You asked three vendors for a quote on the same support program. One priced it per productive hour, one per dedicated agent, and one per resolved ticket. Now you hold three numbers that don’t line up, and each vendor calls its model the cheapest.
The quotes differ because each model leaves a different share of the risk with you. One makes you pay for agents during quiet hours, and another charges again when customers call back. Many buyers would rather pay for results, but few have that option. Recent research from HFS Research shows that 76 percent of the CX buyers it surveyed are open to outcome-based or gainshare pricing, yet only 5 percent hold a contract built that way.
The seven CX outsourcing pricing models are per productive hour, per full-time equivalent (FTE), per minute, per contact, per resolution, outcome-based, and a fixed managed-service fee. Pick per FTE for steady, complex work, per contact or per resolution for high-volume Tier 1 queues, and per minute for short voice transactions. Use per productive hour while you build a volume baseline, and outcome-based pricing once your data is clean enough to measure results.
Helpware publishes this guide and sells customer experience outsourcing under several of the models below. We explain each model the way we would to a client, including where our own options don’t fit. Verify specifics for your use case.
Key Takeaways
- Each model gives your vendor a financial reason to favor certain behavior.
- Under per-resolution pricing, the definition of “resolved” decides what you pay.
- AI tools that cut handle time lower your bill right away under per-minute pricing, but under per-FTE or fixed-fee pricing the vendor keeps the savings until you re-baseline.
- Re-pricing triggers, re-baseline rules, data ownership, and a model-change notice period let you switch models later without renegotiating the whole contract.
CX Outsourcing Pricing Models
Before going model by model, here is how the seven compare on the unit you pay for, who absorbs volume swings, where each fits best, and the main problem to watch for.
| Model | Billing Unit | Who Carries Volume Risk | Best Fit | Main Watch-Out |
|---|---|---|---|---|
| Per productive hour | Agent hour spent on your work | You | New programs, variable Tier 1 work | Padded or idle hours |
| Per FTE (dedicated agent) | Agent per month | You | Steady, complex, or regulated work | Paying for quiet hours |
| Per minute | Handle minute (talk plus wrap-up) | You | Short voice transactions | Longer calls earn more |
| Per contact | Call, chat, email, or ticket handled | Vendor | High-volume Tier 1 queues | Quick closes and repeat contacts |
| Per resolution | Contact resolved, by agreed definition | Vendor | Tier 1 with clear resolution rules | Disputes over what “resolved” means |
| Outcome-based | KPI attainment or shared gains | Shared | Mature programs with clean data | Gaming the measured metric |
| Fixed managed-service fee | Monthly fee for an agreed scope | Vendor, within the band | Stable scope, long planning horizon | Quiet understaffing |
Three Questions That Decode Any CX Pricing Model
The table shows that the models differ mainly in who carries the risk, and most quotes you receive will bill for agent time or per unit of work. In the same HFS Research study, unit-based and time-and-materials pricing account for more than 70 percent of primary contracts. To see where a quote leaves the risk, ask three questions before you compare rates.
- What is the billing unit? Time (hours, minutes, seats) or output (contacts, resolutions, KPIs). Time units reward the vendor for effort; output units reward it for throughput.
- Who absorbs volume swings? On time-based models, you pay for capacity whether contacts arrive or not. On unit-based models, the vendor carries the cost of idle agents and the scramble to staff a spike.
- Who absorbs quality risk? If the invoice doesn’t change when customers call back, you carry the quality risk. Repeat-contact clawbacks, resolution definitions, and KPI-linked fees move part of it to the vendor.
If your volume is steady, carrying volume risk costs you little, and a time-based model is often the simpler choice. If your volume swings or you have no history yet, you have more reason to push that risk to the vendor. Keep your answers in mind as you read through the models.
The 7 CX Outsourcing Pricing Models in Detail
The first three models bill for agent time, so you carry the volume risk. The next three bill for output or results and shift more of that risk to the vendor. The last one is a single fee for an agreed scope. For each model, we cover how it bills, what the rate usually includes, where it fits, what to watch, and what to ask before you sign.
1. Per productive hour
How it bills: You pay for each hour an agent spends logged in and available for your work. “Productive” usually excludes breaks, training, and meetings, but vendors define it differently, so read the definition.
Usually included: Agent wages, supervision, standard QA, and workforce management. Technology, onboarding, and after-hours premiums often sit outside the rate.
Best fit: Pilots, new product launches, and Tier 1 work where you don’t yet know your volume curve. The hours you buy become the baseline data for a better model later.
Watch-outs: You pay for idle time when contacts dip, unless the vendor cuts scheduled hours to match. The vendor also earns more when work takes longer, so you need handle-time reports to see whether the hours you pay for are well used.
Ask the vendor: How do you define a productive hour? Who approves schedule changes, and how much notice do you need to cut hours?
2. Per FTE (dedicated agent)
How it bills: You pay a monthly rate for each full-time agent assigned only to your account. The rate usually bundles wages, benefits, management share, facilities, and standard tools.
Usually included: Recruiting, training, team leaders at a set ratio, and QA. Ask about buffer staff for absences, which some vendors price as an add-on.
Best fit: Steady volumes and work that rewards product knowledge: SaaS technical support, regulated healthcare lines, and VIP queues. A dedicated customer service team learns your product, policies, and repeat customers, which agents shared across several clients rarely have time to do.
Watch-outs: A seat costs the same on a slow Tuesday as on your busiest day. Scaling down takes notice periods, and scaling up takes hiring time.
Ask the vendor: What is the notice period to add or remove agents? Which roles are dedicated and which are shared across clients?
Dedicated vs. shared agents: Shared agents split their time across several clients. Rates run lower, and the vendor absorbs idle time, but agents know your product less well. Shared pools often pair with per-minute or per-contact billing.
3. Per minute
How it bills: You pay for each minute agents spend handling your contacts. Contracts differ on whether that means talk time only or talk time plus after-call work (wrap-up), and the difference changes your invoice.
Usually included: Agent time on contacts. Monthly platform fees, minimum commitments, and setup fees often appear separately.
Best fit: Short, scripted voice transactions such as order status, appointment confirmations, and after-hours message taking.
Watch-outs: The vendor earns more when calls run longer. Without a handle-time ceiling, you pay for every minute a customer spends on hold or waiting through a slow transfer.
Ask the vendor: Do billable minutes include wrap-up and hold time? Is there a monthly minimum, and do unused minutes roll over?
4. Per contact (per transaction)
How it bills: You pay a fixed price for each call, chat, email, or ticket the vendor handles. Prices often vary by channel and by contact type.
Usually included: Staffing, management, and QA. The vendor carries the cost of idle agents, so you stop paying for quiet hours.
Best fit: High-volume Tier 1 queues with predictable contact types: password resets, order tracking, billing questions, and returns.
Watch-outs: The vendor earns more by handling more contacts, which rewards quick closes, unnecessary transfers, and customers who have to call back.
Ask the vendor: Does a transfer or a follow-up on the same issue count as a new contact? What happens to the price when the contact mix shifts toward harder issues?
5. Per resolution
How it bills: You pay only for contacts resolved under an agreed definition, for example no repeat contact on the same issue within seven days. Unresolved contacts cost nothing or carry a lower rate.
Usually included: Everything in per-contact pricing, plus the vendor takes the risk of contacts that don’t resolve on the first attempt.
Best fit: Tier 1 programs with clear resolution rules and reliable data from your ticketing or CRM system.
Watch-outs: The definition of “resolved” decides what you pay, so expect it to be one of the hardest points to negotiate. Vendors also have a reason to steer hard contacts back to you as escalations.
Ask the vendor: Who measures resolution, you or the vendor? How do escalations, reopened tickets, and customer-abandoned contacts count?
6. Outcome-based (KPI-linked or gainshare)
How it bills: Part of the fee depends on results. Common structures put a share of the base fee at risk against KPIs, pay a bonus above target, or split measurable gains such as retained revenue or recovered sales.
Usually included: A base fee under another model (often per FTE or per hour) plus the at-risk or gainshare portion.
Best fit: Mature programs with a clean baseline, KPIs both sides control, and a vendor that owns the levers behind them. Retention, collections, and sales programs suit gainshare; service programs suit KPI-linked fees.
Watch-outs: Agents and team leads focus on whatever the fee is tied to. A fee tied only to customer satisfaction (CSAT) invites long, generous calls; a fee tied only to handle time invites rushed ones.
Ask the vendor: How do we agree the baseline? Which KPIs depend on things only we control, such as product bugs or policy changes?
7. Fixed managed-service fee
Every model so far changes your invoice when volume or results change. A fixed managed-service fee keeps the invoice steady and leaves day-to-day staffing decisions to the vendor.
How it bills: You pay one monthly fee for an agreed scope, usually a volume band plus service levels. The vendor decides how to staff and run the work.
Usually included: Agents, management, QA, workforce planning, reporting, and often technology. Contacts above the band trigger an overage rate or a change request.
Best fit: Stable, well-documented scopes and buyers who want one predictable line item. Back-office queues and mature service desks fit well.
Watch-outs: The vendor keeps any savings from running leaner, so it has a reason to staff your program with fewer agents than you expect. Scope changes mid-contract turn into negotiations.
Ask the vendor: What volume band does the fee cover, and what is the overage rate? What staffing data will you share each month?
Hybrid CX Pricing Structures
Each single model leaves you exposed somewhere, which is why contracts past the pilot stage often blend two. A hybrid puts the predictable part of your volume on a fixed model and bills the unpredictable part in a way that rises and falls with demand. These four structures cover most hybrid deals.
| Hybrid Structure | How It Works | Best For |
|---|---|---|
| Base plus variable band | A fixed FTE core covers baseline volume; contacts above it bill per hour or per contact | Seasonal programs with a known floor |
| Tiered volume pricing | The per-contact price drops as monthly volume crosses agreed thresholds | Growing programs that want scale savings |
| Bonus-malus overlay | An FTE or hourly base, with a percentage added or deducted on two or three KPIs | Programs moving toward outcome pricing |
| AI tier plus human tier | Automated contacts bill per AI resolution; escalations go to humans billed per FTE or hour | Programs adding AI agents to Tier 1 |
Whether you pick a single model or a hybrid, the billing unit also shapes the everyday decisions your vendor makes, from how team leads coach agents to how many people they schedule.
What Each Pricing Model Pays Your Vendor to Do
Every pricing model gives the vendor a financial reason to favor certain behavior. Under per-minute billing, a team lead has no financial reason to coach agents toward shorter calls. Under per-contact billing, a ticket closed too early earns a second fee when the customer writes back. If you don’t name that behavior and counter it in the service level agreement (SLA), you end up paying for it. The table pairs each model with the problem it invites and the guardrail to write into the contract.
| Model | Vendor Earns More When | Failure Mode to Watch | SLA Guardrail |
|---|---|---|---|
| Per productive hour | Agents log more hours | Padded hours, slow work in quiet periods | Occupancy reporting and a handle-time ceiling per contact type |
| Per FTE | Seats stay filled | Idle capacity, slow right-sizing | Utilization floor and a quarterly staffing review |
| Per minute | Calls run longer | Long holds, slow wrap-up | Handle-time ceiling by contact type and QA on call control |
| Per contact | More contacts get handled | Quick closes, transfers, repeat contacts | No charge for repeat contacts within a set window, plus a first-contact resolution floor |
| Per resolution | More contacts get marked resolved | Loose definitions, easy contacts kept, hard ones escalated | Customer-confirmed resolution, reopen clawback, and a monthly audit sample |
| Outcome-based | The measured KPI rises | Gaming one metric at the cost of others | A balanced scorecard of two or three KPIs with an agreed baseline |
| Fixed managed fee | Delivery costs fall | Quiet understaffing, stalled improvements | Minimum staffing disclosure, SLA penalties, and annual improvement targets |
Two guardrails belong in almost every contract, whatever the model. First, track repeat contacts, because they expose quick closes under any model. Second, review the CSAT and QA trend each month next to the invoice, so you see quality and cost together. An independent quality assurance outsourcing team gives you a second view of both, separate from the vendor’s own reports.
Same Program, Seven Invoices: A Volume-Swing Stress Test
Guardrails protect quality, but the biggest cost differences between models appear when your volume changes, and a rate card doesn’t show them. To make them visible, we priced one hypothetical support program under all seven models, then moved volume 30 percent down and 30 percent up.
Illustrative inputs (not market rates): 20,000 contacts a month, six minutes average handle time including wrap-up, 80 percent agent occupancy, 150 productive hours per agent per month, and 80 percent resolution. The FTE team is sized for baseline volume at 17 agents. Rates are round numbers chosen so the models cost about the same at baseline:
- Per productive hour: $20, with hours flexing to volume
- Per FTE: $3,000 per agent per month for 17 agents
- Per minute: $0.40 per handle minute
- Per contact: $2.50
- Per resolution: $3.10 per resolved contact
- Outcome-based: the FTE fee with 15 percent at risk on service-level KPIs
- Fixed managed fee: $50,000 for 16,000 to 24,000 contacts, plus $2.50 per contact above the band
| Model | 14,000 Contacts (−30%) | 20,000 Contacts (Baseline) | 26,000 Contacts (+30%) | Who Absorbed the Swing |
|---|---|---|---|---|
| Per productive hour | $35,000 | $50,000 | $65,000 | You pay for volume; the vendor must find hours for the spike |
| Per FTE | $51,000 | $51,000 | $51,000 | You pay for idle seats on the dip; customers wait on the spike |
| Per minute | $33,600 | $48,000 | $62,400 | You, minute by minute |
| Per contact | $35,000 | $50,000 | $65,000 | The vendor carries staffing risk both ways |
| Per resolution | $34,720 | $49,600 | $64,480 | The vendor carries staffing and resolution risk |
| Outcome-based | $51,000 | $51,000 | $47,175 | Shared: the vendor forfeits half the at-risk fee when service levels slip |
| Fixed managed fee | $50,000 | $50,000 | $55,000 | You pay the full fee below the band; overage applies above it |
The results show where each model puts the cost of a volume swing.
- The dip exposes fixed capacity. At 14,000 contacts, the FTE invoice works out to about $3.64 per contact and the fixed fee to $3.57, against $2.50 under per-contact pricing.
- The spike exposes capacity limits. The 17-agent FTE team covers about 2,550 productive hours, while 26,000 contacts need about 3,250. The invoice stays flat, but wait times grow unless you add agents, and hiring takes time.
- The variable models move with volume and pass the staffing problem to the vendor. You still pay for that protection, because per-unit rates usually include a margin to cover the vendor’s idle-time risk.
Run the same test with your own volume history and quotes before you choose. To compare quotes on a single unit, use the cost-per-resolved-contact method in our CX outsourcing cost guide. Published rate examples show how far outsourced customer service cost varies by provider and region.
How AI Is Changing CX Outsourcing Pricing
Volume isn’t the only thing that moves your invoice. As AI takes over simple contacts, the work left for your human agents changes, and so does the way vendors price it. Gartner predicts that by 2029, agentic AI will autonomously resolve 80 percent of common customer service issues without human intervention, cutting operational costs by 30 percent. When automation handles the quick, simple contacts, the ones that reach human agents take longer and need more skill, which tends to push per-contact prices for human work up.
Plan for three changes in how contracts get priced:
- AI-resolved contacts get their own unit. Vendors price automated resolutions separately from human ones, usually per resolution or per conversation. The definition of an AI resolution needs the same rigor as a human one, including what happens when the customer comes back.
- Productivity gains need an owner. Agent assist tools that reduce average handle time cut your bill right away under per-minute pricing. Under per-FTE or fixed-fee pricing, the vendor keeps the gain until you re-baseline. Write a re-baseline trigger into the contract.
- Blended contracts price the two tiers separately. The AI tier plus human tier hybrid described earlier lets you pay per resolution for automated contacts and per agent or per hour for the rest. Our comparison of AI vs human customer support covers which contacts belong in each tier.
If you plan to add automation mid-contract, agree up front how AI-handled contacts count toward volume bands, minimums, and KPIs. Ask providers to show you that mapping before you sign.
Which CX Outsourcing Pricing Model Fits Your Program?
Volume swings, vendor incentives, and AI all point back to two facts about your program: how steady your volume is and how complex the work is. The matrix below matches common combinations of the two to a model.
| Your Program | Recommended Model | Why |
|---|---|---|
| New program or launch, no volume history | Per productive hour or per FTE for a pilot | Builds the baseline you need for any per-unit model |
| Steady volume, technical or product-heavy work | Per FTE (dedicated) | Product knowledge outweighs idle-time cost |
| Steady volume, Tier 1, clean ticket data | Per contact or per resolution | You stop paying for quiet hours |
| Seasonal peaks | Base plus variable band | The core covers the floor; the band absorbs the peak |
| Short, scripted voice calls | Per minute, with a handle-time ceiling | You pay only for talk time |
| Regulated work (HIPAA, PCI) | Per FTE or fixed managed fee, plus a compliance KPI | Avoids rewarding speed over compliance |
| Mature program with agreed KPIs | Outcome-based overlay on an FTE or hourly base | Ties part of the fee to results you both measure |
Your industry adds constraints the matrix doesn’t show:
- SaaS: Tier 2 and Tier 3 tickets reward dedicated teams, while password resets and billing questions fit per-contact pricing. That’s why splitting the queue by tier often makes sense for SaaS programs.
- Healthcare and telehealth: Patient lines carry protected health information, so the contract must reward accuracy and privacy over speed. A HIPAA-compliant call center priced per FTE keeps trained agents on your account.
- Ecommerce and retail: Holiday volume runs well above the yearly average, and base plus variable band pricing absorbs that peak. Planning peak season support early keeps the variable band staffed with trained agents.
- Fintech: Payment and fraud queues need a PCI-compliant call center and accuracy KPIs. Per-contact pricing works for card activations, but disputes belong with dedicated agents.
Moving Between Pricing Models During a Contract
The matrix describes your program as it is today, and that picture changes once you have a few months of data. A vendor with no volume history to go on builds that uncertainty into any per-unit or outcome-based price. If you start on time-based pricing and move toward output pricing as your data matures, you avoid paying for that uncertainty. This four-stage path suits most mid-market programs.
- Run a pilot on time-based pricing. Use per productive hour or per FTE for 30 to 60 days. Capture volume by hour, handle time by contact type, resolution rate, and contact reasons.
- Set a core plus a band. Size an FTE core for your floor volume and price the rest per hour or per contact.
- Move stable queues to per-unit pricing. Once a queue has consistent contact types and reliable resolution data, shift it to per contact or per resolution.
- Add outcome terms last. Put a share of the fee at risk on two or three KPIs both sides agreed during the earlier stages.
Write the path into the first contract, so moving to the next stage doesn’t require a full renegotiation. Four clauses make the moves possible:
- Re-pricing triggers: Volume outside the agreed band for two consecutive months opens a pricing review.
- Re-baseline rules: New automation, a product launch, or a policy change resets the baseline that KPIs and rates rely on.
- Data ownership: You own the contact, quality, and workforce data, so you keep the evidence to benchmark the next model or the next vendor.
- Model-change notice: Either side has the right to propose a model change with an agreed notice period, without triggering termination fees.
The same clauses reduce outsourcing risks such as vendor lock-in (because you keep your data and the right to change models) and contract drift (because pricing gets reviewed when volume moves outside the band). If you are still choosing a CX outsourcing partner, weigh these clauses next to price, quality, and security.
How Helpware Structures CX Outsourcing Engagements
Our own engagement models cover several stages of the path above, from hourly billing when you need extra agents to outcome-based and gainshare terms when we run the full program. All three are listed on our pricing page:
- HW.Talent: On-demand dedicated agents to fill gaps or add capacity, billed hourly or by subscription. Operations, onboarding, and HR are shared; team leaders and QA are shared or dedicated.
- HW.Team: A ready-made support team with dedicated agents and built-in leadership, billed hourly.
- HW.Hub: A full-service model in which we lead strategy and execution, billed hourly, per transaction, outcome-based, or through gainsharing.
Before you commit to a larger team, we run 30-to-60-day pilot programs. When a pilot works, we have scaled programs from five to ten FTEs to more than 500 within 90 to 120 days. To keep quality consistent as a program grows, our hub and spoke model runs management, QA, and workforce planning from central hubs and adds capacity across 19 locations on four continents, in more than 45 languages.
For the regulated programs in the matrix above, our compliance certifications include SOC 2 Type II, ISO 27001, HIPAA, and PCI DSS. Scaling speed matters for the volume spike in the stress test: in one engagement, a furniture brand doubled support headcount in three weeks and cut its abandon rate by 20 percent.
Our limits: We don’t publish a rate card, so you need a scoping call for a quote. Per-minute billing isn’t one of our listed options. Hiring for a new team takes at least two to four weeks.
Choose the Model, Then Write the Guardrails
The rate on a quote tells you less than the model behind it. Each model decides who pays when volume drops, who has to find extra agents when it spikes, and who covers the cost when customers need to call back. Start with the model that fits the data you have today, which is usually time-based while you build a baseline. Then write the guardrail against the behavior that model rewards, and add the clauses that let you move to per-unit or outcome-based pricing once your numbers are reliable.
Want to see how your program would price under different models? Talk to our team about a scoped quote, or compare our engagement options on the Helpware CX pricing page.











