Per-minute, per-hour, per-call, per-resolution and dedicated-seat call center pricing models compared by what each rewards, penalises and hides.
The model matters more than the rate
Two providers can quote what looks like the same price and bill you completely differently, because the unit underneath the number is doing most of the work. Before you compare a single figure, you have to know what you are being charged for — a minute of talk time, an hour of an agent's day, a completed call, a resolved issue, or a dedicated person. Each of those models rewards a different behaviour and hides a different risk, and the cheapest-looking one is regularly the most expensive outcome.
This guide walks through the models a call center actually uses, what each is good and bad at, and which shape of program each one fits. It deliberately quotes no rates: what a program costs depends on volume, hours, languages, complexity and the service levels attached, and a number invented here would mislead more readers than it helped. The point is to make you fluent in the models so that when you do get quotes, you can compare them on the same terms.
Per-minute pricing
You are billed for the time an agent spends connected, usually rounded to an increment. Per-minute pricing is common on answering and lighter inbound programs, and its appeal is that you pay in proportion to use: a quiet month costs less. Its risk is in the increment and in what counts. Billing rounded up to the next minute on calls that average well under a minute inflates the effective rate without any headline number looking unusual, and whether hold, transfer, wrap-up and the greeting before the caller speaks are billable varies by provider. If you consider per-minute pricing, the increment and the definition of billable time matter more than the rate.
Per-call pricing
You are billed a flat amount per handled call regardless of length. It is simple to forecast and it rewards brevity, which is either a virtue or a problem depending on your program: on a straightforward inquiry it aligns everyone on efficiency, but on a call that should have been handled thoroughly it quietly pressures the agent to close. Per-call pricing also makes the definition of a chargeable call the thing to negotiate — whether wrong numbers, hang-ups and repeat callers count is where two per-call quotes stop being comparable.
Per-hour pricing
You are billed for agent time whether or not calls are arriving, so you carry the cost of idle capacity and the provider carries none of the volume risk. That sounds worse than it is: on a busy, well-forecast program per-hour is often the most economical model, because you are not paying the margin a provider builds in to absorb volume uncertainty under per-minute or per-call pricing. Per-hour rewards you for accurate forecasting and punishes you for staffing to a peak you rarely hit, which is exactly where a staffing calculation earns its keep — our staffing calculator shows the seat count a given volume and service level actually implies, so you can sanity-check what you are being asked to pay for.
Dedicated versus shared agents
Cutting across all of the above is whether your agents are dedicated to you or shared across several clients. Dedicated agents cost more per hour and know your business deeply; they suit complex programs, regulated work and anything where product knowledge or brand voice is the point. Shared agents cost less because their idle time is spread across clients, and they suit simpler, lower-volume or spiky work where a dedicated team would sit idle. Many programs are a blend — a dedicated core with shared overflow for peaks — and the blend is often the honest answer for a business whose volume swings.
Per-resolution and outcome pricing
Increasingly, buyers ask to pay for resolved issues rather than for time or calls, and on the right program it aligns the provider with the thing you actually value. The catch is definitional: a resolution has to be defined precisely and measured honestly, or the model rewards closing tickets rather than solving problems. Pure outcome pricing also shifts risk onto the provider, who prices that risk back in, so it is rarely as cheap as it looks and rarely offered on programs where the provider cannot control the outcome. Treat it as a model to earn into once quality is proven, not a way to make a new relationship cheaper on day one.
How to choose the model before you compare the rate
Work from your own call pattern. If volume is steady and well forecast, per-hour usually wins. If volume is low and spiky, per-minute or per-call keeps you from paying for idle time. If calls are short and numerous, watch the per-minute increment; if they are long and valuable, per-call pressure works against you. If product knowledge is the point, you are buying dedicated agents regardless of the unit. Then judge every quote on cost per resolved contact rather than cost per unit, because a model that costs more per minute and resolves twice as much is cheaper where it counts. The outsourcing cost guide covers the drivers behind the rate, and the guide to comparing quotes shows how to normalise two offers onto the same basis.
Two of these models get their own treatment: how inbound programs are priced — where the service-level target does most of the work — and the answering-service versions of per-minute and per-call, where increments and unwanted calls decide the bill. If the quote in front of you is hourly, what an hourly rate actually buys takes it apart line by line.
When you are ready, the fastest way to a real number is to give a provider a representative month of your call data and ask them to price it in the model that fits — which is what our team will do rather than quote a rate off a template.
Frequently asked questions
Which call center pricing model is cheapest?
None of them in the abstract — the cheapest model depends on your call pattern. Per-hour usually wins on steady, well-forecast volume because you avoid the risk margin built into usage-based pricing. Per-minute or per-call protects you when volume is low or spiky and you would otherwise pay for idle time. The only reliable way to know is to price one real month of your calls in each model and compare the totals, because a model that looks expensive per unit can be cheapest per resolved contact.
What is the difference between dedicated and shared agents?
Dedicated agents work only your program and cost more per hour, because you carry their idle time; they suit complex, regulated or brand-sensitive work where deep product knowledge matters. Shared agents are spread across several clients and cost less, because their idle time is spread too; they suit simpler or spiky programs. Many programs blend the two — a dedicated core with shared overflow for peaks — which is often the most economical answer for volume that swings through the week or the year.
Should we ask for performance or per-resolution pricing?
It can be the right model once quality is proven and a resolution can be defined and measured honestly, because it aligns the provider with the outcome you value. But pure outcome pricing shifts volume and quality risk to the provider, who prices that risk back in, so it is rarely as cheap as it first appears and rarely offered where the provider cannot control the result. Treat it as something to earn into after a program is stable, not a way to make a new relationship cheaper on day one.
Why won't you publish per-minute or per-hour rates?
Because any figure we published would be wrong for most programs. Price depends on volume, hours, languages, the complexity of the work, whether agents are dedicated or shared, and the service levels attached, and every one of those moves the number. A rate card would mislead more readers than it helped. What we can do is price your actual call pattern in the model that fits it, which is both more accurate and more useful than a headline number.