For twenty years, business process outsourcing sold labour arbitrage. The pitch was arithmetic: the same work, performed by people who cost less, in a location where property and infrastructure also cost less. Providers competed on rate card and scale, and the economics were compelling enough that few buyers asked what else they were buying. By around 2010 that model was running out of room. Wage inflation in established offshore locations was compounding, attrition was eroding delivery quality, and the margin available from moving a seat from one country to another had largely been captured. The response reshaped the industry. Providers stopped selling headcount and started selling platforms — technology that performed the work, with people attached to handle what the technology could not. Back office as a service, in the marketing language of the period.
Why the Labour Model Stalled
Three pressures arrived at once. Arbitrage narrowed. Salaries in mature delivery centres rose faster than in client markets year after year. A ten-year contract priced on a labour differential that shrank annually became progressively less attractive to the buyer and less profitable to the provider. Attrition undermined quality. High turnover in transactional roles meant that process knowledge never accumulated. Clients paid for experience and received a rotating population of recently trained staff, with error rates that reflected it. Buyers stopped being impressed by cost alone. After a decade of outsourcing, finance leaders had learned that a lower unit cost on a badly designed process is still a badly designed process. Demand shifted toward providers who could improve the work rather than merely relocate it. The providers who saw this early made a strategic bet: build the technology once, deploy it across many clients, and price on outcomes rather than hours. A platform amortises across a client base in a way that a delivery centre never can.
What a Platform Model Actually Changes
The difference is not that the provider has software. Every outsourcer had software. The difference is where the intelligence sits and who owns the improvement. In a labour model, the process lives in the client's systems and the provider supplies hands to operate it. Improvement means the provider suggests changes the client must approve and implement. Productivity gains accrue slowly and are contested at every review. In a platform model, the provider brings a workflow, rules engine, document capture, exception routing and reporting layer that sits between the client's systems and the work. The provider controls the process, which means it can improve it without a client change programme, and it absorbs the technology investment because the same platform serves dozens of customers. That shift has real consequences for buyers. Pricing can move to outcomes. Cost per invoice processed, per claim settled, per employee paid — or in more mature arrangements, gain-share against error reduction or cycle time. This aligns incentives better than an hourly rate, which rewards the provider for taking longer. Implementation is faster. A configured platform deploys in weeks where a bespoke transition took quarters. Benchmarking becomes real. A provider running the same process for forty clients knows what good looks like, which is information no single client possesses about itself. And lock-in deepens considerably. This is the part buyers consistently underweight. When the process logic lives in the provider's platform, the client no longer owns an operable description of how its own work is done. Exit means rebuilding capability that was never documented internally.
| Question | Labour model | Platform model |
|---|---|---|
| Where does the workflow live? | In the client's systems | In the provider's workflow and rules layer |
| What does the provider supply? | People operating the process | Technology with people handling exceptions |
| Who can change the process? | Client approval and implementation required | Provider controls its platform improvements |
| What must the buyer retain? | Understanding of its own process | Process definitions, configuration knowledge and exit rights |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
The Build, Buy or Outsource Decision
The platform model changed the decision framework but did not remove the need for one. Three questions determine the answer. Is the process differentiating? If how you do it affects what customers choose, keep it and invest in it. If it is invisible to customers and identical across your industry — payables, payroll, statutory filing, reconciliations — there is little strategic value in owning the machinery. Do you have the scale to justify the technology? A platform for a process you run two hundred times a month will never repay its build cost. A provider running it two hundred thousand times across its client base can. Volume, not capability, is usually the deciding factor. Can you tolerate the exit cost? Every platform arrangement creates dependency. The question is not whether it exists but whether you have priced it, documented enough internally to rebuild, and negotiated the terms that make leaving possible. The common failure is answering the first two and ignoring the third, then discovering at renewal that the alternative to accepting the provider's price increase is an eighteen-month insourcing programme nobody has budgeted.
Practical Rules for Platform-Based Outsourcing
- Own your data model, not just your data. Contractual rights to extraction are insufficient if the process logic and configuration live only in the provider's environment. Require documented process definitions as a deliverable.
- Price on outcomes, and define them precisely. Cost per unit is only meaningful if the unit is defined and the quality standard is measurable. Otherwise you have moved the argument from hours to definitions.
- Keep retained capability deliberately. A small internal team that understands the process, reads the exception reports and challenges the provider is the difference between managing a service and receiving one.
- Negotiate exit before implementation. Data formats, transition assistance obligations, duration, cost and the provider's duty to document configuration. Cheapest to secure before you have migrated.
- Measure the leading indicators yourself. First-pass yield, exception categories and aging tails should reach you directly, not filtered through a provider's monthly deck.
- Test the benchmark claim. Ask which specific metrics the provider's client base achieves and how yours compares. A platform's main advantage is comparative data; make them show it.
- Watch sub-processing and residency. Platform providers use cloud infrastructure and specialist subcontractors. For GCC-regulated work, where the platform runs and who can access it are gating questions.
- Review the make-or-buy decision every renewal. Volume, technology cost and internal capability all change. A decision that was right five years ago may not be right now.
Where It Ended Up
The platform model won, and the logic behind it has only intensified. Providers now compete on automation rate, data quality and analytics rather than on seat cost, and the best of them run processes measurably better than most clients could in-house. The current iteration is AI-enabled service delivery, and the economics rhyme exactly with 2010. Investment in models, training data and orchestration amortises across a client base. Pricing moves further toward outcomes because the labour input becomes harder to relate to the result. And the dependency deepens again, because the provider's advantage is increasingly embedded in accumulated process data that the client never sees. The buyer's discipline is unchanged. Outsource the work, keep the understanding. An organization that cannot describe how its own processes function has not bought a service; it has rented a capability it can no longer evaluate.
Common Questions
What does back office as a service mean?
A delivery model where the provider supplies a technology platform — workflow, rules, capture, exception handling and reporting — that performs the process, with people handling exceptions, rather than supplying labour to operate the client's systems.
Why did labour arbitrage outsourcing decline?
Wage inflation in offshore locations narrowed the cost differential, high attrition eroded process knowledge and quality, and buyers increasingly wanted process improvement rather than relocation of the same work.
What is the main risk of platform-based outsourcing?
Deeper lock-in. When process logic and configuration live in the provider's platform, the client loses an operable internal understanding of its own processes, making exit expensive and slow.
How should organizations decide what to outsource?
By asking whether the process differentiates them commercially, whether their volume justifies owning the technology, and whether they can tolerate and have contractually managed the exit cost.
BOaaS Strategy Session — Outpace tests which of your processes genuinely belong on someone else's platform, negotiates the outcome pricing and exit terms before you lose leverage, and keeps enough capability in-house that you can still evaluate what you are buying.