Contact Center Workforce Management (WFM)

WFM forecasting predicts the volume of contacts that will arrive at each specific hour of each day of the week, enabling schedule generation that positions the right number of agents in the right place at the right time. Forecasting inputs: Historical contact volume: the most important input — typically 13+ months of historical data (to capture full seasonal cycles) by channel, hour, and day of week. The volume pattern repeats with high reliability: Monday mornings generate higher volume than Thursday afternoons; January generates higher volume than August; post-release days generate volume spikes. These patterns are the foundation of the forecast. Business drivers: planned marketing campaigns (email blasts generate immediate contact spikes), scheduled product maintenance windows, upcoming holidays, and known seasonal drivers in the customer type. Growth trend: if the customer base is growing by 3% monthly, adjust baseline volume up by 3% monthly. Shrinkage factors: the percentage of scheduled time agents are unavailable for contacts — vacation, training, team meetings, coaching sessions, unplanned absence. Typically 25–35% of total scheduled time. The forecast outputs a staffing requirement by 15-minute or 30-minute interval. WFM tools (Calabrio, Verint, Assembled, NICE WFM) automate the forecast generation and schedule optimization, but the underlying inputs still require human validation by the WFM analyst.

Erlang C is a queuing theory formula developed in 1917 by Danish mathematician Agner Krarup Erlang — and despite its age, it remains the mathematical core of contact center staffing because it accurately models the queue dynamics of random arrival traffic. What Erlang C calculates: given a known average contact arrival rate, average handling time, and number of agents, Erlang C calculates the probability that a contact will be queued (rather than immediately answered) and the expected wait time in queue. Working backwards: given a target probability of queuing (e.g., "only 5% of contacts should wait more than 20 seconds") and the traffic parameters, Erlang C tells you the minimum number of agents required to meet that target. Why it matters: naive staffing (contacts per hour ÷ contacts per agent per hour = agents needed) systematically understaffs because it doesn't account for the variability of random arrival. Contacts don't arrive in a perfectly smooth flow — they cluster randomly, creating moments of queue backlog even when average volume is "covered." Erlang C accounts for this randomness, producing the higher staff count required to maintain service levels during typical arrival spikes. WFM tools have Erlang C calculations built in; WFM analysts must understand the model well enough to validate its outputs for unusual scenarios.

Real-time management is the discipline of monitoring actual contact arrival, queue status, and agent adherence during a live shift and making micro-adjustments to maintain service levels as reality deviates from the forecast. RTM responsibilities: Queue monitoring: continuously watching the queue dashboard — contacts in queue, oldest contact age, available agents, current service level (% of contacts answered within target time in the last 30 minutes). When the queue builds beyond a threshold, the RTM analyst acts. Agent adherence management: comparing the schedule (when agents should be available) to actual status (when agents are actually logged in and ready). An adherence rate below 90% indicates significant off-schedule activity. Low adherence in a high-volume period is a primary service level risk. Intraday reforecasting: when actual volume is significantly different from forecast (±15%), the RTM analyst adjusts the staffing requirement estimate for the remaining hours of the shift and identifies whether overtime or voluntary early departure adjustments are warranted. Escalation communication: when a service level breach is predicted or occurring, the RTM analyst notifies the Support Manager immediately with the specific breach status, estimated recovery time, and options for immediate response (pulling agents from other queues, initiating voluntary overtime). RTM is the operational nerve center of large contact center operations; in smaller teams (< 20 agents), RTM is typically a part-time responsibility of the team lead or supervisor.