Nearly every hold-time statistic online traces back to a vendor blog citing another vendor blog.
There is one durable industry convention worth knowing, and a great deal of circular sourcing around it.
The one figure in this area with real staying power is not a research finding at all; it is a target. The 80/20 service level, answer 80% of calls within 20 seconds, has been the standard contact centre benchmark for decades. Beyond it, the numbers circulating about how long callers tolerate hold before abandoning come overwhelmingly from vendor-published aggregates that cite one another, with methodology rarely stated. They are not necessarily wrong. They are not evidence either.
What is the 80/20 service level?
A target, not a measurement: answer 80% of incoming calls within 20 seconds. It has been the default contact centre goal for long enough to function as a shared reference point, which is exactly what makes it useful even though no study established it.
Its value is as a common unit. Two businesses discussing whether their phone is working can compare against 80/20 and mean the same thing, which is more than can be said for "our answer rate is good".
Its limitation is that it was designed for staffed queues in call centres, where holding is the expected experience. A four-person clinic does not have a queue. Its calls either get answered or go to voicemail, and 80/20 has nothing to say about the second outcome.
Why are the abandonment statistics weak?
Because most trace to vendor content marketing rather than to research. Follow a widely quoted abandonment percentage back through its citations and it usually terminates at another blog post, or at a vendor's own customer data with no methodology attached.
Vendor aggregates are not worthless: a platform reporting on its own traffic is describing something real. But the population is its own customers, which is a selected group, and the definitions are frequently unstated. Many contact centres exclude very short calls from abandonment calculations to avoid counting misdials, and a benchmark that does not say whether it did is not comparable to one that did.
This is the same pattern as missed-call statistics generally, and it is why we publish our own call data with the arithmetic shown. It is a small dataset, and it has a traceable origin, which we would rather have than a large number with none.
What should a small business measure instead?
Its own numbers, against its own definition. Three quantities are worth having and all three are usually available from a phone provider: how many calls went unanswered, how long the answered ones took to pick up, and how many callers hung up before anyone reached them.
One caution on the third: excluding sub-ten-second calls is standard practice and it is also a place where a number can be quietly improved by moving the threshold. Whatever you pick, keep it fixed, or the trend becomes meaningless.
- Unanswered calls, by hour of day. The pattern matters more than the total.
- Time to answer on the calls that were answered.
- Calls that ended before connection, excluding very short ones that are likely misdials.
- Whether any of the above changes on days you are busiest, which is when it costs most.
Follow almost any published call-abandonment statistic back through its citations and it terminates at another blog post, which is a good reason to measure your own phone rather than benchmark it against a number nobody can source.
Your own hold and abandonment figures are usually already sitting in your phone provider's reporting. They are worth more than any benchmark.
Start free pilotWhat this page will not do
It will not repeat abandonment percentages it could not trace to a stated methodology.
- No abandonment benchmark quoted. Widely circulated figures exist. None that we could trace stated its methodology clearly enough to be worth republishing, so this page names the convention and stops.
- 80/20 is a target, not a finding. It is an industry convention of long standing. No study established it as the point at which callers abandon.
- No first-party hold data. Our own lines answer rather than queue, so we have no hold-time data to publish.
Questions this raises
- How long will callers wait on hold before hanging up?
- There is no well-sourced general answer. Figures circulate widely but most trace to vendor blogs citing each other rather than to research with a stated methodology.
- What is the 80/20 rule in call centres?
- A service level target: answer 80% of calls within 20 seconds. It is a long-standing industry convention rather than a research finding, and it was designed for staffed queues.
- What is a good call abandonment rate?
- Commonly cited healthy ranges sit in the low single digits, but the sourcing behind those numbers is weak and definitions vary. Your own trend against a fixed definition is more useful than any published figure.
- Does the 80/20 target apply to a small business?
- Only loosely. It assumes a queue callers wait in. A small business without one has calls that are answered or not, and voicemail is the outcome 80/20 does not describe.
- Where do I find my own numbers?
- Most phone providers report unanswered calls, time to answer and calls ended before connection. Fix your definitions once and watch the trend rather than the absolute value.
Sources
- Lead Response Management Study (Dr. James Oldroyd)MIT / InsideSales.com2007, consulted 2026-08-09
Want an AI receptionist built around these rules rather than despite them?
We build a working receptionist on your calls, your booking rules and your escalation rules, and you call it yourself before committing to anything. About 14 days, no call limit during the pilot.
Start free pilot