Anyone who asks for three outsourcing quotes almost always receives three documents that cannot be laid side by side. One quotes per agent per month, another per productive hour, the third per resolved case. The three figures look similar and none of them means the same thing.
The classic mistake is picking the lowest. The less obvious and more expensive mistake is failing to notice that the three describe different scopes.
The unit of measure is the first trap
Before comparing prices you have to equalise units. There are four common ways to quote an operation, and each one shifts the risk to a different side of the table.
- Per FTE (dedicated agent per month). You are buying capacity, not outcome. It is the most transparent option and the right one when volume is stable and processes are still changing. If volume drops, you pay anyway.
- Per productive hour. It sounds precise, but everything depends on the definition of productive. Ask whether it includes training, coaching, quality sessions, breaks and idle time. Every exclusion raises the real cost per contracted hour.
- Per case, ticket or transaction. It aligns incentives and protects against over-consumption, but it requires a minimum volume and a rigorous definition of what counts as a case. Without that, the month-three conversation is whether a reopened case gets billed twice.
- Hybrid. A fixed floor covering the structure plus a variable component by volume. It is what usually gets signed once both sides understand real demand.
An FTE quote and a per-case quote are only comparable once you convert both to the same basis: cost per unit of work actually delivered. That almost always requires estimating productivity, which is exactly why the provider's assumptions should be requested in writing.
What almost never comes itemised
A per-agent rate is not a salary plus a margin. Several cost items sit between the two, and they exist whether or not they appear in the PDF. If they are not explicit, they are either buried in the rate or not covered at all — and both possibilities matter.
Absence and unavailable time
Holidays, sick leave, training, attrition. No operation delivers 100% of contracted hours: there is always a share of unavailable time the provider has to cover with extra staff. The useful question is blunt: how many hours of effective coverage am I buying, and who pays for the replacement when someone is out?
The ramp curve
A new agent does not produce on day one what they produce on day sixty. Someone pays for those weeks of lower output. If the quote does not say who, you will find out from the invoice or from your first-quarter metrics.
Supervision and quality
How many agents per team lead, and whether the quality auditor is included or billed separately. An operation without sample auditing does not have quality control; it has self-reporting.
Technology and licences
CRM, telephony, recording, automation tooling. Define whether the client or the provider supplies the licences, and what happens to your data and configurations if the contract ends.
Attrition
Replacing and retraining someone is a real, recurring cost. A provider who cannot tell you last year's annual attrition probably is not measuring it, and an operation that does not measure attrition does not control it.
Four questions that clarify any quote
- What is excluded? The exclusion list says more about scope than the inclusion list. Ask for it explicitly.
- What happens if volume rises 40% in a month? The answer reveals whether there is real surge capacity or whether the plan is overtime until quality breaks.
- How does the rate adjust year to year? In Colombia the norm is indexing to inflation or to the minimum wage. Having the formula written down avoids an awkward renegotiation every December.
- How do I exit? Notice period, handover of process documentation, data return, transition window. A contract that is easy to leave is a contract that is safer to enter.
Why the lowest rate usually costs more
A rate well below market is being sustained somehow, and the ways are few and well known: less supervision per agent, hiring below the required profile, trimmed training, or an engagement model whose employment liabilities get claimed later. None of those levers is invisible — they show up in attrition, in escalation rate and in rework, two or three months late.
That is where price and metrics connect. Cost per resolved case, not cost per hour, is what reveals whether the cheap rate was actually cheap. It is one of the six metrics worth agreeing before go-live; the other five are in what to measure in a BPO operation from month one.
What to ask for before signing
- The rate broken down into its components, even if aggregated into blocks.
- The productivity and unavailable-time assumptions it was calculated on.
- The proposed shift pattern set against your demand curve by time slot.
- The ramp plan with milestones and who absorbs partial output.
- The committed metrics, how they are measured and who audits them.
- The data protection framework: processor or controller role, sub-processors, where data sits.
A provider who answers those six without discomfort has already told you something important about how they operate. One who answers with a rate and a promise has told you something too.
What you are buying is not a rate
It is the ability to run an operation at a defined quality, sustained over time, with someone accountable for keeping it there. That is the object of the comparison. The rate is only how it gets paid.
At smartBPO we quote after the diagnosis, not before, and we hand over the assumptions along with the number. Not out of generosity: a rate built on the wrong assumptions breaks by month four, and breaking it costs both sides.