How Many Customer Support Agents Does My Business Need?
A step-by-step method for turning your contact volume and handling time into an honest agent count, including why the naive calculation understates it.
Most businesses answer this question by feel. Someone says "we probably need a couple of people" and that becomes the plan. The feel is often roughly right, but it is impossible to defend to a finance team and impossible to adjust when volume changes. The calculation underneath it is not complicated, and doing it properly takes about twenty minutes.
Start with three numbers
Everything follows from three figures. Get these before you do anything else, and get them from your systems rather than from memory.
- Contact volume: how many calls, chats, emails or tickets you receive in a typical day, broken down by channel.
- Average handling time: how long a contact takes from first response to fully closed, including any notes or CRM updates afterwards.
- Coverage hours: the hours you need someone available, which is not the same as the hours you are open.
The third one catches people out. If you want coverage from 8am to 8pm, that is a twelve-hour window, and no single person works twelve hours. Coverage hours drive staffing far more than volume does at small scale.
The naive calculation, and why it is wrong
The obvious approach is to multiply volume by handling time and divide by the working day. Say you receive 60 calls a day, each averaging 6 minutes. That is 360 minutes, or 6 hours of talk time. A single agent working an 8-hour day covers it with room to spare.
This is wrong in two ways, and both matter.
First, an agent is not available for 100% of their shift. Breaks, briefings, training, system issues and the simple fact that nobody is productive for eight consecutive hours mean realistic availability is closer to 75-85% of paid time. Our 6 hours of talk time needs roughly 7.5 hours of paid time to absorb it.
Second, and more importantly, contacts do not arrive evenly. They arrive in bursts. Your 60 calls are not one every eight minutes — they are twelve between 9am and 10am, three across the whole of lunch, and fifteen in the hour after your email campaign lands. A single agent perfectly sized for the daily average will have a queue of eight people at 9:15 and nothing to do at 2pm.
A better calculation
Work through it in this order.
Step 1: find your peak hour
Pull contact volume by hour for the last month. Find the busiest hour of a typical busy day — not the single worst hour of the year, but the level you hit regularly. If your busiest hour handles 12 calls at 6 minutes each, that is 72 minutes of work arriving in a 60-minute window.
Step 2: convert to concurrent agents
72 minutes of work in a 60-minute window needs 1.2 agents just to break even, before any allowance for availability. Divide by your availability factor — 0.8 is a reasonable planning number — and you get 1.5 agents needed during that hour.
Step 3: multiply by coverage hours
If you need 1.5 agents during peak and want coverage across twelve hours, you cannot staff that with 1.5 people. Two agents covering staggered shifts gives you twelve hours of single coverage with an overlap in the middle — which is precisely why a two-agent minimum exists. One agent cannot cover an extended day, take a holiday, or be ill.
Step 4: add the channel adjustment
Channels are not equivalent. A phone agent handles one conversation at a time. A chat agent typically handles two to three concurrently while keeping replies quick. Email and tickets are asynchronous, so they absorb into gaps rather than requiring dedicated capacity at a specific moment.
| Channel | Realistic concurrency | Planning note |
|---|---|---|
| Phone | 1 conversation | Peak-hour driven. Queues form fast. |
| Live chat | 2-3 conversations | Push past 3 and response quality drops noticeably. |
| Email / tickets | Sequential, not real-time | Fills the quiet gaps between live contacts. |
| Outbound calling | 1 conversation | Volume driven by dial attempts, not incoming contacts. |
A practical consequence: if you have moderate call volume and a steady ticket queue, one agent can often cover both, because the tickets absorb the gaps between calls. Trying to do the same with two phone queues does not work, because both demand the agent at the same moment.
Worked example
A home services company receives 45 calls and 20 web enquiries per day. Calls average 5 minutes including CRM notes. Enquiries take about 8 minutes each to call back and qualify. They want coverage from 8am to 6pm, six days a week.
- Daily workload: (45 × 5) + (20 × 8) = 225 + 160 = 385 minutes, roughly 6.4 hours.
- With 80% availability: 8 hours of paid agent time per day.
- Peak hour: 9 calls between 8am and 9am = 45 minutes of work in a 60-minute window. Comfortable for one agent, tight if two enquiries arrive at once.
- Coverage need: 10 hours a day, six days a week — impossible for one person.
The answer is two agents on staggered shifts. Daily workload alone suggests one; coverage requirements and the six-day week make two the honest number. The second agent is not idle — the overlap absorbs peaks and the quieter periods go into enquiry follow-up, which is where the revenue actually is.
When to add the third agent
Add capacity when the data says so, not when it feels busy. The signals worth watching:
- Your peak-hour queue regularly exceeds your target wait time.
- Abandoned calls or chats climb above the threshold you set.
- Follow-up work — the callbacks, the emails, the CRM updates — is consistently pushed to tomorrow.
- Handling time is creeping up, which usually means agents are rushing and creating rework.
- Coverage gaps appear whenever someone takes leave.
The last one is worth taking seriously. A two-agent team where both are fully loaded has no slack. The moment one takes a week off, you are running at half coverage. If that happens more than occasionally, the third agent is not a luxury.
What to do if you do not have the numbers
Plenty of businesses cannot pull contact volume by hour because the phone system does not report it and the enquiries arrive in a shared inbox. In that case, estimate deliberately for two weeks: count contacts by channel and rough handling time. Two weeks of imperfect data beats a year of guessing, and it will change the answer more than you expect.
Then start small. Two agents on a defined scope, with reporting from the first week, tells you more about your real staffing need than any amount of upfront modelling. You will discover that the volume is higher than you thought, or that the handling time is longer, or that half the contacts are one question you could answer on your website. All three are useful.
This guide reflects how we actually run campaigns and what we see go wrong. We have tried to be useful whether or not you ever work with us — including where that means recommending you do something other than outsource.
Related services
Inbound Customer Support
Dedicated agents answer your incoming calls, resolve routine questions and document every conversation in your systems.
From $4.99/hrLive Chat Support
Agents staff your website chat, answer questions in the moment and capture the details your sales team needs.
From $2.99/hrEmail & Ticket Support
Agents work your support inbox and helpdesk queue, replying accurately and keeping the backlog under control.
From $2.99/hrKeep reading
How Much Does Outsourced Customer Support Cost?
The pricing models providers use, what moves the rate, which costs sit outside the hourly rate, and how to compare two quotes that look different on purpose.
9 min readRead →Live Chat vs Phone Support
What each channel is genuinely better at, the cost difference and why it exists, and a decision framework for businesses that can only staff one.
8 min readRead →How to Measure Customer Support Agent Performance
Which metrics to track, which ones create perverse incentives, and how to combine them into a scorecard that survives contact with real agents.
9 min readRead →