Every wholesale voice route is sold on two things: a price and a promise about quality. The price is easy to compare. The promise is where buyers get caught, because "high quality" means nothing until it is attached to a number.
Four metrics do most of the work. None of them is meaningful alone, and each one can be gamed in isolation, which is exactly why they are worth understanding together.
Key takeaways
- ASR tells you how many calls connect, but says nothing about whether a human answered.
- ACD is the honesty check on ASR: high ASR with low ACD is the signature of false answer supervision.
- PDD is mostly a measure of how many carriers a call passed through before reaching its destination.
- NER separates network failures from ordinary human behaviour like a busy line or an unanswered phone.
- Compare metrics only across the same destination and the same traffic type. Anything else is noise.
ASR: Answer-Seizure Ratio
ASR is the percentage of call attempts that result in an answered call. Send a thousand calls, get four hundred answered, and the route reports 40% ASR.
It is the headline number on every rate deck, and the most misleading one in isolation, for two reasons.
The first is that ASR is dominated by factors that have nothing to do with the carrier. Cold outbound dialling to a market where nobody recognises the number produces low ASR on a perfect route. Inbound traffic to a busy support line produces high ASR on a mediocre one. Comparing ASR across different traffic types tells you about the traffic, not the route.
The second is that ASR is trivially inflated. A carrier that answers calls before they reach the destination (false answer supervision) reports excellent ASR while delivering silence or a recorded message to the caller.
ACD: Average Call Duration
ACD is the mean length of answered calls, and its main job is to keep ASR honest.
Real conversations have a characteristic length for a given traffic type. When a route's ACD drops well below what the same traffic achieves elsewhere, the calls are connecting but not working: poor audio, one-way audio, or an answer event that fired before any human picked up.
The combination is what matters:
| Pattern | Likely meaning | What to do |
|---|---|---|
| High ASR, healthy ACD | The route is genuinely working | Nothing, this is the target |
| High ASR, very low ACD | False answer supervision, or bad audio on connect | Test manually, then move the traffic |
| Low ASR, healthy ACD | Calls struggle to connect, but succeed when they do | Usually traffic or destination, not the carrier |
| Low ASR, low ACD | The route is failing in both directions | Remove it from rotation |
PDD: Post-Dial Delay
PDD is the gap between the caller finishing dialling and hearing ringback. It is the most immediately human of the four metrics, because it is the one a caller actually experiences as the phone apparently doing nothing.
Long PDD is usually a routing topology problem rather than a capacity problem. Every carrier a call passes through on its way to the terminating network adds a routing decision and a signalling round trip. A call that changes hands four times before delivery accumulates four sets of that delay.
This makes PDD a useful proxy for something a rate deck will never tell you directly: how many times the route you are buying has already been resold.
NER: Network Effectiveness Ratio
NER exists because ASR punishes carriers for things that are not their fault.
A call that reaches the destination network and gets a busy signal is a successful delivery. So is a call that rings out unanswered. Both count against ASR, and neither indicates anything wrong with the route.
NER counts a call as effective if the network delivered it, including when the outcome was busy, no answer, or an explicit rejection by the called party. It excludes only genuine network failures: congestion, unavailable circuits, and routing errors.
The practical use of NER is diagnostic. When ASR falls and NER holds steady, the network is fine and something changed about the traffic. When ASR and NER fall together, the route itself is degrading.
Reading them together
None of these numbers means anything without a comparison. The same route, carrying the same traffic type, to the same destination, over time. That is the only comparison that isolates the variable you care about.
Two habits make the metrics useful rather than decorative:
- Segment by destination before you draw any conclusion. Blended, all-destination averages hide exactly the problems worth finding.
- Watch the trend, not the snapshot. A route that has quietly drifted from 38% to 24% ASR over three weeks is a more urgent problem than one that has sat at 30% since the day you turned it up.
Where reporting is only available as a daily or weekly export, the drift in point two is very difficult to catch. Real-time, per-destination visibility is what makes these four numbers actionable rather than retrospective.