Telephony fundamentals for WFM analysts
WFM analysts who misunderstand their telephony data make systematic errors throughout every output. Forecasting from offered contacts instead of handled contacts, calculating AHT without ACW, or interpreting SL using the wrong abandon methodology — each produces a predictable, persistent error.
Six telephony concepts every WFM analyst must understand
Automatic Call Distributor (ACD)
What it is
The telephony system that receives inbound contacts, queues them, and routes them to agents based on defined rules (skills, priority, longest idle agent, etc.). The ACD produces the interval-level data (calls offered, handled, abandoned, AHT, SL) that the WFM function uses for forecasting and reporting.
WFM significance
All WFM volume data comes from the ACD. Understanding how the ACD counts contacts — specifically what it classifies as 'offered' vs 'handled' vs 'abandoned' — is essential for building an accurate forecast. Different ACD systems use different definitions and counting methodologies.
Common WFM error
WFM team uses default ACD reports without understanding the counting logic. If the ACD counts a call as 'offered' when it enters the IVR rather than when it enters the agent queue, the offered volume overstates the demand actually reaching agents. Forecasting from this figure will produce overstated staffing requirements.
Interactive Voice Response (IVR)
What it is
The automated menu system that customers interact with before reaching an agent. Customers may self-serve in the IVR (resolving their issue without agent contact), select a routing destination, or abandon the IVR before reaching any destination.
WFM significance
IVR deflection (contacts self-served in IVR) reduces the volume that reaches agents. The WFM forecast must be based on post-IVR volume (contacts that exit the IVR and enter the agent queue), not pre-IVR volume (total calls entering the IVR). If the IVR deflects 15% of contacts, the agent staffing requirement is based on the 85% that reach the queue — not the full 100%.
Common WFM error
Forecasting from total IVR entry volume rather than agent queue entry volume. When IVR changes are made (new menu options, new self-service journeys), the deflection rate changes without the WFM team being notified — producing a persistent forecast error that is not obviously attributable to an IVR change.
Average Handle Time (AHT) in ACD data
What it is
AHT = Talk Time + Hold Time + After Call Work (ACW). The ACD records: talk time (agent connected to customer); hold time (customer on hold while agent remains on call); ACW (agent in wrap/disposition state after call ends, before becoming available for the next contact).
WFM significance
The ACD breaks AHT into its components. WFM staffing models typically use total AHT as the input — but understanding the component split helps identify AHT reduction opportunities (reducing hold time vs. ACW vs. talk time requires different interventions) and forecast AHT changes when process changes are made.
Common WFM error
Calculating AHT from total talk time only, excluding ACW. This understates AHT and overstates the number of contacts each agent can handle per hour. A WFM model that uses talk time (6 minutes) instead of AHT (8 minutes, including 2 minutes ACW) will understate the required staffing by 25%.
Service Level calculation in ACD
What it is
SL = (Contacts answered within threshold ÷ Total contacts offered) × 100. The SL threshold (e.g. 20 seconds) is configured in the ACD. Some ACDs calculate SL as a percentage of answered contacts (excluding abandoned); others include abandoned contacts in the denominator. The methodology significantly affects the reported SL figure.
WFM significance
The SL calculation methodology determines how abandons affect the reported SL. If abandons are excluded from the denominator, a high abandon rate can artificially inflate the reported SL (because the numerator and denominator both exclude abandoned contacts, keeping the ratio high). If abandons are included, a high abandon rate will suppress the reported SL even if answered contacts are fast.
Common WFM error
Assuming the SL metric from the ACD report is calculated using the same methodology as the SL target agreed with the business. Before reporting SL to stakeholders, confirm: (1) whether the threshold matches the agreed SL definition; (2) whether abandons are included or excluded from the denominator.
Real-time ACD queue metrics
What it is
Metrics updated in near-real-time (typically every 15–30 seconds): calls in queue (current count waiting to be answered); oldest call in queue (how long the longest-waiting contact has been waiting); average wait time (estimated wait for a new contact entering the queue now); agents available (agents ready to receive a contact).
WFM significance
Intraday management uses real-time ACD metrics to identify queue deterioration early. Oldest call in queue is typically the most useful single metric — a call that has been waiting 5 minutes in an operation with a 20-second SL target is a significant signal that requires immediate action, even if the interval-level SL has not yet been calculated.
Common WFM error
Intraday team monitoring interval-level SL reports (updated every 30 minutes) rather than real-time queue metrics. By the time the 30-minute SL report shows a problem, the affected interval has already ended. Real-time queue metrics allow intervention within the interval — not after it.
Abandon before and after threshold
What it is
Contacts can abandon the queue before or after the SL threshold. 'Abandon before threshold' — customer hung up before the SL target wait time. These are typically excluded from SL calculations in most ACD configurations (because the customer had not yet waited longer than the target). 'Abandon after threshold' — customer waited beyond the SL threshold and then hung up. These count as SL failures.
WFM significance
Short abandons (before threshold) may represent customers who dialled incorrectly, changed their mind, or misdialled. Long abandons (after threshold) represent customers who genuinely wanted help but left because the wait was too long. WFM should monitor both separately — a spike in long abandons indicates a real queue crisis; a spike in short abandons may be a data artefact or a system issue.
Common WFM error
Treating all abandoned contacts the same way in the abandon rate metric. A 5% abandon rate composed entirely of short abandons (before threshold) is a very different operational situation from a 5% abandon rate composed entirely of long abandons.
Telephony data questions
What is the difference between offered contacts, handled contacts, and abandoned contacts in ACD data?
Offered contacts — all contacts entering the ACD queue (answered + abandoned + IVR-deflected). Handled contacts — contacts connected to an agent and handled; excludes abandoned and IVR self-served contacts. Abandoned contacts — contacts that entered the queue but hung up before being answered. For WFM forecasting: use handled contacts as the staffing model input, because each handled contact drives agent time at the AHT rate. Use offered contacts to understand total customer demand. Do not forecast from offered contacts in a high-IVR environment — the IVR deflection will overstate the agent staffing requirement.
Related guides
Volume forecasting
Applying telephony data to forecasting
AHT guide
Understanding and managing AHT
Service level explained
How SL is defined and calculated
Abandonment rate
Managing and interpreting abandon data
Real-time adherence
Using ACD data for intraday management
WFM technology
WFM systems that integrate with ACD data
Erlang C calculator
Apply telephony data inputs to calculate FTE and SL
Service level target calculator
Derive the SL target from the call volume and AHT your telephony provides