What is call abandonment rate?

Call abandonment rate is the percentage of contacts that disconnect before reaching an agent or completing what they called to do. A customer waits in the queue, decides it's taking too long, and hangs up.

It's the metric that punishes understaffing the fastest. A queue can look fine on paper right up until wait times cross the point where people give up, and then abandonment rate moves before almost anything else does.

This page covers how to calculate abandonment rate, what should and shouldn't count as abandoned, what a credible benchmark looks like, how it relates to average speed of answer and service level, and what to do when it's rising.

Abandonment rate in one sentence

Abandonment rate is the share of customers who give up before anyone answers.

How to calculate abandonment rate

Abandonment rate = (abandoned contacts ÷ total contacts offered) × 100

A worked example. In a day, 2,000 contacts enter the queue. 1,860 are answered and 140 hang up before an agent picks up.

Abandonment rate = (140 ÷ 2,000) × 100 = 7%

Most platforms apply a short grace period, commonly 10 seconds, before a hang-up counts as a real abandonment. A caller who dials the wrong number and hangs up in two seconds isn't reporting anything about your queue, and counting them muddies a number that's otherwise a clean read on staffing.

What counts as abandoned, and what doesn't

Counted: any contact that enters the queue and disconnects, past the grace period, before an agent answers.

Not counted, in most implementations: contacts that never reached the queue at all, because the system was full and refused them outright. Those are blocked calls, a separate and often worse problem, and folding them into abandonment rate understates how many customers failed to get through.

Also worth separating: callback requests. A customer who accepts a callback isn't abandoning, but some reporting setups miscount them that way if the callback flow isn't wired into the same queue logic as a live wait.

What's a good abandonment rate

The most credible figure available comes from ACXPA's Australian Call Centre Rankings, which measures abandonment through an ongoing mystery-shopping program rather than self-reported figures. Its 2024 Contact Centre Best Practice Report, drawing on more than 300 Australian contact centres, put the national voice-channel average at 8%, up from 7% in 2022 and down slightly from a peak of 9% in 2023. Synchronous channels (voice and live chat) averaged 7% abandonment; asynchronous channels (email and messaging) averaged 3%, which tracks, since nobody "abandons" an email the way they hang up a phone.

SQM Group, a customer-experience benchmarking firm, puts the commonly cited working target lower: under 5% as generally acceptable, and 3% or lower among centers with the strongest CSAT scores. That range comes with less published methodology than the ACXPA figures, so treat it as directional, the way you would any number quoted without a named study behind it.

Both figures point the same direction: a queue abandoning fewer than 1 in 20 contacts is doing well by current standards, and the double-digit rates some vendor blogs cite as "typical" describe a bad quarter, not an average one.

Abandonment rate vs. average speed of answer

These two move together and hide each other's problems. Average speed of answer (ASA) only averages the contacts that were answered, so every customer who abandons drops out of that calculation entirely. A badly understaffed hour can post a flattering ASA precisely because the people who waited longest gave up and stopped counting.

Which means the two need to be read side by side. A falling ASA next to a rising abandonment rate is the same staffing shortfall showing up in two places, with one of the two numbers hiding it.

Abandonment rate vs. service level

Service level (the share of contacts answered within a set threshold, like 80% in 20 seconds) is normally the reported commitment, and service level targets are usually set assuming a certain amount of abandonment happens along the way. If abandonment rate climbs, service level looks worse for a reason that has nothing to do with how fast agents are answering, because fewer of the contacts that would have blown past the threshold are still in the count by the time an agent picks up.

Why abandonment rate matters

An abandoned contact almost never means the issue is resolved. The customer's problem is still there after they hang up. Most either call back later, adding a second contact to the workload that the first one should have covered, or give up on the channel entirely and stay unresolved.

It moves before almost anything else does. Abandonment reacts to a staffing gap in minutes, while satisfaction scores and churn take days or weeks to reflect the same problem. Watching abandonment by interval catches a staffing miss while it's still small.

For ecommerce, the timing compounds the cost. A where's-my-order call abandoned during a peak shipping week doesn't just disappear. It comes back as a chat, an email, and sometimes a chargeback dispute, all during the exact week that contact volume is already stretched thinnest.

How to reduce abandonment rate

Offer an accurate wait estimate or a callback. Customers who know how long they're waiting, or who can trade the wait for a scheduled callback, abandon less even when the underlying wait doesn't change.

Staff to the interval. Abandonment spikes are usually concentrated in specific half-hour windows where volume outran staffing. A daily average can look healthy while those windows are quietly bleeding customers.

Resolve more before the queue. Self-service that genuinely answers common questions, order-status lookups and return initiation in particular, keeps those contacts from ever entering a queue where they could abandon.

Watch the interval right before your worst abandonment windows. A spike at 10 minutes past the hour often traces back to a staffing gap that started 5 to 10 minutes earlier. Fixing the trigger is more durable than reacting to the spike itself.

Abandonment rate when AI answers first

When automation picks up every contact immediately, voice-channel abandonment can look like it's dropped to nearly zero, because there's no live queue for a customer to give up on. That number is real but incomplete: it says nothing about customers who abandon during the AI conversation itself, before getting the resolution or the human handoff they wanted.

The metric worth tracking in an AI-first setup is escalation abandonment: the share of customers who ask for a person, get placed in a queue for one, and give up before that handoff completes. Those customers have already spent time with the assistant, so their patience for a second wait runs thinner than a first-time caller's, and a queue that treats escalation abandonment as a rounding error will miss exactly the contacts most likely to end in a complaint.