Many carriers publish one quality number for their whole network. It looks reassuring and tells you almost nothing about the calls you are about to send. Quality is a property of a destination and a route, not of a company.
The mix problem
Network averages are weighted by volume. If most of a carrier’s traffic goes to a few well-run destinations, those destinations decide the headline figure. A route that performs badly but carries little traffic barely moves it.
| Destination | Attempts | Answered | ASR |
|---|---|---|---|
| Destination A | 900,000 | 432,000 | 48% |
| Destination B | 100,000 | 18,000 | 18% |
| Network total | 1,000,000 | 450,000 | 45% |
The network total of 45% is close to destination A. If your campaign is aimed at destination B, the figure you were shown is more than double what you will get.
What to ask for instead
- ASR and ACD for the specific destinations you plan to dial, ideally split by mobile and fixed.
- The time window the figures cover. A week of data is more useful than a quarter.
- Post-dial delay, because long PDD pushes callers to hang up and lowers ASR.
- Whether caller ID was delivered on the measured calls. Good ASR on a route that drops CLI is not a CLI route.
You can work these out from your own CDRs with the ASR, ACD & NER calculator. The definitions are explained in ASR, ACD, PDD and NER.
Your traffic changes the number too
ASR also depends on what you dial. A fresh, consented list produces a higher answer rate than an old one on the same route. That is why we suggest a route acceptance test with a sample of your real traffic before you commit volume, rather than relying on any published figure, including ours.