A foreign company wants a few hundred people in India doing finance operations, support or analytics. It has three broad routes. It can outsource the work to a BPO, which runs it under a contract. It can set up its own captive centre, now commonly called a Global Capability Centre (GCC). Or it can use a Build-Operate-Transfer arrangement, where a partner builds and runs the centre for a defined period and then hands it to the company. I will compare them on what actually differs, and then say which one I think is overrated and which underrated.

What each model really is

Outsourcing. The vendor owns the premises, hires the staff, and the client buys an outcome: seats, transactions, resolved tickets. The client does not carry the lease or the payroll. It carries the vendor relationship.

Captive or GCC. The company establishes its own legal entity or branch in India, hires its own employees and runs the work as an extension of the parent. Everything is owned: people, process, intellectual property, and also every headache from compliance to canteen contracts.

Build-Operate-Transfer. A specialist provider handles the early phases: finding space, hiring, setting up governance and operating the centre. After an agreed period, often a few years, the client takes it over, typically by buying the entity or taking on the employees and assets. The terms of that transfer, including price and what is included, should be fixed at the start, not at the end.

Where they differ

QuestionOutsourcingCaptive / GCCBOT
Speed to startFast, vendor already has capacitySlowest, entity, hiring and leadership from zeroFaster than captive, slower than outsourcing
Upfront commitmentLow, mostly contract termsHighModerate, rising towards transfer
Control over people and cultureLimitedFullShared at first, full after transfer
Knowledge stays withVendor, partlyCompanyCompany, increasingly
Exit flexibilityEasier, if contract allowsHard; closing a centre means severance and reputationBuilt into the plan
Suitable scaleAny, including smallUsually needs enough scale to justify its own leadershipMid scale and growing

The scale row deserves attention. A centre has fixed costs: a head, finance, HR, facilities, compliance, IT. Spread across 40 people those fixed costs are a heavy tax; spread across 800 they are noise. Any comparison that ignores this is a brochure.

The case people make for captives, and where it frays

The pitch for a GCC is control and capability. The company wants engineers, analysts and specialists who understand the business deeply, who can take on more valuable work over time, and who are not shared with other vendor clients. This is a good argument for knowledge-heavy work: product engineering, analytics, risk, core finance. For high-volume transactional work it is weaker. Running a contact centre well is a specialised craft with its own forecasting, workforce, quality and attrition management. A company whose business is insurance or retail does not automatically get good at that by owning seats. Some captive centres end up doing just what a vendor would, at higher cost and with weaker operations discipline, because the parent never made it a priority.

Also, owning people means owning the cost of unwinding. If the business changes direction, a vendor contract can be reduced or ended according to its terms. A captive centre must be restructured, and that is slow, expensive and visible.

The case for outsourcing, and where it frays

Outsourcing wins on speed, flexibility and operational maturity. A good vendor brings tools, recruitment machinery, trained managers and a view of what other clients are doing. It frays in two places. First, dependency: after a few years the client may no longer know how the process truly works, and the vendor holds the knowledge. Switching becomes a transition project with real risk. Second, alignment: the vendor earns from seats or volumes, while the client wants outcomes and falling cost. Contracts try to bridge that gap with SLAs and incentives, and they only partly succeed.

Why BOT is attractive, and what it is not

BOT answers a particular worry: “we want our own centre eventually, but we do not know how to start one in a country where we have no team.” The partner absorbs the early learning, and the client gets a runway to build its own leadership. When it works, the transfer is a handover of something already stable, not a leap.

But BOT is not a magic trick. The partner’s incentive to run the centre well is clear; its incentive to prepare the client for transfer is less so, since the transfer ends its revenue. That is the central design problem. Good contracts define transfer readiness milestones, require the client to place its own managers in shadow roles during the operate phase, and spell out the price formula. Vague BOTs turn into long outsourcing deals with a promise attached. Also be careful with the legal and tax side: how the entity is structured and what is transferred can have tax and employment consequences, and it should be reviewed with advisers, not assumed.

My verdict

For a company entering India with fewer than a few hundred people and uncertain plans, outsource first, with a contract that protects documentation and exit. For knowledge-intensive work the company expects to grow and keep for a decade, go captive, but only when it can name a real India leader and commit the scale. In between, BOT is the most underused option, and I think it suits the “we are fairly sure but not certain” case better than either extreme, provided the transfer terms are written early and tested.

The overrated option, in my view, is the captive for routine transactional work. It sounds strategic. Often it is outsourcing done by a team that has never done it. The best operators will say the same thing about their own models in reverse, which is why the question to ask is not “which model is best” but “what do we want to be good at ourselves”.

Leave a Reply

Trending