You run an outbound floor. The dialer connects, your agent says "hello," and there is nothing on the other end. Do that forty times an hour across thirty seats and you are paying for silence and burning out agents who spend half a shift talking to no one.
So you scrub the list again. Fresh data, tighter filters, new numbers. The dead air is back inside a day. It comes back because the list was rarely the cause. The route was.
This post is for the owner or floor manager staring at a dead-air rate north of 30% who keeps getting told it is a data problem. We will cover what dead air actually is, why it lives in the carrier route instead of your list, and how to prove it in your own call detail records in about an hour.
What "dead air" actually is
Dead air is a call that connects, so you get billed for it, with no live audio or no live human on the far end. There are two flavors and they get blamed on each other constantly.
The first is technical. The signaling went through, the call "answered," but the audio path never formed. This is the classic one-way audio and firewall problem: the instructions to ring the phone arrived, the actual voice packets did not. That one is on your side of the wire most of the time and it is fixable in an afternoon.
The second flavor is the one that survives every list scrub you throw at it. The route tells your dialer the call was answered before a real person ever picked up.
False answer supervision, in plain English
False answer supervision, or FAS, is when the answer signal for a call gets sent before, or without, a real connection. Your billing starts on silence. The more aggressive variant routes your call into a recording: a fake ringback, then a message, anything to keep the line open and the meter running while your agent waits for a human who is never coming.
Here is the part operators miss, and it is the whole point of this post: FAS happens on perfectly valid numbers. It is generated by the route that carries your call, not by the number you dialed. That is why scrubbing does not fix it. You can remove every disconnected and flagged number in your database and a clean route will still be clean while a bad one keeps manufacturing false answers on good cell numbers.
Why cheap routes manufacture dead air
Most dead air traces back to how your minutes get carried. A lot of carriers protect their margin by sending your traffic down whichever downstream path is cheapest at that moment. Your call lands on whatever was cheapest that second, and you inherit that path's behavior. Some paths return false answers on a slice of their terminated traffic. You never see it on a single test call. You see it when the day's numbers come back and a third of your "connects" lasted one second.
The fix is not another list scrub. It is a route chosen on whether calls actually complete, not on how cheap the minute looks on a rate sheet. And that is not something you should take on faith from any carrier, ours included. It is something you measure on your own traffic, which is the only proof that counts.
Prove it's the route, not the list
You do not have to take anyone's word for this, including ours. Three checks, all in the CDRs you already have:
- Chart average call duration. A cluster of calls that "answered" and ended in zero to one seconds is your answer-and-drop population. Real conversations do not average one second. If that cluster is large, you have a route problem, not a data problem.
- Run the same list across two routes. Take one scrubbed list, split it, send half through your current route and half through a test route. If the dead air is the list, both halves look the same. If one comes back clean, the route was the cause and the list was the alibi.
- Watch post-dial delay. PDD over seven or eight seconds means the person hangs up before your agent is even bridged in. Under five seconds is the target. High PDD is its own form of dead air, and it is also a route characteristic.
None of this needs new software. It is sitting in your call detail records right now.
What to demand instead
There is nothing magic here, and you should not take any carrier's word for it, ours included. Demand the things you can actually verify:
- A route you can test on live traffic before you commit, so you judge it on completion instead of promises.
- CDRs you can pull yourself, so you can see the zero-to-one-second cluster if it is there.
- STIR/SHAKEN signing on every outbound call.
- Same-day turn-up, so a test costs you a morning, not a month.
- The option to buy clean, health-checked DIDs and termination on the same network, so origination and outbound are not stitched together across vendors who each blame the other.
- A human to call when something breaks.
What to do this week
- Pull one day of CDRs and chart average call duration. Flag the zero-to-one-second cluster and get your real dead-air number, not the dialer's optimistic one.
- Split your next campaign across your current route and one test route, same list on both.
- Compare dead-air rate, ACD, and PDD side by side at the end of the day.
- If the test route wins on the same data, you found your problem, and it was never the list.
Dead air is the most expensive number on an outbound floor that nobody puts on a report. It costs you talk time, it costs you agents, and you cannot scrub your way out of it because the thing generating it is not in your database. It is in the pipe. Fix the pipe.