Every predictive dialer makes the same bet thousands of times an hour: it dials more numbers than it has free agents, gambling that by the time someone answers, an agent will be ready. Win the bet and your agents stay busy. Lose it and a live person picks up to silence, then hangs up. That silent hang-up is an abandoned call, and it is both a wasted contact and, past a certain rate, a compliance problem. Pacing is how good auto dialer software keeps that bet in your favor.
Let me walk through how pacing works, what pushes abandonment up, and how to tune the dial without turning your campaign into a liability.
What pacing really means
Pacing is the ratio of calls placed to agents available, and the dialer adjusts it continuously. If you have 10 agents and the dialer places 13 calls, it is pacing at 1.3. The logic behind that number is statistical: most dialed calls will not connect to a live person at all. They hit voicemail, no answer, or a busy tone. So dialing one-to-one would leave agents idle most of the time, waiting through dead calls.
The dialer watches your live answer rate and average call length, then sets the pace so that, on average, an agent comes free just as a call connects. When it works, the customer hears an agent immediately and the agent never waits. The whole thing rides on prediction, which is exactly why it is called predictive dialing.
Why calls get abandoned
Abandonment happens when the prediction overshoots. The dialer expected an agent to be free, the call connected first, and there was nobody to take it. A few things make this worse:
- Pacing set too aggressively. Push the ratio too high to squeeze out idle time and you will connect more calls than you can staff.
- Short, unpredictable call lengths. When handle times swing wildly, the dialer’s forecast gets noisy and misfires more often.
- Small agent pools. With only a handful of agents, one or two long calls throw the whole average off, so small teams see more volatility.
- Sudden list quality changes. A fresh list with a high live-answer rate behaves nothing like a tired one, and a fixed pace built for the old list will over-dial the new one.
That small-team point is the one people underestimate. Predictive dialing shines at scale and gets twitchy with tiny agent counts. If you are running five agents, a preview or progressive mode is often the saner choice.
The compliance line you cannot cross
Abandonment is not just an efficiency metric. Many jurisdictions cap how many answered calls a campaign may abandon, commonly around three percent measured over a period, and they require a recorded message when a call is abandoned so the person is not left in silence. Cross that line and the efficiency you gained is wiped out by the risk you took on.
So the real objective is not zero abandonment, which would mean dialing one-to-one and wasting agent time. It is keeping abandonment under the legal cap while keeping agents productive. The two goals fight each other, and pacing is the dial that balances them. If you want the broader regulatory picture, the dialer should also enforce calling windows and suppression, not just pacing.
Tuning the dial without getting burned
Good pacing is mostly about giving the dialer good information and sensible limits. A few practical moves:
- Set a hard abandonment ceiling below the legal cap and let the dialer throttle itself to stay under it. Treat the legal number as a wall, not a target.
- Let the system adapt by campaign rather than running one global pace. A collections list and a survey list do not behave the same.
- Always attach a compliant abandoned-call message, so the rare miss is handled gracefully instead of as dead air.
- Watch the rolling abandonment rate during the campaign, not just in the post-mortem. By the time the report lands, the damage is done.
This adaptive, per-campaign approach is how ICTDialer handles predictive dialing. Built on the FreeSWITCH telephony platform, it adjusts pace against live answer behavior while holding to the abandonment ceiling you set, so the safe pace and the productive pace end up being the same one.
A quick example of the payoff: a 12-agent outbound team running a flat, aggressive pace was abandoning around five percent and getting numbers flagged. Switching to an adaptive pace capped just under the legal limit pulled abandonment down to roughly two percent. Agent idle time barely moved, because the dialer was simply predicting better, not dialing slower across the board.
Frequently asked questions
What is a good abandonment rate for a predictive dialer?
Aim to stay comfortably below the legal cap that applies to you, which is often around three percent of answered calls. Setting your own ceiling a point or so under the limit gives you a safety margin against normal variation.
How is predictive dialing different from preview or progressive dialing?
Predictive dialing places more calls than there are free agents and predicts availability, maximizing talk time at scale. Preview and progressive modes dial one call per available agent, trading some efficiency for near-zero abandonment, which suits smaller teams and high-value lists.
Does aggressive pacing always mean more productivity?
No. Past a point, a higher pace just converts saved idle time into abandoned calls and flagged numbers. The productive setting is the one that keeps agents busy while holding abandonment under your cap, not the highest ratio you can force.
Can a small team use a predictive dialer?
It can, but predictive mode gets volatile with few agents because one long call skews the forecast. Small teams often get steadier results from progressive dialing until the agent pool is large enough for prediction to settle.
Is open source auto dialer software able to handle compliant pacing?
Yes. Open platforms like ICTDialer enforce abandonment ceilings, attach compliant abandoned-call messages, and adapt pace per campaign, so being open source has no bearing on whether pacing stays within the rules.