The conversation about outsourcing customer support almost always starts the same way. Someone pulls up a spreadsheet, compares agent salaries in two columns, and concludes that keeping things in-house is cheaper. It is a reasonable instinct and it is almost always wrong. Running support in house at scale involves a category of expenses that rarely appear in that initial comparison, and by the time they become visible they have already done considerable damage to the operating budget.
I want to be clear that this is not an argument against in-house teams as a concept. There are scenarios where retaining certain functions internally makes genuine strategic sense. What I am arguing is that the decision to run support in house needs to be made with accurate cost information, not a simplified version that excludes the expenses that compound most aggressively with scale. Conversations with call centre outsourcing South Africa operations and other offshore models have given me a different view of where the real cost gaps emerge.
- The salary figure is the starting point, not the full cost of support in house
- Technology, infrastructure, and the costs that nobody budgets for upfront
- Opportunity cost is the expense that almost never makes it onto the spreadsheet
- Scaling reveals the structural weaknesses of in-house support models
- Support In House: Turnover is not a line item, it is a dynamic that affects everything else
- What the full cost picture actually looks like when you run the numbers honestly
- Continue exploring the operational economics of customer support
The salary figure is the starting point, not the full cost of support in house
When businesses calculate the cost of keeping support in house, the headline figure is almost always the salary. It is visible, easy to compare, and feels like the primary cost driver. But research shows that hidden costs push in-house expenses 25 to 40% beyond base salaries once recruitment, turnover, and opportunity cost are factored in. That gap is significant and it widens considerably as the team grows.
Recruitment alone is a material cost that many operations managers dramatically underestimate. Sourcing candidates, running assessments, conducting interviews, and onboarding successful hires takes time from existing team members whose time has a cost. When attrition is running at industry-standard rates, which in contact centres can reach 30 to 45% annually, that recruitment cycle does not stop. It becomes a permanent overhead funded by the team’s productivity that never appears on a clean budget line.
Technology, infrastructure, and the costs that nobody budgets for upfront
A functioning support in house operation requires technology infrastructure that most businesses initially underestimate. CRM licences, ticketing platforms, quality assurance tools, workforce management software, telephony systems. Each of these carries both a direct cost and a management overhead. Someone has to own vendor relationships, manage renewals, troubleshoot issues, and evaluate upgrades. In a dedicated outsourced operation that cost is distributed across many clients. In an in-house team it sits entirely with the business.
Office space is another category the headline comparison often ignores. Not every support in house team works remotely, and even those that do require equipment, security infrastructure, and IT support. The global outsourced customer care market reached $77 billion in 2024 and continues growing, which reflects a broad business recognition that the in-house total cost often does not hold up to honest scrutiny. That scale of market adoption is not driven by fashion. It is driven by finance.
Opportunity cost is the expense that almost never makes it onto the spreadsheet
There is a cost category that almost never appears in a budget comparison but it is arguably the most significant one for growth-focused businesses: opportunity cost. When senior leaders are managing a support in house operation they are not doing something else. The time spent on hiring, training, QA, tooling decisions, and people management is time not spent on product development, commercial relationships, or strategic planning.
This compounds at scale. A small in-house team might absorb a modest amount of leadership attention. A large one becomes a significant operational management commitment in its own right, requiring dedicated management layers, reporting infrastructure, and governance systems that all carry cost. Many of the businesses I have worked with have found that once they account for opportunity cost honestly, the case for keeping support in house at scale becomes much harder to justify on a purely financial basis.
Scaling reveals the structural weaknesses of in-house support models
The hidden costs of support in house operations do not compound linearly. They accelerate. When contact volume is predictable and relatively modest, an in-house team can manage with reasonable efficiency. When volume spikes seasonally, in response to a product launch, or following a service disruption, the model struggles to respond. Hiring takes weeks. Training takes longer. Quality typically drops before it recovers.
Specialist outsourced providers do not face the same constraint. They maintain pools of trained agents who can be deployed quickly, absorb volume spikes without quality degradation, and scale back when demand normalises. For businesses whose contact volumes vary throughout the year, which is most businesses, support in house means either overstaffing to cover peaks or understaffing and absorbing quality problems. Neither is a satisfying answer. For more on how this plays out operationally, our piece on scaling support without losing control gets into the specifics.
Support In House: Turnover is not a line item, it is a dynamic that affects everything else
High turnover is one of the most damaging hidden costs of support in house operations, and it tends to be accepted as an unavoidable feature of the contact centre environment. Industry figures put annual attrition between 30 and 45% in many in-house operations. At that rate, a 50-seat team is replacing 15 to 22 people every year. Each replacement carries recruitment cost, onboarding cost, and a period of reduced performance while the new hire builds competence.
The quality cost of that cycle is significant. Experienced agents resolve more issues on first contact. They know the product, the systems, and the common failure points. When this kind of support turnover is high, that experience is constantly being drained from the team and replaced with learners. The customer experience reflects this even when the metrics lag behind. Satisfaction scores do not capture it immediately but they do capture it eventually.
What the full cost picture actually looks like when you run the numbers honestly
When you build the complete cost model for support in house at scale, including salaries, benefits, recruitment, onboarding, technology, infrastructure, management overhead, opportunity cost, and the cost of quality degradation during scaling events, the comparison with outsourced models shifts considerably. The figure that appeared to favour in-house in the initial spreadsheet looks quite different once all the variables are included.
None of this means the answer is always to outsource everything. But it does mean the decision deserves a full analysis rather than a salary comparison that excludes the majority of the real cost. The businesses that make the best decisions are the ones that go into the analysis with accurate numbers and honest assumptions. In house process can make sense in specific contexts but those contexts are narrower than most initial calculations suggest.
Continue exploring the operational economics of customer support
If you want to go further with this analysis, Customer Experience Online has a growing library of content that looks at the real economics of customer support across different operating models. Whether you are currently running support in house and questioning the model, or considering a shift to outsourced or hybrid delivery, there is practical, detailed thinking available that can help inform the decision.
The goal is not to steer you toward any particular answer. It is to make sure you are working with the complete picture. Decisions made on incomplete cost information tend to produce outcomes that were never really in the budget. Getting the analysis right at the beginning is considerably less expensive than correcting it later.
And if you have been running support in house for a while and suspect the real cost is higher than the spreadsheet shows, you are probably right. The good news is that it is not too late to build a more accurate picture and make a decision based on it.
Frequently Asked Questions
The most commonly missed costs include recruitment cycle expenses, technology infrastructure and licence fees, management overhead, the cost of quality degradation during scaling events, and opportunity cost.
Costs do not scale linearly. Each growth stage adds management layers, QA complexity, technology requirements, and recruitment overhead. The costs that were manageable at small scale compound aggressively, particularly when contact volumes are variable and the team needs to flex up and down with demand.
Support in house makes genuine sense in specific contexts, typically where the work is highly complex, deeply embedded in proprietary systems, or requires regulatory oversight that makes third-party access inappropriate. Outside those contexts the full cost comparison often does not support the in-house model at scale.
It is often the largest cost category that does not appear in a budget. When senior leaders are managing a support in house team, that time is not available for product development, commercial strategy, or growth-focused work.
Build a full cost model that includes salaries, benefits, recruitment, onboarding, technology, infrastructure, management overhead, QA costs, and opportunity cost. Then compare that against the all-in cost of an outsourced model including transition costs and ongoing management time. The comparison often looks quite different from the initial headline salary comparison.
