What do customer service outsourcing companies publish about pricing?
Published 23 September 2026
Customer service outsourcing companies almost never publish their pricing. Of the 29 providers scored on public evidence, five publish how they charge at all, and only two of those, HGS and iSON Xperiences, go as far as a figure a buyer can compare.
The other 24 publish neither. These are the providers whose published materials this research found stating terms; a provider absent from this count may publish something we did not find.
How they charge, and what it costs, are different questions
A buyer needs two things before a procurement conversation, and providers treat them very differently.
How they charge
This is the shape of the bill. The common models across this industry are per-FTE (a flat rate per seated agent), per-transaction or per-ticket, per-minute, outcome-linked (paid against a resolution or a KPI), and a flat retainer. 5 of 29 publish which of these applies to them.
What it costs
This is a number. 2 publish one.
The gap between those two counts is the finding. A published commercial model without a published rate tells a buyer the mechanism and nothing about the magnitude, and it is the more common of the two by a factor of more than two.
What it actually costs
Two real figures exist in this field, and neither is a rate card for CX outsourcing in the sense a buyer usually means.
HGS: a real price, for a narrower question than it looks like
HGS publishes Agent X at £39 per user per month on the UK Government's G-Cloud 14 procurement registry, with stated termination terms and a UK data-location commitment. That price covers the Agent X software module specifically, not seat-based customer-service pricing. It is a real, checkable number, and it answers a narrower question than "what does an agent cost."
iSON Xperiences: a directory-published minimum and hourly band
iSON Xperiences publishes a $50,000+ minimum project size and a $200–300 per-hour rate band, on its Clutch profile rather than its own site. Under the same authorship test this page applies throughout, a provider-supplied directory field counts.
Both figures are independently corroborated rather than company-stated on a page nobody else can see. Neither is comparable to the other (a per-seat software price and a project minimum answer different buyer questions), which is itself the finding: even the two providers that publish something concrete don't publish the same kind of number.
What actually drives the number, even though this field won't publish it
Every generic pricing guide names the same cost drivers, and they hold here too: region (onshore, nearshore or offshore delivery), channel mix (voice costs more than async chat or email), service complexity, required volume and seasonality, and whether the engagement is a dedicated or shared team.
None of that is specific to this field's 29 providers; it's the general shape of outsourcing cost, and it's why "how much does outsourcing cost" has no single honest answer even before a specific provider enters the conversation.
What this research adds isn't a number for any of those factors. It's which of the 29 actually publish enough to make comparison possible in the first place, which is the structure vs. figure gap above.
For the delivery-region driver specifically: this field's own registered-HQ breakdown is on the BPO ranking, and the nearshore-versus-offshore trade-off for a European buyer is covered in the nearshore-Europe finding. For the channel-mix driver, see the channel-coverage scores on the call-center guide.
In-house vs. outsourced isn't the question this page answers
Comparing an outsourced rate against an in-house fully-loaded cost per agent is a real exercise, and it depends entirely on a buyer's own headcount, benefits and overhead numbers: nothing this research could verify about a third party.
What this research can say is narrower and checkable: whether a specific provider publishes enough for that comparison to even start. For 27 of 29, it doesn't.
Which providers publish what
| Provider | Publishes a structure | Publishes a figure | What is published |
|---|---|---|---|
| HGS | Yes | Yes | An actual price on a public procurement registry, with stated termination terms |
| iSON Xperiences | Yes | Yes | A minimum project size and an hourly band, on its own directory profile |
| Sutherland | Yes | No | Scaling bands, outcome-based and shared-risk models, priced against agreed volumes |
| Infosys BPM | Yes | No | FTE-based, transaction-based and outcome-linked pricing, with no inclusions or exit terms |
| Alorica | Yes | No | All-in pricing with no hidden fees, on a productised entry offering |
Every other provider in the field publishes neither.
Why a directory profile counts
Two of the five publish through a third-party directory rather than their own site. That still counts, because the commercial fields on those profiles are supplied by the provider: the platform hosts them, the provider writes them. The test this publication applies throughout is who authored a claim, not who hosts it.
It also cuts the other way. A directory profile that leaves those fields blank is a provider choosing not to publish, and that is scored as such.
What this does not say
It does not say 24 providers refuse to discuss price. Every one of them will quote on request. It says the information is not available before a conversation, which means a buyer cannot shortlist on cost without entering a sales process with each provider first.
It also does not say the two who publish figures are cheaper. Publishing a rate is a transparency behaviour, and this framework scores transparency because a buyer must verify everything anyway. But a published price is not a low price.
How this was established
Each flag is set from a claim in that provider's own evidence file, and the phrase that sets it is recorded there. A build check fails the deploy if the phrase ever leaves the ledger, so the count cannot drift from the evidence behind it.
The full method is on the methodology page, and you can read how we score commercial transparency. Every provider record carries its own claims with an evidence grade against each: the 29 data sheets.
Related: how many providers publish a control over their AI (five of 32 assessed across both editions).
Birchy operates commercially in the outsourcing sector and may work with providers named here. It makes no difference to the scores, because every provider is judged against the same published standards, from evidence anyone can check.