntary overtime for the remainder of the day if available. There is no intraday fix that replaces a structurally missing headcount without incurring cost.
Medium-term fix
Rebuild the schedule for the affected period using corrected staffing requirements. Investigate whether the understaffing is due to a planning model error (wrong shrinkage assumption, wrong Erlang input) or a headcount shortfall (actual headcount below plan due to attrition). The root cause determines the fix.
SL, occupancy, and cost: the management triangle
SL, occupancy, and cost are linked through Erlang C — you cannot improve one without affecting at least one other. Understanding the relationship prevents management decisions that improve one metric while making another worse.
Add agents to improve SL
effect on occupancy
Occupancy falls — agents spend more time available between calls. At low enough occupancy (<65%), agents may disengage.
effect on cost
Cost increases in direct proportion to agents added.
Reduce agents to cut cost
effect on sl
SL falls — the queue clearing rate reduces. The SL decline is non-linear: small reductions in staffing below the optimal level cause disproportionate SL deterioration.
effect on occupancy
Occupancy rises — agents become busier. Above 87%, agents have insufficient recovery time between calls.
Reduce AHT (without reducing quality)
effect on sl
SL improves — agents clear each call faster, so they become available sooner. The same headcount handles more volume.
effect on occupancy
Occupancy rises slightly — agents are more consistently on calls. Net effect is positive because the SL improvement more than compensates.
Service level management questions
What causes service level to fall in a contact centre?
Five distinct causes, each with a different response: (1) Volume above forecast — fix: intraday flexible resource deployment; (2) Below-target adherence — fix: intraday adherence management, not AHT pressure; (3) AHT above assumption — fix: immediate root cause investigation, then model update if sustained; (4) Skill imbalance — fix: routing adjustments and temporary cross-skilling; (5) Understaffing in the schedule — fix: voluntary overtime today, scheduling rebuild for the next period. Applying the wrong fix to the right cause wastes resource — intraday actions cannot fix a scheduling model error.
Related guides
Service level explained
What SL is and how it is calculated
Setting SL targets
How to set the right SL target
Intraday management
Real-time management of the SL
Queue management
Queue-level SL response
Adherence improvement
Adherence as an SL driver
Occupancy explained
The SL-occupancy relationship
Erlang C calculator
Calculate the SL your current staffing level delivers
SL target calculator
Decide what SL target is realistic for your operation