WFM guideHeadcount lifecycle

Contact centre agent lifecycle and WFM

Q: What WFM data should be used in an agent performance review?

Four WFM data sources are relevant to an agent performance review: (1) Schedule adherence trend — the agent's adherence percentage over the last 3–6 months compared to the team average and the target. A deteriorating adherence trend that predates a formal performance concern is often the first quantitative signal of an agent at risk. (2) Absence frequency and pattern — the number of absence episodes and whether they show a pattern (e.g. consistently on Mondays, consistently before/after shifts they dislike, consistently on busy days). The Bradford Factor is a useful trigger for a more detailed absence conversation. (3) Ramp progression (for agents in their first 12 months) — whether the agent's productivity has progressed in line with the ramp curve applied in the staffing model, or whether it has plateaued below the expected trajectory. (4) AHT trend — whether the agent's AHT has moved significantly relative to the team average. A sustained increase in AHT may indicate a knowledge or confidence gap; a sustained decrease may indicate quality shortcuts that will show up in quality scores. None of these WFM metrics should be used in isolation — they provide context for the performance conversation, not a verdict.

The WFM function has a role at every stage of the agent lifecycle — not just when the headcount plan is built and when adherence deteriorates. Active engagement at each stage produces fewer capacity surprises and contributes to better retention outcomes.

The six lifecycle stages and WFM's role at each

Candidate pipeline (pre-hire)

WFM inputs at this stage

The approved headcount plan provides the recruitment target by month and quarter. WFM must specify the required start date for each cohort, working backwards from the production-ready date using induction duration and ramp time.

WFM outputs at this stage

Cohort sizing (how many agents to hire per intake); required offer date (accounting for recruitment lead time + induction + ramp); start date configuration in the WFM system for schedule generation.

Common engagement gap

Recruitment and WFM operate on different timelines. WFM produces the headcount plan in October; recruitment starts hiring in January based on a Q1 headcount gap without knowing the ramp-time implications of a Q2 production date. The offer date required to hit the production target is rarely communicated explicitly from WFM to recruitment.

Induction (weeks 1–4 off-phone)

WFM inputs at this stage

Confirmed cohort start date and size from HR/recruitment. Induction duration and schedule from L&D.

WFM outputs at this stage

Induction agents do not appear in the scheduled staffing position — they are off-phone and must be excluded from the productive capacity calculation. WFM must confirm that induction timing does not create a capacity gap during peak periods by pulling agents off the phone who were planned as productive in the headcount model.

Common engagement gap

Induction agents are included in the headcount count but not the scheduled capacity. Operations management assumes a 10-person cohort joining adds 10 agents to the scheduled position on day one — but they are in induction for 4–8 weeks. This creates a phantom headcount gap that only becomes visible when the schedule is built.

Ramp (weeks 4–16 on-phone at reduced productivity)

WFM inputs at this stage

Ramp factors from historical cohort data (typical: 60%/80%/92% of full productivity at weeks 4, 8, 12 from production start). Actual AHT and throughput from WFM system for each new starter.

WFM outputs at this stage

Adjusted effective FTE in the staffing model — each ramping agent's contribution is multiplied by the appropriate ramp factor rather than counted as 1.0 FTE. WFM must update the ramp factor at each 4-week review based on actual productivity data rather than assuming the default curve applies uniformly.

Common engagement gap

Ramp factors applied at cohort level rather than individual level. A cohort of 10 agents may have 3 progressing ahead of the standard curve and 4 behind it — using the cohort average masks the individuals who need targeted support, and underestimates the capacity contribution of the fast-progressors.

Experienced production (fully productive)

WFM inputs at this stage

Ongoing schedule adherence data; AHT trends; absence frequency. WFM-generated analytics should flag agents whose metrics are moving in a direction that may predict early exit (declining adherence, increasing absence, AHT diverging from team average).

WFM outputs at this stage

WFM data for line manager performance conversations; schedule optimisation based on individual productivity. WFM should proactively supply adherence and absence trend data to team leaders at regular intervals — not wait for a formal performance trigger.

Common engagement gap

WFM data shared only when formally requested for a performance management process. By the time a formal process is triggered, the pattern has often been visible in the WFM data for 3–6 months. Early sharing of WFM trend data enables earlier intervention — before the agent's engagement has deteriorated to the point where exit is likely.

Performance risk (deteriorating metrics)

WFM inputs at this stage

The WFM function flags agents whose adherence has fallen below the agreed threshold consistently for 4+ weeks, or whose absence frequency has exceeded the Bradford Factor trigger. This is a referral to the line manager and HR, not a WFM-owned performance management process.

WFM outputs at this stage

For capacity planning purposes: flag the agent as an attrition risk in the rolling headcount forecast. Update the probabilistic attrition projection to include the agent in the expected leaver pool, rather than waiting for the resignation to be confirmed. This moves the headcount gap forward by 2–4 months in the forecast.

Common engagement gap

The WFM function does not update the headcount forecast when agents enter a formal performance process. An agent on a performance improvement plan has a materially higher probability of leaving than a general agent — but the forecast treats them as if they will be in role for the full planning horizon until a resignation letter arrives.

Exit (notice period and last working day)

WFM inputs at this stage

Confirmed resignation notice and last working date from HR. WFM must immediately update the headcount forecast to remove the agent from the productive headcount position from their last working date, and confirm whether the resulting headcount gap requires a recruitment action.

WFM outputs at this stage

Updated rolling headcount forecast; revised recruitment trigger assessment; knowledge transfer plan for any specialised skills or responsibilities the departing agent held. If the departing agent was a skill-pool anchor (e.g. the most senior agent for a specific language or product), flag this to Operations as a scheduling risk, not just a headcount metric.

Common engagement gap

Agents serve out notice periods at reduced engagement and effectiveness. The WFM model counts them as 1.0 FTE in the headcount forecast during their notice period — but in practice, agents on notice have higher absence rates, lower adherence, and lower throughput than productive agents. Apply a notice-period discount (typically 60–80% of standard productivity) when the agent is confirmed as a leaver.

Agent lifecycle WFM questions

What WFM data should be used in an agent performance review?

Four sources: (1) Schedule adherence trend — the agent's adherence percentage over the last 3–6 months vs team average and target. Deteriorating adherence often precedes formal performance concerns by months. (2) Absence frequency and pattern — episode count and whether absences cluster on specific days or shift types. Bradford Factor provides a single-number trigger. (3) Ramp progression (for first-year agents) — whether productivity has followed the expected ramp curve or has plateaued below it. (4) AHT trend — sustained AHT increase may indicate a knowledge gap; sustained AHT decrease may indicate quality shortcuts. None of these metrics is a verdict — they provide context for the conversation.