WFM guideReal-time operations

Contact centre queue management

A queue that is growing and a queue that is already broken require completely different responses. Queue management is the real-time discipline of recognising which situation you are in — and acting before SL crosses the threshold where customer experience is damaged.

Queue signals, thresholds, and response actions

Effective queue management requires pre-agreed thresholds — not reactive eyeballing of the wallboard. For each signal below, the threshold is the point at which action should already be taken, not the point of crisis.

Calls in queue rising faster than normal arrival pattern

Threshold

Calls in queue >150% of planned 15-min interval forecast for more than 5 minutes

Action

Pull flexible resource (discretionary break hold, training postponement, back-office with call skills). Alert team leaders to hold ready-to-call agents.

Do not

Do not wait for SL to break before acting. By the time SL falls below target, the interval is usually already lost.

Longest wait time approaching SL threshold

Threshold

Oldest call in queue within 30 seconds of SL target (e.g. if SL is 80% in 20s, oldest call >15s)

Action

Open overflow to secondary skill group if available. If not, alert TLs to maximise availability (brief agents on calls to close if safe to do so).

Do not

Do not route overflow to agents not proficient in the queue type — a routed contact handled poorly does more damage than a slightly extended wait.

Abandonment rate rising above expected range

Threshold

Abandonment >8% in current 15-min interval (or above your operation-specific target)

Action

Review call in queue distribution — if calls are clustering in specific time slots, the issue may be forecast-based, not adherence-based. Trigger callback offer if system supports it.

Do not

Do not assume rising abandonment is always a staffing issue — check IVR functionality, recent routing changes, and whether a repeat contact event is generating re-queuing.

Occupancy above 90% for sustained period

Threshold

Occupancy >90% across the skill group for >15 minutes continuously

Action

Agents cannot clear the queue faster because they are unavailable between calls. Add resource from flexible pool — the only solution is more agents taking calls.

Do not

Do not instruct agents to reduce AHT under these conditions. AHT pressure during a queue crisis increases repeat contacts and worsens the medium-term position.

Single skill group in crisis while others are stable

Threshold

One queue at >200% forecast volume while adjacent queues are under 80% forecast

Action

Cross-skill agents from the light queue to the heavy queue if proficiency permits (minimum proficiency level 3 on target skill). Temporarily suspend low-urgency contact types.

Do not

Do not cross-skill agents with no prior training on the queue type. A proficiency 1 agent on a complex queue increases AHT and potentially produces errors.

Queue clearing after a volume spike

Once a volume spike has passed — volume has returned to forecast or below — the queue still needs to be actively cleared. The mistake is standing down flexible resource the moment volume normalises. The queue that built during the spike still exists and will continue to drive abandoned calls and repeat contacts until it is cleared.

Confirm volume has normalised below forecast before standing down

Check the current 15-minute arrival rate against the forecast for this interval. If volume is still 10% above forecast, the queue will continue growing. Only stand down flexible resource once volume has fallen to or below forecast for two consecutive 5-minute intervals.

Hold flexible resource until the queue depth has halved

Even after volume normalises, maintain flexible resource until the queue depth (calls waiting) has fallen to at least half of its peak. A queue of 80 calls after a spike does not clear instantly — with 50 agents at 75% occupancy, clearing 80 extra contacts takes approximately 12–15 minutes.

Release resource in stages, not all at once

Standing down 20 agents simultaneously creates an immediate dip in staffed capacity that can restart the queue. Release in groups of 3–5 agents every 5 minutes, monitoring abandonment and wait time after each release before releasing the next group.

Re-open discretionary activities only after the queue is fully clear

Training sessions, team meetings, 1-to-1s, and back-office periods that were suspended during the spike should not resume until the queue is clear and SL has returned to within 5pp of target for two consecutive intervals.

Log the spike and spike response in the intraday log

Record: when the spike started, peak queue depth, response actions taken, time to queue clearance, and whether the spike was forecast or unforecast. Repeated unforecast spikes at the same time period are a forecasting accuracy problem — the intraday log creates the evidence base to correct it.

Priority routing during a queue crisis

Routing lever	How it works	Risk	Best for
Skill priority rebalancing	Temporarily lower the skill priority threshold so more agents are offered calls from the overloaded queue	Lower-proficiency agents handle contacts they may not handle well — watch FCR and QA scores post-crisis	Queues where most agents have some proficiency but are assigned to other primary skills
Overflow to secondary queue	Route contacts that have waited beyond a threshold to an overflow group — typically a team with broader skills	Overflow group agents may lack full context or system access for the contact type — briefing is required	Transactional queues where agent capability difference is small
Suspend low-urgency contact types	Temporarily halt outbound campaigns, callback handling, or email processing to release those agents to the voice queue	Creates a backlog in the suspended channel — must be managed in the post-crisis session	Operations with blended inbound/outbound or multi-channel agents
Callback / scheduled callback offer	IVR or queue system offers customers a callback rather than holding — flattens the real-time queue	Callbacks must be honoured reliably; a callback offer that is not delivered destroys trust	Operations where wait times are long (>5 minutes) and callback fulfilment SLAs can be met

Queue management questions

What is the difference between queue management and intraday management in a contact centre?

Intraday management is the broader real-time function: it includes schedule adherence tracking, break management, and updating the day's staffing position. Queue management is one component within intraday management — specifically the monitoring of queue length, wait time, and abandonment, and the response actions taken when the queue is building. A centre can have excellent intraday management (every agent in adherence, breaks running to plan) and still have a queue crisis if volume has arrived faster than forecast or a skill group is unbalanced. Queue management is about the contacts waiting, not just the agents working.