Skip to main content
TurnellaBeta
WFM guideWFM governance

Contact centre WFM metrics

A WFM function measured only on SL is not measured on what it controls. SL is affected by adherence, AHT variance, and unforecast volume — all outside WFM control. WFM-specific metrics hold the planning function accountable for what it directly controls.

The six WFM function metrics

Forecast accuracy (WAPE) — daily level

What it measures

How accurately the WFM team forecast daily contact volumes. Daily WAPE is the most operationally relevant level — it drives the schedule.

Target

Below 10% excellent; 10–15% acceptable; above 20% requires model review

How to calculate

Sum of |actual − forecast| across all days ÷ sum of actual across the same days × 100

Common failure mode

Centre reports monthly WAPE (which looks good) while daily WAPE is significantly worse — monthly accuracy masks intraday distribution errors.

Schedule efficiency

What it measures

What percentage of scheduled agents are actually required to meet the SL target. High efficiency means the schedule is tight to the requirement; low efficiency means there is excess planned headcount.

Target

85–95% efficiency. Below 85% suggests significant over-scheduling; above 97% leaves no buffer for absence or AHT variance.

How to calculate

Erlang C requirement for each interval ÷ scheduled agents for that interval × 100, averaged across all intervals

Common failure mode

Schedule efficiency is not tracked at all — the WFM team has no systematic view of whether schedules are efficient or carrying significant hidden overhead.

Scheduling cycle lead time

What it measures

Whether the WFM team publishes schedules within the agreed lead time window — giving agents and team leaders enough notice to plan.

Target

Weekly schedule published at least 2 weeks ahead (4 weeks for operations with higher change management requirements). Intraday adjustments published with at least 24 hours notice where possible.

How to calculate

Track publication date for each schedule period vs. the agreed lead time standard. Report as % of schedule periods published within standard.

Common failure mode

Schedules published with 3–5 days notice rather than 2–4 weeks — forcing last-minute agent arrangements and reducing schedule adherence as agents cannot organise their outside commitments.

Shrinkage model accuracy

What it measures

How closely the shrinkage assumption used in the staffing model matches actual shrinkage each month.

Target

Actual shrinkage within ±2pp of the model assumption. More than 3pp deviation for 2+ consecutive months indicates the assumption needs updating.

How to calculate

Actual shrinkage % − model assumption % = shrinkage model error. Track by shrinkage component (external vs. internal) to identify which component is driving deviation.

Common failure mode

Shrinkage assumption is set once and never reviewed — an assumption set at 25% for 3 years while actual shrinkage has risen to 31% produces systematic understaffing without a visible root cause.

Intraday response time to queue crisis

What it measures

How quickly the WFM team identifies a developing queue crisis (volume or AHT above threshold) and takes a response action.

Target

Identification within 10 minutes of threshold breach. First response action taken within 15 minutes of identification.

How to calculate

Log the time of threshold breach (from wallboard data), the time of WFM identification, and the time of first response action for each intraday event. Track average and maximum response times.

Common failure mode

Queue crises are identified only after SL has been broken for a full 30-minute interval — the WFM team is responding to outcomes rather than monitoring for early warning signals.

Headcount plan accuracy

What it measures

How accurately the WFM team forecasts the required headcount at the planning horizon (typically monthly, 2–3 months forward).

Target

Actual FTE required within ±5% of the plan at the 1-month horizon; ±10% at the 3-month horizon.

How to calculate

Planned FTE requirement for the month (from the planning model) vs. actual FTE required (calculated retrospectively from actual volume, AHT, and shrinkage). Report as percentage deviation.

Common failure mode

Headcount plans are not retrospectively reviewed — the planning team has no feedback loop on whether their plans were accurate, so model errors persist without correction.

WFM metrics questions

What metrics should a contact centre WFM team be accountable for?

Six metrics the WFM function directly controls: (1) Forecast accuracy (WAPE) at daily level; (2) Schedule efficiency (Erlang requirement ÷ scheduled agents); (3) Scheduling cycle lead time (schedules published within agreed notice period); (4) Shrinkage model accuracy (actual vs. modelled shrinkage); (5) Intraday response time (time from threshold breach to first response action); (6) Headcount plan accuracy (±5% at 1-month horizon). The WFM function should not be held directly accountable for SL or cost per contact — both are affected by factors outside WFM control.

Related guides