Every outsourcing vendor has two numbers, and they publish neither.
The first is the smallest account they will take. The second is the largest they can hold before the wheels come off. The distance between those two numbers is the only thing that tells you whether this vendor can still be your vendor in three years.
Most lists of customer service outsourcing companies rank providers by size. Size is a point. What you are actually buying is a range — and the day your volume crosses the edge of it, you are not renegotiating. You are re-procuring.
Key Takeaways
- Vendors have a floor and a ceiling. Almost none publish either one.
- The vendor that can take you at 3 agents usually cannot hold you at 30. The one that can hold you at 30 would not have answered the phone at 3.
- Crossing that edge is a re-procurement, not a renegotiation: new selection process, retraining from zero, rebuilt quality baseline, roughly two to three months of seam.
- Hourly rate is a trap comparison. Hours you must buy, billable idle time, and who pays for ramp-up matter more.
- Ask every vendor the nine questions at the end, including us.
The cost nobody puts in the spreadsheet
Picture the ordinary version of this. You start with 3 agents on a subscription plan because that is what your volume justified. Two years later you are at 30. Your provider’s model was built for the first number, not the second, and the service starts to show it.
So you run a new search. That means a new selection process, re-documenting your product and policies for a new team, training agents from zero, rebuilding a quality baseline, and losing every piece of tribal knowledge the last team accumulated — the edge cases, the customers with history, the workarounds nobody wrote down. Buyers who have done it describe a degraded-service window of two to three months.
None of that appears in the rate card you compared two years earlier.
It is the same arithmetic that decides whether to build the team in-house or outsource it in the first place: the visible cost is the hourly rate, and the expensive part is everything that happens when the arrangement has to change.
Search “customer service outsourcing companies” and every result on the first page is a ranked list. Not one publishes a floor or a ceiling. That is not an oversight — a narrow band is a weakness, and no vendor volunteers it.
How the market actually splits
Three groups, described using each provider’s own published positioning. Where a commercial term is not published, this guide says so rather than guessing — the silence is itself the finding.
Enterprise BPOs — high floor, high ceiling
Who is here: Teleperformance (TP), Concentrix, Foundever, Alorica.
The largest operators in the category. Concentrix describes itself as “a global technology and services leader”; Alorica positions as a “customer service CX leader.” Several are publicly listed. They run programs for telecoms, airlines, banks, and large platforms.
Their band: enormous at the top, and it starts high. Their economics are built on large dedicated teams with workforce management layered on top. A well-run enterprise BPO absorbs thousands of contacts a day at a predictable service level, and does it better than anyone.
Where the floor bites: below the size where a dedicated team with its own supervisor makes sense, the model has no shape for you. That is design, not failure. We compared what actually changes when you go with a large BPO if you are weighing that path.
Mid-market specialists — medium floor, medium ceiling
Who is here: TaskUs, Helpware, SupportYourApp, SupportNinja.
This tier grew up serving technology companies and marketplaces rather than telecoms. TaskUs describes itself as “BPO & Digital Transformation Services.” SupportYourApp positions as “AI-Powered Customer Support Outsourcing.” SupportNinja leads with “seamless outsourcing that helps companies grow.”
Their band: narrower at both ends than it looks. Entry points typically sit in the small-team range rather than the single-agent range, and most quote dedicated FTEs. At the top, they hand off to enterprise BPOs.
Where the edge bites: when your volume cannot keep a small dedicated pod busy, or when it outgrows the account structure they staff around.
Subscription support-as-a-service — low floor, and the ceiling is the tier you bought
Who is here: Influx (“Customer Support, Done.”), Hugo (“Outsourcing for Customer Support, Ecommerce & AI”), ShyftOff (an “Agentic Workforce Delivery Platform”).
This group replaced the headcount conversation with a plan. You buy a tier or a block of hours and they handle staffing behind it. It genuinely lowers the entry point, and for many businesses it is the right answer.
Their band: the floor is the most accessible in the market. The ceiling is whatever tier you are on — and moving up tiers is a pricing conversation, not always a capability one. Agents may be shared across accounts.
Where the edge bites: long coverage windows with thin volume, where you pay for a tier you use a fraction of. And at the top, where the plan structure stops describing your operation. If you are comparing plan structures, it is worth reading any provider’s pricing page — including ours — for what the tier actually includes rather than what it costs.
Where Callnovo’s band sits
Founded in 2004. The working range is half an agent to 200+ seats — including a 200-seat government engagement at the top end and fractional accounts at the bottom. Both ends run through the same service: you interview and approve the agents, whether that is one person for twenty hours a week or a named team of forty.
Half an agent, or two hundred. Same phone call.
Both ends are the point, and they are one claim rather than two. A low floor on its own reads as “small vendor.” A high ceiling on its own means you cannot start until your volume justifies a full team. Together they mean you are not scheduled for a re-procurement.
What makes the low end workable is the delivery footprint. Our own operating centers are in Markham, Canada; Yueyang and Shenzhen, China; La Paz, Bolivia; Managua, Nicaragua; and Ecija, the Philippines — and the wider network delivers from 13 countries across North America, Latin America, Europe, Africa, and Asia — the full footprint and how we got here is on the about page. Partial allocations get assembled across those time zones instead of asking one person to stretch across a fourteen-hour day. What makes the high end workable is that the same operation runs dedicated named teams when the account calls for it.
Two commercial terms change the arithmetic, mostly at the small end:
- Pay-per-resolution, from $1, with unresolved contacts not billed. You are charged for outcomes rather than time. This suits businesses where the contact genuinely ends in a resolution — a refund, an order change, an account fix. It suits poorly where the call is intake and the real work happens later in the field; for those, hourly is the honest structure and we say so on the industry pages where it applies.
- Five days is the normal window to find and place an agent who fits the business. That is a staffing timeline, not a software timeline. Product and systems training runs on top of it.
If your program needs a single onshore delivery site with formal workforce-management certification and a contractually named site director, an enterprise BPO is built for that and we are not. That is a structural difference, not a scale one — it applies at 20 seats as much as at 200.
Why hourly rate is the wrong comparison
Published rates in this market run from single digits for offshore delivery to the mid thirties and above for onshore or specialist work. Buyers line these up in a spreadsheet and pick something in the middle. On its own that comparison means almost nothing, because the rate does not tell you:
How many hours you must buy. A $9 rate with a 160-hour monthly minimum costs more than an $18 rate with no minimum, if you only need 60 hours of real work.
Whether idle time bills. In a dedicated model you pay for the seat, not the contact. At 40 contacts a day across 14 hours, a meaningful share of the paid hours is waiting.
Who pays for ramp-up. Some contracts bill training hours at full rate, some absorb them, some bill at a reduced rate. Across a three-month engagement this can outweigh the gap between two vendors’ headline rates. This is also the line where a managed team stops looking expensive next to freelancers — the freelance rate excludes the recruiting, training, and coverage gaps you end up absorbing yourself.
What happens when volume moves. Ask what a 30% drop costs you, not only what a 30% spike costs. Spikes are the fun question. Drops are where dedicated models hurt.
Customer service outsourcing pricing models, in plain terms
| Model | You pay for | Works when | Hurts when |
|---|---|---|---|
| Per hour / per FTE | Agent time | Volume is steady and fills the hours | Volume is uneven or thin |
| Per ticket | Each contact handled | Contacts are short and similar | Complexity varies a lot |
| Per resolution | Each issue actually solved | The contact ends in an outcome | The call is intake only |
| Subscription tier | A capacity ceiling | Volume is predictable and inside the tier | You use a fraction of the tier |
Most operations end up using more than one of these — hourly for the dedicated core, per-resolution for the overflow. If you are trying to get the blend right before you talk to anyone, our take on structuring CX cost walks through where each model earns its keep.
Nine questions to ask any outsourcing provider — including us
Take these to every vendor on your shortlist. The answers separate providers far more sharply than any ranked list does.
- What is the smallest account you will take, and the largest you currently run? Both numbers. Ask for the largest you run, not the largest you could.
- What happens to my account if I triple in eighteen months — same team structure, or a handoff?
- Is idle time billable?
- Who pays for ramp-up, and how long is it?
- Which of your delivery locations would this account run from, and can I speak to someone there?
- Which languages could you actually staff for my account, starting next month? A headline count answers a different question than the one you are asking. We staff 65+ languages with human agents and cover 100+ with AI voice; what you need to know is which of those we can put on your account next month, and that is a shorter, more useful list. Ask every vendor for that list, not the number.
- What integrations do you support with my existing stack? Get it in writing. Our agents work inside HeroDash, the platform we built and run ourselves, which connects natively to Amazon Selling Partner API, Shopify, and eBay, plus WhatsApp, LINE, and TikTok on the messaging side; for field-service businesses, Jobber, ServiceTitan, and Skimmer. Whatever system you run, ask for a yes or a no rather than “we integrate with anything.”
- If we handle health information, will you sign a Business Associate Agreement? We will — that is covered on our healthcare support page. Note that signing a BAA and holding a third-party security certification are different things — ask each vendor which one they are actually offering.
- What does a 30% volume drop cost me?
How to run the shortlist
Start from your own numbers, not from a list of company names.
Write down four figures, not three: daily contact volume, the hours you need covered, how much volume swings week to week, and where you expect to be in two years. The fourth is the one buyers skip, and it is the one that decides whether you pay for a switch later.
Then check the band, not the tier. Ask each vendor for both ends. A vendor whose band contains your number today but not your number in two years is a decision you will revisit — price that in now, or pick differently.
Then compare inside comparable bands only. Putting an enterprise BPO and a fractional provider side by side on hourly rate produces a number that means nothing, because they are not selling the same unit.
If you want a second opinion on where your four numbers land, send them to us. We will tell you which band you belong in and which kind of provider fits — and if that is not us, we will say so. An account that outgrows its vendor in eighteen months is not a win for either side. More on how we structure outsourced support if you want the mechanics first.
If you want this market cut a different way — by published scale, language coverage, and stated minimum team size, provider by provider — there is a companion comparison of six named providers on callnovo.com. This guide sorts by band; that one sorts by profile.