What outsourced customer service actually costs, and how providers price it

Customer service outsourcing pricing is usually compared on one number: the rate. Two providers quote, one is lower, the lower one looks better.

The rate is the least interesting part of the agreement. The pricing model, meaning the unit you buy, determines what your provider optimizes for over the next three years. Every model rewards something and hides something, and the gap between what you pay for and what you want is where most disappointing outsourcing relationships originate.

Here is what each model does, in plain terms.

Per hour: you buy time, and you carry the productivity risk

Per-hour pricing is the simplest. You pay for agent hours at an agreed rate.

What it rewards: hours. Nothing in the model rewards resolving a contact faster, deflecting a contact that should never have happened, or improving a process so less work arrives.

What it hides: productivity. Two providers at the same hourly rate can deliver materially different output, and the model gives you no visibility into that. You are also carrying the risk of shrinkage assumptions, since an hour paid is not an hour available to handle contacts once breaks, training, coaching, and absence are taken out.

When it is right: short engagements, highly variable scope, staff augmentation where you are directing the work, and pilots where nobody yet knows what the volume looks like.

Per FTE: predictable, and the shrinkage question decides it

Per-FTE pricing charges a monthly amount per full-time equivalent. It is the most common model in managed contact center work and it budgets cleanly.

What it rewards: stability. It is neutral on productivity in the same way per-hour pricing is, but it usually comes with a managed service wrapper, meaning the provider owns supervision, training, quality, and scheduling. That wrapper is the actual value.

What it hides: the definition of an FTE. This is the single most important question in a per-FTE agreement and the one least often asked. How many productive hours does one FTE deliver after breaks, lunches, training, coaching, meetings, paid time off, and unplanned absence? Two providers quoting the same per-FTE price can differ by a meaningful share of delivered capacity based on shrinkage treatment alone. Get the shrinkage assumption in writing and ask what happens contractually if actual shrinkage exceeds it.

When it is right: ongoing programs with reasonably predictable volume, where you want a managed team rather than borrowed hands.

Per contact: aligned on volume, misaligned on quality

Per-contact pricing charges a fixed amount per call, email, or chat handled. It sounds like outcome pricing and is not quite.

What it rewards: handling contacts. More precisely, it rewards handling more contacts per unit of cost, which means the provider has an incentive toward shorter contacts.

What it hides: the difference between a handled contact and a resolved one. If a customer calls three times about the same issue, that is three billable contacts. The model pays your provider more when their work generates repeat contacts, which is the opposite of what you want.

When it is right: high-volume, low-complexity, well-defined work where resolution is unambiguous. It needs a first call resolution commitment alongside it, otherwise the incentive runs the wrong way.

Per resolution and outcome-based: aligned, and harder to write

Outcome-based pricing ties payment to a defined result: a resolved issue, a completed transaction, a qualified appointment, a processed claim.

What it rewards: the thing you actually want. It also transfers real risk to the provider, which is the point, and it is the structural argument behind outcomes-based BPO contracts.

What it hides: nothing about incentives, and a great deal about definitional complexity. Every outcome model lives or dies on the definition. What counts as resolved? Who adjudicates a dispute? What happens to contacts that fall outside the defined outcome set? These agreements take longer to write and require clean data both sides trust.

When it is right: mature programs where the work is well understood, volume is stable enough to model, and both parties have the reporting to adjudicate honestly. It is a poor first contract with a new partner and a good second one.

Mpathic contracts on outcomes rather than hours worked, with service level objective and service level agreement terms and monthly executive reviews.

Blended and tiered: usually the practical answer

Most real agreements combine models. A base per-FTE charge covering core staffing, a variable component for volume above a threshold, and an at-risk portion tied to service level or quality attainment.

This is generally the right structure for an ongoing program, because it gives you budget predictability, absorbs volume variation without renegotiation, and puts real money behind the performance commitments. Our guidance on building an ROI model that tells the truth covers how to compare structures on the same basis. The proportion at risk is the negotiation. Too little and the commitments are decorative. Too much and the provider prices in the risk, which you pay for anyway in the base rate.

What the rate does not include

Four costs sit outside most quoted rates and belong in your comparison.

Implementation and transition, which is real work: knowledge transfer, systems integration, training build, and parallel running. Technology, meaning who pays for the contact center platform, the CRM seats, and the reporting layer. Change requests, which is where a low base rate is frequently recovered, so ask how a scope change is priced before you sign rather than after. And your own management cost, because a lower-priced provider that requires more of your team's oversight is not cheaper.

For internal comparison against in-house delivery, remember that a wage is not a cost. The Bureau of Labor Statistics reported that private industry employer costs for employee compensation averaged $46.89 per hour worked in June 2026, of which $32.82 was wages and salaries and $14.07 was benefits (U.S. Bureau of Labor Statistics). Benefits, facilities, technology, recruiting, training, and supervision all belong in an in-house comparison, and leaving them out is the most common reason build-versus-buy analyses come out wrong. Our decision framework for in-house versus outsourced walks through the full comparison, and the same omissions distort offshore comparisons, as covered in the true total cost of offshore BPO.

Three questions to ask every provider

Regardless of model, ask these. What does one unit of what I am buying actually deliver, defined precisely? What behavior does this pricing model reward, in your own words? And what is not included in this rate?

A provider who answers the second question thoughtfully, including where their model's incentives diverge from yours, is telling you something useful about how they will behave when something goes wrong.

Customer service outsourcing pricing is a behavior contract. Per-hour and per-FTE models buy capacity and leave productivity risk with you, with the FTE definition and shrinkage assumption deciding what you actually receive. Per-contact pricing can reward the repeat contacts you are trying to eliminate. Outcome pricing aligns incentives and demands rigorous definitions. Blended structures are usually the practical answer.

Before your next renewal, do one thing: write down what behavior your current pricing model rewards, then compare it to the outcomes on your scorecard. Where those two disagree is where your agreement is working against you.

Talk to our team about structuring an agreement