Blockage Rate

Blockage rate is the share of inbound calls that never reach your queue because no trunk, port or licence was free. The caller hears a busy tone. The call never hits your ACD, so it never shows up in any report you read.

That invisibility is the whole problem. Abandonment, service level and average speed of answer all count calls your switch actually saw. A blocked call was never seen, so every one of those numbers stays clean.

Two other OA glossary entries measure this same thing under different names, and both are linked in Related terms below. This entry covers what they leave out: why blockage is a capacity fault, and why hiring more agents will not fix it.

Key takeaways

• A blocked call never enters the ACD, so it is missing from abandonment, service level and volume reports.
• Blockage is a capacity fault. Trunks, ports and concurrency licences cause it, not headcount.
• Busy-hour demand sets the number of lines you need, never the daily average.
• Erlang B is the standard model for sizing a trunk group against a target blocking probability.

How it works

A call is blocked when the path into your contact center is full before the switch can queue it. No trunk, no session licence, no free port. The carrier returns a busy tone and the attempt disappears.

The arithmetic is plain. Divide blocked call attempts by total attempts offered across the same window — then read the answer as a percentage.

Sourcing the numerator is the hard part. Your ACD cannot supply it, because it only logs calls that arrived. Blocked attempts live in carrier records, session border controller logs and platform rejection counters.

That is why so many teams quietly report a blockage rate of zero. Zero is what you get when nothing is counting.

Capacity planners size trunk groups with Erlang B — one of the traffic formulas built on the Erlang unit of call traffic. Erlang B assumes a blocked caller gives up rather than retrying.

Feed it two inputs and it does the rest. You supply offered traffic in Erlangs for the busy hour and the number of circuits available. It returns the probability that a call finds every circuit taken.

The trade-off is money against reach. Each extra circuit costs something and cuts the blocking probability by less than the one before it, so the curve flattens fast.

Size against the busy hour, never the daily mean. A line group that looks half empty across a shift can still hit the ceiling for twenty minutes after a broadcast, an outage notice or a billing run.

Cloud telephony has softened this, not solved it. Physical trunks gave way to elastic session paths, which is a real gain for anyone whose peaks arrive without warning.

The cap moved rather than vanished. Contracts still limit concurrent sessions, agent seats and API channels — hit that limit and the platform rejects the call exactly as a full trunk group did.

One diagnostic beats all the others. Watch the gap between attempts offered at the carrier and calls in queue inside the platform.

When offered volume climbs and queue entries flatten, your ceiling is capacity. When both climb together and waits stretch, the fault is staffing or routing instead.

Examples

Blockage turns up in three familiar shapes: a demand spike the lines were never sized for, an outbound dialler eating shared capacity, and a cloud contract whose concurrency cap nobody read. Each one hides behind healthy dashboards.

A retailer emails its whole list at 9am. Calls triple for twenty minutes against a trunk group sized for the daily average, and the surplus callers hear a busy tone.

Nothing in the service report moves. Abandonment holds steady, service level barely dips, and the team closes the day pleased with itself.

Marketing then judges the campaign on the responses it can see. The people who heard a busy tone and bought elsewhere are counted nowhere — not in the campaign report, not in the service report.

The second shape is a shared trunk group. An outbound dialler and the inbound service line draw from the same pool, and every dial the dialler places is a path an inbound customer cannot use.

Regulation shapes the outbound half of that picture.

16 CFR part 310 treats a call as abandoned when a person answers and the telemarketer fails to connect them to a sales representative within two seconds of the greeting.

The safe harbour in the Federal Trade Commission’s Telemarketing Sales Rule protects a seller whose technology keeps abandonment at no more than 3 percent of calls answered by a person.

That 3 percent is measured across a single campaign running under 30 days, or separately over each successive 30-day period. The same rule says to let the phone ring at least 15 seconds or four rings before disconnecting.

Notice the asymmetry. Regulators put a hard ceiling on outbound calls the dialler drops, while inbound calls your own capacity refuses carry no equivalent cap.

The third shape follows a cloud migration. Complaints about busy signals filter back, the operations lead reads them as understaffing, and a hiring round begins.

That is an expensive misdiagnosis. The US Bureau of Labor Statistics put the median hourly wage for customer service representatives at $20.59 in May 2024.

The same handbook projects employment in the role to decline 5 percent from 2024 to 2034, with about 341,700 openings a year on average over the decade. Hiring is neither cheap nor quick.

None of that spending opens a single extra line. If every session licence is already in use, your newest agent sits idle while customers still hear a busy tone.

Related terms

These five entries sit closest to blockage rate. Two of them measure the same quantity under different names, which is worth knowing before you write a metric definition that your reporting team then has to maintain for years.

• Calls Blocked in Percent: the same measure carried under an older reporting label.
• Percent Blocked Calls: another name for the blocked share of offered calls.
• Busy Hour: the sixty minutes of peak traffic that capacity gets sized against.
• Erlang Models: the traffic formulas, Erlang B among them, used to size lines and agents.
• Queue Management: the routing and waiting rules that apply once a call clears capacity.

FAQ

What counts as a good blockage rate?

There is no regulatory floor for inbound blockage, so the target is whatever you agree with your carrier and write into the contract. Treat any measurable blockage in the busy hour as a fault worth fixing, since a blocked caller receives no service at all.

Why doesn’t blockage rate show up in my ACD?

Because the call never reached the ACD. Blocked attempts stop at the carrier, the gateway or the session licence, so you have to pull them from those logs and compare them against offered volume yourself.

Is blockage rate the same as abandonment rate?

No. Abandonment counts callers who got into the queue and then hung up, while blockage counts callers who were refused entry before the queue existed for them.

Does moving to cloud telephony remove blockage?

It reduces it without removing it. Elastic session paths absorb peaks that fixed trunks could not, but concurrency caps in the contract still reject calls once the licensed session count runs out.

Will hiring more agents lower blockage rate?

No, because extra agents cannot answer callers who never reach a line — fix the capacity ceiling first.

Compare delivery partners who publish their capacity and trunk sizing in the Outsource Accelerator hubs.