Why call abandonment rises and how to reduce it

The short answer

Call abandonment — a caller ending the call before an agent answers — rises for four main reasons: the wait is long, the caller does not know how long they will wait, there is no alternative, or the call came out of hours and nobody answers. To reduce it, first split abandonment by wait time: a call dropped in the first 5 seconds and one dropped in the third minute are different problems.

This article shows how to split abandonment, how to find the root cause for each group, and which actions — wait information, a callback offer, business hours, answering without a queue — fit which cause.

How to calculate abandonment

A simple definition: the share of calls that enter the queue but end before reaching an agent. But two decisions must be made in advance. First: do calls ending in the first few seconds count? Some of them are misdials. Second: are calls ending in the menu counted separately? Fix these definitions, or month-to-month comparison is meaningless.

Splitting by wait time

1. 10–10 secondsWrong number, a change of mind, or hanging up on meeting a long greeting or menu.
2. 210–60 secondsImpatience and uncertainty: not knowing how long the wait will be.
3. 31–3 minutesThe classic queue problem: a long wait and no alternative.
4. 4Over 3 minutesA serious capacity problem or a "stuck" line — the caller waited a long time, then gave up.

A root-cause table

• Long introduction and menu → abandonment in the first seconds → shorten the greeting, simplify the menu
• No information about time → abandonment at 10–60 seconds → a hold message, an estimated time
• Not enough capacity at peak → abandonment at 1–3 minutes → shift plan, overflow, automate simple calls
• No alternative → abandonment on long waits → a callback offer
• Out-of-hours calls enter the queue → abandonment on an "empty line" → a night scenario, a note and a callback

Looking at hours and days

Abandonment always clusters in particular hours: Monday morning, lunchtime, the end of the working day, campaign days. A weekly average hides these peaks. Show abandonment in a table by hour and weekday — the three or four reddest cells are usually most of the whole problem.

Wait information

A caller who does not know how long they will wait hangs up sooner. Honest information about queue position or estimated wait — if the system can calculate it reliably — can reduce abandonment in the 10–60 second group. An unrealistic promise has the opposite effect: saying "one minute" and keeping them three minutes annoys callers even more.

The callback offer

The most effective fix for abandonment on long waits is to give the caller the option not to wait: "Shall we keep your number and call you back?" This turns abandonment into a callback list — the caller is not lost, just served later. The condition: the callback must actually be made, at the agreed time. Otherwise abandonment falls but complaints rise.

Calling back abandoned calls

If the numbers of abandoned calls are kept, they can be called back — especially on sales lines. But before calling each number, check: has the customer already called again and been answered? Calling the same customer from two sides is confusion, not service. Timing matters too: a few hours later the customer may already have found a solution elsewhere.

Illustrative example: a medical centre

Not a real customer case. A medical centre's monthly abandonment is 18%. The split shows: 5% at 0–10 seconds, 4% at 10–60 seconds, 7% at 1–3 minutes, 2% over 3 minutes. The largest group is 1–3 minutes, clustered between 9:00 and 11:00 — the booking peak. The cause of the 5% in the first 10 seconds turns out to be a 25-second greeting.

Actions: the greeting is cut to one sentence; for the morning peak, simple booking requests are taken by an AI agent and complex calls are passed to reception; after two minutes a callback is offered. Results are tracked separately for each group.

Abandonment and service level together

Abandonment read on its own can mislead. For example, when a callback offer is added, abandonment falls but the service level does not change — callers are simply waiting in a different form. Or the opposite: the service level improves but abandonment stays the same, because the problem is in the first 10 seconds — the long greeting. So always show abandonment alongside the service level (the share of calls answered within a set number of seconds) and the repeat-call rate.

A weekly abandonment review

Fifteen minutes once a week: the share of each of the four wait-time groups, the worst three hours, and the effect of last week's change. Look at the numbers of a few abandoned calls — did those customers call again later and get served? That shows whether abandonment is a real loss, and helps choose the next step.

Common mistakes

• Not splitting abandonment by wait time
• Looking at the weekly average and missing the peak hours
• Promising a callback and not making it
• Letting out-of-hours calls into the queue
• Not checking complaints and repeat calls when abandonment falls

Limits

The definition of abandonment can differ by telephony platform, and comparison only makes sense within the same definition. Some abandonment is unavoidable: wrong numbers, callers changing their mind. Set the target against your own baseline, not zero. Calling back abandoned calls must take customer consent and local rules into account.

Reducing abandonment with Vexvon

In Vexvon AI Call Center a call is answered by the AI agent without hold music, and even when several calls arrive at once nobody hears a busy tone (the parallel-call ceiling depends on the plan). Simple questions finish with the agent, a difficult call is passed to a live operator, and for a night-time call the caller's number and request are on the team's panel in the morning. So the "long wait" and "out of hours" abandonment groups fall directly. Switching the agent on only for certain peak hours is possible with routing rules.

The general queue diagnosis is covered in reduce call queue time, and scheduling callbacks in callback scheduling.

First step

Split last month's abandoned calls into the four wait-time groups and see which hours the largest group clusters in. Start there. More articles are in the call center operations section; to run the analysis together, get in touch.