ASR, ACD and PDD Explained: Reading a Wholesale Voice Quality Report

Wholesale voice quality is argued about in three acronyms. Understanding what each one actually measures, and how each can be made to look better than it is, is most of what you need to evaluate a supplier.

ASR: answer-seizure ratio

ASR is the proportion of call attempts that are answered, expressed as a percentage of attempts. It is the headline number on every quality report and the easiest one to misread.

A low ASR can mean the route is poor. It can equally mean your traffic is outbound cold-calling into a market that does not answer unknown numbers, or that you are dialling at the wrong hour, or that your list quality is bad. ASR is only comparable between routes when the traffic is comparable, which is why a supplier quoting a global ASR figure without asking about your traffic is quoting a number that means nothing.

What a genuinely bad ASR looks like: a sharp drop on one destination while your other destinations hold steady, or a consistent gap against a second supplier carrying the same traffic profile.

ACD: average call duration

ACD is the mean length of answered calls. It is the audio-quality proxy: users hang up on calls they cannot hear.

A short ACD on a route carrying conversational traffic is the classic signature of transcoding problems, packet loss or excessive latency. If your ASR is fine and your ACD has collapsed, people are answering and then leaving, which is a media problem, not a signalling one.

Watch for: a cluster of answered calls under five seconds. That is not a conversation, it is someone hearing nothing and hanging up. Some suppliers include these in ASR, which flatters the answer rate while ACD quietly tells the truth.

PDD: post-dial delay

PDD is the time between the call being placed and ringback being heard. It is the metric most often left out of a quality discussion and frequently the one users actually notice.

Long PDD means the caller hears silence and, past three or four seconds, starts assuming the call has failed. They hang up before the destination has finished ringing. That shows up in your reports as a low ASR, which sends you looking at the wrong problem.

High PDD usually means too many hops. Every intermediate carrier adds signalling latency, so a route being resold three times over will have PDD that no amount of media optimisation fixes.

How routes are made to look better than they are

• Cherry-picked reporting windows. Off-peak numbers presented as typical. Always ask for the peak-hour figure for the destination.
• Short calls counted as answers. Inflates ASR while ACD absorbs the damage. Ask for the distribution, not the mean.
• Aggregated destinations. A regional average hides a specific destination performing badly. Ask per destination.
• Excluding failed attempts. Some reports quietly drop certain SIP failure codes from the denominator.

Reading the three together

The combination tells you where the problem is:

• Low ASR, normal ACD, normal PDD: probably your traffic or your list, not the route.
• Normal ASR, low ACD: media quality. Transcoding, loss, or one-way audio.
• Low ASR, high PDD: callers giving up during setup. Too many hops.
• All three degrading at the same hour each day: congestion, either on the route or on your own transport.

That last one is worth isolating carefully, because it is as often the customer’s own link as the terminating route. If quality tracks time of day rather than destination, look at your transport before you blame the carrier, that is what a managed tunnel is for.

What to ask for

Per-destination ASR, ACD and PDD, at peak hour, on your own traffic, in a format you can export. Anything less specific is marketing. Our wholesale termination puts all three into your own CDRs so the rate sheet can be checked against reality.