Dialer capacity calculator

Outbound sizing usually gets done backwards. Somebody picks a number of lines because that is what the carrier quoted, turns the dialer up until agents stop complaining about idle time, and finds out months later that the campaign was pacing against a ceiling nobody wrote down. This runs it forwards instead: agents, handle time and your dial-to-contact ratio decide the dials, and the dials decide the lines.

Three answers come out. Dials per hour is the workload the dialer has to place. Concurrent lines is what you have to provision to place it. Live contacts per hour is what the floor should hand back, and it is the one to check against yesterday's report before you trust either of the other two.

Outbound capacity planner

Dialer capacity calculator

Start from the floor you have, not the one a vendor quotes. Enter your agent count and how long a contact actually takes, and this works out the dials per hour the campaign has to place, the concurrent lines that needs, and the live contacts it should return.

What this floor has to place

Dials per hour required

2,372

Concurrent lines needed

Projected live contacts per hour

118.6

Handle time per contact: 255 sec
Dials per agent per hour: 198
Contacts per agent per hour: 9.9
Average channels in use: 23.4
Lines per agent: 2.5

Pressure-test these numbers with us

How this is calculated

Talk plus wrap is what one contact costs an agent. Occupancy says how much of the paid hour is available to spend on that, which fixes the contacts one agent can carry; multiply by agents for the floor, and by your dial-to-contact ratio for the dials those contacts take. Lines are a separate question with a separate answer: a channel is busy while a dial rings and while a connected call is talking, and it is free during wrap.

handle_seconds  = talk_seconds + wrap_seconds
contacts_per_hr = agents x (3600 x occupancy / 100) / handle_seconds
dials_per_hr    = contacts_per_hr x dials_per_contact
line_seconds_hr = dials_per_hr x ring_seconds
                + contacts_per_hr x talk_seconds
lines_needed    = ceil(line_seconds_hr / 3600 x (1 + headroom / 100))

Note: every figure above is computed from the inputs you supply. The starting values are a placeholder to replace with your own dialer report, not a benchmark, an industry average or a claim about anyone's results. Lines needed is average demand plus the headroom you set -- it is a provisioning floor, not a guarantee that no dial ever queues.

How to read the result

Read contacts per hour first, not dials. It is the only output you can check against something that already happened: pull a day off your own dialer report, put that day's agent count and handle time into the form, and see whether the projection lands near what the floor actually produced. If it does not, the input that is wrong is almost always occupancy or the dial-to-contact ratio, and the calculator has just told you something more useful than a capacity number.

Dials per hour is a workload, not a target. Nobody should be managed against it. It exists so you can ask the dialer vendor a specific question -- can this configuration sustain that rate on this list -- instead of the vague one everybody asks, which is whether the system is fast enough.

Lines per agent, in the small print under the headline figures, is the number worth remembering. It is the ratio that tells you whether a quote is priced for the campaign you are actually running, and it moves far more with talk time than with agent count. Halve the talk time and the same floor needs roughly twice the dials and close to twice the lines.

Occupancy is the input that lies

Every other field on the form is something you can measure. Occupancy is something you choose, and choosing it high is how a capacity plan turns into a retention problem. It is the share of the paid hour an agent spends on a contact or wrapping one up, so the remainder is not waste -- it is breaks, coaching, system time, and the pause between a call ending and the next one arriving.

Push it toward the top of the slider and the arithmetic obligingly tells you that fewer agents can carry the same campaign. What it cannot tell you is that the same setting, held for a quarter, is the standard way a floor loses the people who were good at the job. Treat a high occupancy figure the way you would treat a high utilisation figure for a crew: fine for a storm week, not a staffing model.

The honest way to use the field is to run it twice. Once at the occupancy you are getting today, which gives you the floor you actually have, and once at the occupancy you would be willing to defend, which gives you the floor you would need to hire toward.

Why the line count has a ceiling as well as a floor

This tool gives you a provisioning floor: enough channels to carry the average traffic your dial rate produces, plus whatever headroom you set. It deliberately does not tell you to buy more than that, because on a predictive campaign extra lines are not free capacity. They are extra calls placed ahead of agents who are not yet ready, and the person who answers one of those hears silence.

That is regulated, and the rule is specific rather than a matter of taste. Under the FTC Telemarketing Sales Rule, a call is abandoned if a person answers and is not connected to a live representative within 2 seconds of their completed greeting; the safe harbor caps abandonment at 3% of calls answered by a person and requires at least 15 seconds or 4 rings before disconnecting.

Quoted from FTC Telemarketing Sales Rule, 16 CFR 310.4(b)(1)(iv) and 310.4(b)(4), https://www.ecfr.gov/current/title-16/chapter-I/subchapter-C/part-310/section-310.4. Read it as the reason line count and pacing are one decision and not two: the channels you provision set the dialer's room to over-dial, and the over-dial is what produces the abandoned call. This is a description of published rules, not legal advice, and the calculator does not model pacing -- ask whoever runs your dialer what the campaign's measured abandonment rate is, and treat a vendor who cannot answer that as having answered it.

Where the dial-to-contact ratio comes from

It is the input people guess at, and it is the one that swings the answer hardest -- everything downstream of contacts per hour scales with it linearly. Do not take a number from a forum. Take total dials and total live contacts off one week of your own dialer export and divide.

One thing worth knowing before you do: the ratio is not a constant across a list, because the chance of reaching a person changes with which attempt it is. We published what that looks like on our own dialer history -- across attempts 1 to 10 the human-answered rate runs 5.443%, 5.663%, 5.070%, 4.640%, 4.169%, 4.042%, 3.635%, 3.417%, 3.110% and 2.840%. Highest to lowest is a factor of 1.994, and the median campaign in each bucket tracks it: 6.546% at attempt 1, 6.970% at attempt 2 and 3.207% at attempt 10.

Source: apps/ccdocs-astro/src/data/dial-cadence-study.json (`byAttempt.emitted` and `spread.attempt`), generated by apps/ccdocs-astro/scripts/dial-cadence/dial_cadence_distribution.py, 2026-05-25/2026-07-26. The working consequence for capacity is that a floor dialing a fresh list and a floor grinding an aged one need different ratios in this form, so size them separately rather than blending a quarter into one average. The full dataset, including the hour-of-day cut and the buckets that were withheld, is at the outbound dial cadence benchmark.

Worked example, using the numbers the page loads with

The panel opens with twelve agents, a two-hundred-fifty-five-second handle time made of two hundred ten seconds of talk plus forty-five seconds of wrap, twenty dials per contact, seventy percent target occupancy, a twenty-five-second ring window and twenty-five percent line headroom. One agent at seventy percent occupancy and two hundred fifty-five seconds a contact carries about ten contacts an hour, so twelve agents produce roughly a hundred nineteen contacts an hour. At twenty dials per contact that is about twenty-three hundred seventy dials an hour -- close to two hundred per agent.

Every one of those dials ties up a line for its ring window, and every contact ties one up for the talk portion, which works out to about twenty-three concurrent lines on average. Add the twenty-five percent headroom and the floor should provision thirty lines -- two and a half per agent, not one each, because most of the channel demand is dials that never connect rather than the conversations that do.

Where this fits

Outbound call center -- what an outbound campaign involves beyond the dialer, if the answer here is that you need more floor than you have.
Build a call center -- the other side of the decision, once you know the seat and line count you would be standing up.
Call center cost estimator -- puts a monthly figure on the seats this calculator says the campaign needs.
Pricing -- our published rates, if renting the capacity is cheaper than provisioning it.

Bring us a week of dialer export

We will work out the ratio and the occupancy you are really running, tell you where the capacity is going, and say plainly if the honest answer is that the list is the problem rather than the floor.

Book a Free Consultation