ASR and ACD: what VoIP operators measure

4 min read

You open a chat with a supplier and the first numbers on the table are ASR and ACD. You open Aggregate on your own switch and the same two columns decide whether a terminator stays in the LCR or gets pulled. Those numbers are how wholesale and retail VoIP operators argue about route quality.

The formulas sit below. So does the “ACD number” mix-up that shows up in search, and where the numbers live when you run MOR or M4.

Rate file excerpt with Destination, Price, ASR, and ACD columns

Two formulas, two jobs

ASR (answer-seizure ratio) is the share of call attempts that get answered.

ASR % = (answered calls / total call attempts) × 100

Example: 156 attempts, 62 answered → ASR = (62 / 156) × 100 = 39.7%.

In telecom language, each attempt is a seizure. Busy, reject, and no-answer all count against ASR. ASR moves with user behavior, not only with network faults.

ACD (average call duration) is the average length of the calls that did answer.

ACD = total duration of answered calls / number of answered calls

Example: 62 answered calls, 145 minutes of answered talk time → ACD = 145 / 62 = 2.34 minutes.

On Kolmisoft, ACD uses billable seconds (billsec) of answered calls, then divides by the answered count. You average paid conversation time, not ringing time.

ACD is not an “ACD number”

Search often shows ACD number next to answer-seizure ratio. That phrase mixes two different industries.

Phrase Meaning in this article
ACD Average call duration: length of answered calls
ACD (contact centre) Automatic call distributor: a queue that hands calls to agents
ACD number Usually a queue or DID that feeds that distributor, not a softswitch quality metric

If you buy and sell minutes, you want average call duration. If you run a contact-centre queue, you want a different product conversation. This article is about the quality metric on voice routes.

What counts as “good”

Rough industry starting points (not SLAs):

Metric Rough floor Strong
ASR 40–50% 60%+ (some VoIP guides cite 60–70% as very good)
ACD about 3–5 minutes above ~6 minutes on many retail mixes

Treat those as a starting point, not a scoreboard. Destinations differ. A typical ASR to some mobile corridors is lower than a typical ASR to Western Europe. Answering machines, local habits, and network quality all move the number. Compare a terminator to itself last week and to other terminators on the same destination, not to one global number.

High ASR means more attempts become conversations. High ACD means those conversations last. Cheap minutes with a collapsed ASR still cost you retries, dialer burn, and a customer who measures you on answer rate.

Why your ASR and your supplier’s ASR disagree

One customer attempt can create several provider attempts under LCR.

Sketch: the customer places two calls. One answers, one fails. Customer ASR = 50%.

The failed call hunted two providers. Provider A handled the answered call (ASR 50% on that sample). Provider B only saw the failed hunt (ASR 0% on that sample).

Same human traffic, different denominators. When you argue ASR with a supplier, agree whether you are looking at originator attempts or terminator attempts. On MOR Aggregate you can view the situation from the user side when an LCR has more than one provider, so one user call can show as several recorded attempts.

What pulls the numbers down

ASR drops when lines are busy, the far end is congested, or callers hang up before answer. Long post-dial delay (PDD) makes people abandon. That shows up as failed attempts, not as short answered calls.

ACD drops when answered calls die early: congestion, looping, echo, jitter, delay, packet loss, weak codecs, too many media hops, or false answer supervision (FAS). FAS often pairs a high ASR with a short ACD: many “answers,” little real talk. That pattern is a fraud signal, not a quality win.

Traffic mix matters too. A dialer that burns short connects will print a lower ACD than a residential route on the same terminator. Read the profile before you drop the supplier.

Price-only LCR ignores both columns

Least cost routing sorts terminators by buy price. That protects margin on the price list. It does not see ASR or ACD. A cheap path with a bad answer rate still wins first.

If you already rank paths on live quality, read The cheapest route that kills your ASR. That post is the next step after you trust the numbers in this one.

Where to read ASR and ACD on Kolmisoft

On MOR (admin): Statistics → Calls → Aggregate. Group by terminator, provider, destination, or day. Turn on the ASR and ACD columns. Export CSV when you need to send evidence to a supplier.

Reseller PRO can use Aggregate when providers are assigned to terminators. A plain reseller login does not. End users do not get ASR as a labeled column; they see graphs and failed vs all calls under their personal call views.

On M4, Aggregate templates and auto-export sit under reports so you can mail terminator and originator cuts on a schedule. Wholesale volume and quality routing formulas belong on M4; retail, resellers, and PBX features belong on MOR. See Class 4 vs Class 5 if you are choosing which switch owns the traffic.

A simple weekly check

  1. Pick one destination that pays the bills.
  2. Group Aggregate by terminator for the last seven days.
  3. Sort by ASR, then glance at ACD.
  4. If the cheapest terminator is also the worst ASR, LCR did what you asked, and the ask is wrong.
  5. For short ACD with high ASR, open CDRs and test for FAS before you renegotiate price.

Carriers often write SLAs around ASR and ignore ACD. You still need both columns. ASR without duration can hide FAS. Duration without answer rate can hide a path nobody connects on.

NER (why you almost never see it)

Network effectiveness ratio (NER) strips user behavior out of the score: answers plus busy, ring-no-answer, and terminal rejects, over seizures. It is a cleaner network metric than ASR. The industry still quotes ASR in almost every interconnect chat, so you will live with ASR even when NER would be fairer. ITU operational guidance for international network management is where ASR sits as a defined term (Recommendation E.411).

FAQ

What does ASR mean in VoIP?

Answer-seizure ratio: answered calls divided by call attempts, as a percentage. It is the usual “did the call connect?” score between VoIP providers.

What is ACD in telecom?

Average call duration: total answered talk time divided by the number of answered calls. On Kolmisoft billing stats, that uses billable seconds of answered calls.

Is “ACD number” the same as average call duration?

No. “ACD number” usually points at a contact-centre automatic call distributor or a queue DID. Softswitch quality reporting means average call duration.

Why is my ASR different from my supplier’s?

One of your attempts can become several of theirs under failover and LCR. Agree on whose attempts sit in the denominator before you escalate.

What ASR should I demand from a supplier?

Start from your own destination history. Use 40–50% as a rough floor and 60%+ as strong on many mixes, then judge the drop against last week and against peer terminators on the same code.

Keep both columns open

ASR tells you whether attempts become conversations. ACD tells you whether those conversations last. Read them per destination and per terminator, not as one global grade. When the cheapest path loses on ASR, move from price-only LCR to a quality rule, and keep Aggregate open while you do it.

If you want the next step on the switch, open quality-based routing or the Class 4 / Class 5 comparison for which product owns the traffic.


Measuring ASR and ACD on your switch?

If you want Aggregate and quality routing set up so supplier numbers match what you see, our team can walk through the reports with you.

Contact us

4 Replies to “ASR and ACD: what VoIP operators measure”

    1. maybe since some first priority routes’ quality are not so good and has almost use up the callers trying time,still not connecting, then the call has so less trying time for connecting to the 2nd routes. it’s hard to connect within a very short pdd. some connects but may be considered not connecting by user. so user hang up or switchs terminated the call

  1. Question.
    If as a partner or user of Kolmisoft MOR, what should I tell my International Wholesales Carriers and my CSPs (directly connected to my facilities as a unique Hub) to do about an ASR between 40 and 50%, because my goal is to have it at 55% Minimum? Thanks for any useful idea.

Leave a Reply

Your email address will not be published. Required fields are marked *