Metric Definitions
Heedify reports classify every Microsoft Teams call into exactly one outcome, and compute every service metric — abandon rate, Service Level, quick drops — from that single classification. This page is the reference: what each term means, which calls each widget counts, and where the thresholds come from.
Read it before comparing two widgets that seem to disagree. In almost every case, they are answering different questions.
The five call outcomes
Every call ends in exactly one of the five outcomes below. They are evaluated in this order — the first one that matches wins. That order is what makes the categories add up to the total, with no call counted twice and none left out.
| # | Outcome | The caller… | Counts as a service failure? |
|---|---|---|---|
| 1 | Answered | reached an agent, who picked up | No |
| 2 | Forwarded | was redirected by the call flow — voicemail, an external number, a fallback, or out of a full queue | No — the call was handled |
| 3 | Quick drop | reached a queue and hung up almost immediately | No — see below |
| 4 | Abandoned | reached a queue, genuinely waited, then gave up | Yes |
| 5 | IVR abandon | hung up during the greeting or menu, before ever asking for a queue | No — but watch it |
Answered
An agent picked up the call. Nothing else matters — even if the agent later transferred it elsewhere, the call was answered.
Forwarded
The call flow — not an agent — sent the call somewhere else. This covers every automatic redirection you configure:
- outside business hours, or on a holiday
- a menu option that routes straight to a person or an external number
- a queue that timed out or was full, and overflowed to voicemail or another destination
A forwarded call was handled. Your flow did exactly what you told it to do. It is never counted as an abandon, and it never penalizes Service Level.
This is the most common source of confusion. A caller who reaches your voicemail outside business hours has not abandoned anything — but older reports counted them as an abandon, which inflated the abandon rate on evenings, weekends and holidays.
Quick drop
The caller entered a queue and hung up before the Quick Drop Threshold (6 seconds by default). Misdials, second thoughts, someone who realizes they called the wrong department.
A quick drop is not a service failure: nobody could have answered a two-second call. Counting these as abandons would blame your agents for wrong numbers. They are reported separately so you can see them without them polluting the abandon rate.
Good to know — Set Quick Drop Threshold to
0in the Analytics Profile if you want every dropped call in a queue to count as an abandon. The quick-drop category then disappears from your reports.
Abandoned
The caller entered a queue, waited at least the Quick Drop Threshold, no agent ever answered, and the flow did not redirect them. They gave up.
This is the only true service failure. It is the metric to watch, and the one that drives the abandon rate.
IVR abandon
The caller hung up during the greeting or the menu, without ever asking to be put in a queue. They were never offered to your agents, so they cannot be an abandon and they do not affect Service Level.
That does not make them harmless. A high IVR abandon count usually means the front door is the problem — a greeting that runs too long, a menu that is hard to follow, or options that do not match what callers actually want. It is a quality-of-welcome indicator, not a staffing one.
Offered calls — the denominator
An offered call is an inbound call that entered a queue.
This is the base for every service metric: abandon rate, Service Level, answer rate. It is the population of calls that were actually presented to your contact center.
A caller who hangs up in the menu was never offered to anyone. Outbound calls are not offered to anyone either. Neither belongs in a denominator that measures how well your agents serve incoming demand — so neither is included.
| Included in offered | Not included |
|---|---|
| Answered | IVR abandons (never queued) |
| Abandoned | Outbound calls |
| Quick drops | |
| Forwarded out of a queue (overflow) | Forwarded before a queue (out-of-hours, menu routing) |
Note the last row. The same “forwarded” outcome sits inside or outside the denominator depending on when it happened. A call that overflowed out of a full queue was offered — your queue simply could not take it. A call redirected outside business hours never reached a queue at all.
Thresholds
Every threshold lives in the entity’s Analytics Profile. None are hard-coded, and changing one re-renders the widgets immediately — no data re-ingestion.
| Threshold | Default | What it decides |
|---|---|---|
| Quick Drop Threshold | 6 s | The line between a quick drop and an abandon |
| SLA Threshold | 60 s | The wait time under which an answered call counts as within Service Level |
| Quick / Standard / Delayed Answer Threshold | 30 / 60 / 90 s | The answer-speed bands in the Answer Time Thresholds widget |
| Target / Max (abandon, talking, waiting) | varies | The green / orange / red colors on metric cards — display only, they do not change any count |
SLA Threshold and Target Waiting Time are different things. The SLA threshold decides whether a call is counted as within Service Level. Target Waiting Time only colors the Max Waiting Time card. Setting one does not affect the other.
What each widget counts
Call volume
| Widget | What it answers | How it is counted |
|---|---|---|
| Total Calls | How much traffic did we get? | Every call in the period, inbound and outbound |
| Answered Calls | How many reached an agent? | Outcome 1 |
| Abandoned Calls | How many callers gave up on us? | Outcome 4. The percentage is out of offered calls |
| IVR Abandons | How many gave up at the front door? | Outcome 5. The percentage is out of inbound calls |
| Outbound Calls | How many calls did agents place? | Agent-initiated calls |
| Transferred Calls | How many calls changed hands? | Calls with a completed transfer, by an agent or by the flow |
| Call Status Distribution | How does traffic split across outcomes? | The five outcomes above. The slices add up to the total, exactly |
| Call Volume by Day / by Hour | When does traffic arrive? | Call count, grouped by the day or hour the caller arrived |
Service level
| Widget | What it answers | How it is counted |
|---|---|---|
| Service Level (SLA) | Are we picking up fast enough? | Calls answered within the SLA Threshold, divided by offered calls |
| SLA Trend | Is it getting better or worse? | The same calculation, one point per day |
| Answer Rate | What share of demand did we serve? | Answered calls, divided by offered calls |
| Daily SLA Statistics | Where do our answer times sit? | Answered calls per day, split into 0-15 / 15-30 / 30-45 / 45-60 / over-60s wait buckets |
| Calls Served by Time Ranges | Which flows answer fastest? | The same buckets, broken down per call flow |
| Answer Time Thresholds | How fast do agents pick up once rung? | Answered calls, split by how long the agent’s phone rang — not by how long the caller waited |
Service Level counts abandoned callers against you. A caller who waited three minutes and hung up lowers your Service Level, because they are in the denominator. This is the standard contact-center definition, and it is deliberate: a metric that only looks at the calls you managed to answer will always flatter you.
Abandons
| Widget | What it answers | How it is counted |
|---|---|---|
| Abandonment Rate | What share of offered callers gave up? | Abandoned calls, divided by offered calls |
| Quick Drops / Quick Drop Rate | How many were misdials? | Outcome 3, as a count and as a share of offered calls |
| Average Abandon Time | How long do they hold on before giving up? | Average wait of abandoned calls only |
| Max Abandon Time | What is the worst case? | The longest wait among abandoned calls |
| Daily Abandon Statistics | When do we lose them? | Abandons per day, split into wait-time buckets |
| Calls Abandoned by Time Ranges | Which flows lose callers? | The same buckets, broken down per call flow |
Wait, talk and handle time
| Widget | What it answers | How it is counted |
|---|---|---|
| Average Wait Time | How long does a caller queue? | Average queue wait across offered calls |
| Max Waiting Time | What is the worst wait? | The longest queue wait among offered calls |
| Average Ringing Time | How long does an agent’s phone ring? | Average ring time on answered calls |
| Average Handle Time (AHT) | How long does an agent spend on a call? | Conversation plus hold time, on answered calls |
| Average Talk Time | How long do they actually talk? | Conversation only — hold time removed |
| Max Talking Time | What is the longest conversation? | The longest conversation, hold time removed |
| Average Call Duration | How long is a typical call? | Average duration, with inbound-only and outbound-only variants |
Average Handle Time does not include after-call work. Heedify measures the time an agent is connected to the caller — conversation and hold. Wrap-up time after hanging up is tracked as an agent status, not as part of the call, so it cannot be attributed to a specific call.
Queues, flows and agents
| Widget | What it answers | How it is counted |
|---|---|---|
| Queue Overflow Rate | How often does one queue spill into another? | Calls that passed through two or more queues |
| Queue Overflow Matrix / Top Overflow Paths | Which queues spill where? | From-queue → to-queue hops, and the most frequent chains |
| Calls by Flow / Calls by Queue | Where does traffic land? | Call count per flow, per queue |
| Flow KPI Summary | How does each flow perform? | The five outcomes plus times and rates, one column per flow. The outcome rows add up to Total |
| Month-to-Date Summary | Same, for the current month | Same calculation, scoped to the month |
| Calls per Agent | Who handled what? | Per agent: answered, outbound, unanswered-when-rung, average ring time |
| Agent Status Statistics | How did agents spend their time? | Ready, In a call and Not Ready — these three add up to 100 % of connected time |
| Peak Hours Analysis | When are we busiest? | Call count and average wait per hour |
Agent Status Statistics reports what agents could actually do, not what they declared. An agent who sets themselves to Ready while Microsoft Teams shows them Busy receives no calls — so that time is not reported as Ready. The report reflects reality, which is what a supervisor needs.
Reading the reports without being misled
Two widgets can disagree and both be right. A call flow that redirects out of hours produces a Forwarded call, not an abandon — so a queue with heavy evening traffic will show a much lower abandon rate than the raw “calls that nobody answered” count you might compute yourself.
The abandon count is deliberately conservative. Quick drops, IVR abandons and forwarded calls are all excluded. If you want the number of callers who did not reach an agent for any reason, add Abandoned + Quick drops + IVR abandons.
Timezone matters at the edges. Days and hours are grouped in the timezone chosen in the filter bar, which follows your browser by default. A call arriving at 23:55 belongs to that day, in that timezone.
Calls are counted on arrival. A call that arrives at 23:55 and is answered at 00:05 belongs to the day it arrived, in every report.