Staffing Reference Guide

Overview of the Forecast-to-Staffing Flow

  • Forecast Generation (Step 1 to Step 3)

    • Inputs:

      • Historical Data Period: Determines the data range used for trend analysis.

      • Forecast Date Range: Future time period for which forecast is needed.

      • Skills: Forecasts are generated per skill (or skill group).

      • Interval: Forecast granularity (e.g., 15-min, 30-min).

    • Output:

      • Volume (calls/chats/etc. per interval per skill)

      • Average Handling Time (AHT) per interval per skill

  • Staffing Generation (Step 4 and Step 5)

    • Inputs:

      • Forecasted Volume and AHT from Step 3

      • Staffing Parameters:

        • SLA (For example, 80% in 20s)

        • ASA (Average Speed of Answer)

        • Shrinkage (For example, 25%)

        • Occupancy Target (For example, 85%)

        • Concurrency (for chat, email)

    • Output:

      • Required Staffing Count per interval per skill

Staffing Formula

The base formula used for calculating raw staffing requirement (before applying modifiers like shrinkage) is:

Required Staff = (Volume × AHT) / Interval Duration (in seconds)

This proivdes Total Handling Time per interval, which is then divided by available handling time per agent (i.e., interval duration in seconds), assuming full occupancy.

For example (for 15-min interval):

  • Forecasted Volume = 100

  • AHT = 300 seconds

  • Interval = 15 minutes = 900 seconds

Raw Staffing = (100 × 300) / 900 = 33.33 ≈ 34 agents

Then adjusted for shrinkage, occupancy, and SLA buffers.

Calculation Considerations

  • Skill Allocation Logic

    If a skill is shared across multiple scheduling units (SUs),

    • Forecast is first generated at the skill level.

    • During staffing, the staffing output is then proportionally split across SUs based on skill-to-SU mapping or weight distribution.

  • Concurrency Handling

    For digital channels (For example, chat email), the concurrency value reduces the required headcount.

  • Shrinkage Application

    After raw staffing is calculated, shrinkage is applied:

  • Final Staffing = Raw Staffing / (1 - Shrinkage %)

Reference Inputs

Parameter

Example Value

Purpose
AHT 300 sec Total average time to handle an interaction
Interval 900 sec Forecast granularity (15 min)
Shrinkage 25% Absenteeism, break, training, etc.
Occupancy 85% Utilization cap
SLA/ASA Configurable Customer experience target

Simulation Step – Mimicking Real-World ACD Behavior

Before finalizing staffing recommendations, the system runs a simulation engine that mimics how an Automatic Call Distributor (ACD) would behave in the real world. This simulation plays a critical role in:

Validating Staffing Against SLA Goals:

  • It tests whether the configured staffing levels for each interval and skill actually meet the target SLA or ASA.

  • Uses queuing theory models under simulated load conditions.

Distributing Staffing Across Scheduling Units (SUs):

  • In cases where multiple Scheduling Units share a skill, the simulation helps in intelligently distributing the required staffing load across SUs.

  • Distribution is based on routing rules, SU capacity, volume weights, and concurrency logic.

  • The simulation identifies how the interaction load flows through the network, reflecting ACD routing decisions in reality.

Why This Matters

This simulation-based approach goes beyond static formulas. It provides staffing recommendations that reflect real-world operations, helping improve SLA adherence and agent efficiency.

Key Talking Points for Customers

  • Staffing is based on skill-level forecast of volume and AHT.

  • We apply industry-standard models to ensure SLA targets are met, accounting for shrinkage and occupancy.

  • The system ensures staffing is aligned to business needs per interval and skill, with flexibility in configuration.