69 contact center terms defined in plain language — metrics, dialing modes, routing, workforce management, compliance and quality assurance. Formulas are included wherever a term has one.
What this glossary covers
A call center glossary is a reference list of the terminology used to run a contact center. This one covers 69 terms across six areas: performance metrics such as AHT, FCR and occupancy; the four dialing modes; routing and queue concepts; workforce management; compliance terminology including TCPA and STIR/SHAKEN; and quality assurance. Terms that have a standard formula include it.
A call that ends before it reaches an agent. On inbound, the caller hangs up while queued; on outbound, the dialer connects a live person but has no agent free to take it. Outbound abandonment is regulated in the US and is capped under the FTC's Telemarketing Sales Rule.
The percentage of connected calls that end before an agent handles them. Formula: (abandoned calls / total calls offered) x 100. A healthy inbound rate is 2-5%; above 8% usually points to understaffing or an IVR that traps callers.
How closely agents work the hours they were scheduled to work, including breaks taken at the assigned times. Formula: (time in adherence / total scheduled time) x 100. Top-quartile centers reach roughly 92%.
The time an agent spends finishing a contact after the caller has disconnected: notes, dispositions, CRM updates, follow-up tasks. ACW is a component of average handle time and is one of the first things automation shortens.
The share of an agent's paid time spent on productive contact work, including talk, hold and after-call work. Typically targeted at 75-85%. Distinct from occupancy, which is measured against logged-in time rather than paid time.
Technology that decides whether an outbound call was answered by a person or a voicemail system, so the dialer can route live answers to agents and drop a message on machines. Aggressive AMD settings reduce agent idle time but increase the risk of hanging up on real people.
The system that queues inbound calls and routes each one to an agent according to rules such as skill, priority, longest idle time or source. The ACD is the core of any inbound operation.
The carrier-supplied number of the calling party, delivered independently of caller ID. Because ANI is harder to spoof than caller ID, it is commonly used for screen pops and account lookup.
The average total time an agent spends on one contact: talk time plus hold time plus after-call work. Formula: (talk + hold + ACW) / contacts handled. Voice AHT commonly runs 4-7 minutes, but the healthy figure is entirely vertical-dependent.
The average time an inbound caller waits in queue before an agent answers. Global benchmarks land near 25-28 seconds. ASA hides its own tail, so always read it alongside abandonment rate.
A supervisor joining a live call so that both the agent and the customer can hear them. Used to rescue a call that is going badly. Contrast with monitoring and whisper coaching.
An operation that handles inbound and outbound contacts on the same agent pool, with the platform shifting agents between the two as volume moves. Blending raises utilization but complicates forecasting.
An outsourced provider that runs contact center operations on behalf of a client brand, either onshore, nearshore or offshore. BPOs are usually measured on contractual service levels rather than internal KPIs.
Contact Center as a Service. A cloud-delivered contact center platform sold on subscription, where the vendor operates the telephony, routing, recording and reporting infrastructure. Replaces on-premise ACD and dialer hardware.
A live connection between the contact center platform and the system of record, so agents get a screen pop on connect, dispositions write back automatically, and lead status stays consistent in both systems.
Automatically moving agents between inbound and outbound work in response to real-time queue conditions, so that agents are not idle while an outbound campaign waits.
The outcome code an agent applies at the end of a contact - sale, callback, wrong number, do not call, no answer. Dispositions drive redial rules, list hygiene and campaign reporting, so a sloppy disposition scheme corrupts every downstream metric.
Capturing audio, and sometimes screen activity, from contacts for quality assurance, training, dispute resolution and regulatory evidence. Consent requirements vary by state and by country.
The logic that decides which agent or queue receives a given contact. Common strategies are skill-based, source-based, priority, round-robin and least-occupied routing.
An arrangement where the center returns the call rather than holding the customer in queue. Scheduled callbacks are agent-initiated at an agreed time; virtual hold callbacks preserve the caller's place in line and ring them when an agent frees up.
A configured outbound calling effort with its own lead list, dialing mode, pacing, scripts, dispositions, caller ID set and calling-hour rules. Campaign design is where most outbound performance is won or lost.
Checking a lead list against the National Do Not Call Registry, internal DNC lists, litigator lists, state DNC lists and reassigned-number data before dialing. Scrubbing is done before every campaign, not once at import.
The percentage of outbound dial attempts that reach a live person. Formula: (live connects / dial attempts) x 100. Connect rate is driven far more by list quality, caller ID reputation and time-of-day strategy than by dialer settings.
The percentage of unique records in a list that were reached at least once during a campaign. Distinct from connect rate, which is measured per dial attempt rather than per record.
The fully loaded cost of handling one contact: labor, technology, telecom, facilities and overhead divided by contacts handled. Assisted contacts commonly run around $13 while self-service contacts run under $2, which is the entire economic case for deflection.
A survey metric asking how much effort the customer had to expend to get their issue resolved, usually on a 1-7 scale. CES predicts repeat contact and churn more reliably than satisfaction scores do.
The percentage of surveyed customers who rate an interaction positively, typically the top two boxes on a five-point scale. A CSAT of 85% or better is the common 2026 target. CSAT is a trailing indicator - when it falls, look at occupancy, then AHT inflation, then FCR.
The system that places outbound calls automatically and delivers answered calls to agents. The four common modes are preview, progressive, power and predictive, which differ in how many numbers are dialed per available agent.
The US national list of consumer numbers that may not receive telemarketing calls. Sellers must scrub against it, keep an internal DNC list, and honor requests. State registries and internal suppression lists apply on top of the federal one.
Under US outbound rules, a call answered by a live person where no agent is connected within two seconds of the person's greeting. Dropped calls are what abandonment limits actually regulate.
The queuing formula used to convert forecast contact volume and handle time into the number of agents needed to hit a service level target. Erlang C assumes callers wait rather than abandon, so it overstaffs slightly for centers with high abandonment.
Transferring a contact to a supervisor, specialist team or higher support tier when the first agent cannot resolve it. High escalation rates usually indicate a training, tooling or routing problem rather than an agent problem.
The percentage of contacts resolved on the first interaction with no follow-up needed. Industry average runs 70-74%; above 80% is top-tier. Retail averages near 78%, healthcare near 71%, financial services 67-71%. Widely treated as the single most important measure of call center effectiveness.
How closely predicted contact volume matched actual volume, usually reported as mean absolute percentage error. Poor forecast accuracy shows up as either abandoned calls or idle agents, and both are expensive.
Time the customer spends on hold during a contact that has already been answered. Counted inside average handle time, and tracked separately because long holds damage satisfaction even when total handle time looks fine.
An automated phone system that collects caller input by keypad or speech and routes, answers or authenticates accordingly. Badly designed IVR trees are one of the most common causes of high abandonment.
Ranking records in an outbound list by likelihood to convert, so the dialer works the best records first and during the best contact windows. Scoring lifts revenue per agent hour without changing headcount.
The percentage of a lead list that has been dialed at least once in a campaign cycle. Formula: (records attempted / total records) x 100. Low penetration with high dial volume means the dialer is re-hitting the same small pool.
A supervisor listening to a live call without either party hearing them. Also called silent monitoring. The listen-only counterpart to whisper coaching and barge-in.
A loyalty metric derived from one question about likelihood to recommend, on a 0-10 scale. Formula: percentage of promoters (9-10) minus percentage of detractors (0-6). NPS measures the relationship, not the individual interaction, so it is a poor substitute for CSAT at contact level.
The share of an agent's logged-in time spent handling contacts rather than waiting. Formula: ((talk + hold + ACW) / logged-in time) x 100. The healthy band is 75-82%. Sustained occupancy above 85% reliably produces satisfaction decline and an attrition spike about 90 days later.
Handling voice, SMS, email, chat and social contacts in one platform with shared context, so a customer who switches channel does not have to repeat themselves. Distinct from multichannel, where each channel runs in its own silo.
Agent-initiated or system-initiated calls to customers or prospects for sales, collections, appointment reminders, surveys or service notifications. Subject to consent, calling-hour and abandonment rules in most jurisdictions.
The number of numbers a predictive dialer places per available agent, for example 2.5:1. Raising the ratio increases connects per agent hour and also increases dropped calls, so pacing is tuned against the abandonment ceiling rather than against talk time alone.
A dialer that places a fixed number of calls per available agent regardless of forecast, for example three at a time. Simpler than predictive dialing and easier to keep inside abandonment limits, but less efficient on large lists.
A dialer that forecasts when agents will become available and dials several numbers ahead of them, so a connected call is waiting the moment an agent frees up. Highest talk-time efficiency of any mode, and the only one that generates meaningful abandonment risk.
A dialer that shows the agent the record before the call is placed and lets them decide when to dial. Slowest mode by connects per hour, best for complex, high-value or sensitive conversations that need preparation.
A dialer that places exactly one call per agent, only once that agent is free. Removes abandonment risk entirely while keeping the agent out of manual dialing. The usual choice for regulated verticals.
The process of scoring recorded or live contacts against a defined rubric, then coaching against the results. Manual QA typically samples 2-5 calls per agent per month; automated QA scores every contact, which changes what the program can detect.
The holding line for contacts waiting on an available agent. Queue configuration - priority, overflow, maximum wait, callback offers, announcements - determines abandonment as much as staffing does.
The FCC-administered database that lets callers check whether a phone number has been disconnected and reassigned to a new subscriber since consent was obtained. Correct use provides a limited safe harbor against liability for calling a reassigned number.
A connect with the specific person the record identifies, rather than anyone who happens to answer. The metric that matters in collections and in any campaign where speaking to a household member is not equivalent to speaking to the debtor or customer.
The caller ID authentication framework US voice providers use to cryptographically attest that a calling number is legitimate. Calls with weak or absent attestation are more likely to be labeled or blocked before they ring.
The warning carriers and handset apps display on inbound calls their analytics score as unwanted. A labeled number collapses answer rates. Remediation involves caller ID reputation monitoring, number rotation, branded calling registration and reducing complaint-generating behavior.
The CRM record that appears on the agent's screen the instant a call connects, matched by ANI or by campaign record. Saves 15-30 seconds of lookup per contact and removes the awkward silent opening.
The percentage of contacts answered within a target time, written as answered/seconds. The classic target is 80/20 - 80% of calls answered within 20 seconds. Voice queues in 2026 commonly sit between 75/30 and 85/20.
The share of paid agent time that is not available for handling contacts: breaks, training, meetings, coaching, absence, system downtime. Commonly 30-35%. Understating shrinkage is the most frequent cause of a staffing model that fails in week one.
Routing each contact to the agent whose skill profile best matches it - language, product, seniority, licensing. Raises resolution rates and lengthens queue times, so skill groups need to be broad enough to stay staffable.
The Telephone Consumer Protection Act, the 1991 US federal statute governing autodialed calls, prerecorded messages and texts. Enforced by the FCC and through private suits, with statutory damages of $500 to $1,500 per violation and no aggregate cap.
The portion of a contact where the agent and customer are actually speaking, excluding hold and after-call work. A component of average handle time, and the numerator most agent productivity measures are built on.
The FTC rule governing telemarketing practice: calling hours are limited to 8am-9pm local time, abandoned calls are capped, most calls to registry numbers are prohibited, and specified records must be retained.
The percentage of contacts an agent passes to another agent, team or tier. Formula: (transferred contacts / handled contacts) x 100. Rising transfer rate is usually a routing or training signal before it is an agent performance signal.
Offering a queued caller a callback that keeps their place in line instead of making them wait on hold. One of the most reliable ways to cut abandonment without adding headcount.
An automated conversational system that handles a voice contact end to end using speech recognition and a language model, escalating to a human when it cannot resolve. Under the FCC's February 2024 guidance, AI-generated voices fall under the TCPA rules for artificial or prerecorded voice calls.
Leaving a pre-recorded message on an answering machine and immediately releasing the agent to the next call, rather than having them record each message live. Saves roughly 30 seconds per unanswered call across a campaign.
Transferring a call after briefing the receiving agent while the customer is on hold, so the customer does not have to re-explain. Contrast with a cold transfer, which passes the call with no handover.
A supervisor speaking to the agent during a live call so that only the agent hears them. The middle option between silent monitoring and barge-in.
The discipline of forecasting contact volume, building schedules that meet it, and managing adherence in real time. WFM is where service level targets either become achievable or stay aspirational.
The structured outcome an agent selects during after-call work. Wrap-up codes feed redial logic, suppression rules and campaign reporting, which is why a bloated code list quietly degrades every outbound metric.
Every metric in this glossary is one you can watch live in DialedIn's dashboards.