The Third Act of the Sourcing Story
The metric you choose is the institution you build — and most organisations default to the metric that justified the business case, which is the cost metric, which is the outsourcing metric.
The Celebration That Feels Familiar
The ribbon-cutting follows a pattern you could set your watch by. A senior leader flies in for the opening of the new Global Capability Centre. The local leadership team presents headcount ramp projections. Facilities tours showcase the collaborative workspaces, the innovation labs, the breakout areas designed to signal that this is not a back-office. The press release lands: strategic investment, talent access, long-term commitment to the region. The business case sits in a board pack somewhere, anchored to a cost-per-FTE comparison that made the investment self-evident.
Twelve months later, the questions start. Why is the centre running at eighty percent of planned headcount but delivering at sixty percent of expected output? Why do the most capable local hires keep leaving? Why does every decision of consequence still route through the home office? The answers, when they surface, are uncomfortable — because they describe an outsourcing arrangement with the vendor removed.
The Sourcing Trilogy
We have been here before, twice. The first act was the great outsourcing wave of the late 1990s and 2000s — cost arbitrage as strategy, delivered through managed-service contracts that promised transformation and delivered transaction processing. The second act was the repatriation movement of the 2010s, when organisations discovered that what they had outsourced was not just cost but capability, and began the painful work of rebuilding what they had hollowed out. The third act is the GCC boom: enterprises building their own captive centres, chasing the arbitrage they once contracted for while retaining the ownership they once surrendered.
The logic is seductive. Why pay a vendor’s margin when you can employ the talent directly? Why accept the vendor’s priorities when you can set your own? The GCC promises the cost advantage of Act One without the capability loss of Act Two — ownership and arbitrage in a single model.
But the promise holds only if the organisation understands what it is actually building. And the pattern I have watched unfold across a dozen capability-centre integrations tells me that most do not.
The Design Test
The difference between a GCC that works and one that quietly fails is not location, not talent pool, not even investment — it is design intent. And design intent reveals itself not in the strategy deck but in four operational realities that are visible within the first year.
Where decisions are made. In a capability institution, decision authority migrates with the work. The centre’s leaders own product roadmaps, make architecture choices, approve releases. They are accountable for outcomes, not activities. In a rebranded outsourcing arrangement, the centre executes against requirements written elsewhere. Every material decision routes back to the home office. The local leadership team is senior in title and subordinate in practice — and the best people notice this within their first quarter.
What gets measured. The most reliable tell is the metric structure. A GCC measured on seats filled, utilisation rates, and cost-per-FTE is an outsourcing contract wearing a different badge. A GCC measured on cycle time, defect rates, customer outcomes, and the revenue impact of what it ships is something else entirely. The metric you choose is the institution you build — and most organisations default to the metric that justified the business case, which is the cost metric, which is the outsourcing metric.
How careers work. The talent markets in Bangalore, Hyderabad, Kraków, and São Paulo are sophisticated, and the best practitioners have choices. A centre that offers career progression only into “senior individual contributor, offshore” will haemorrhage its most capable people to competitors — or to the very vendors the GCC was meant to replace. The centres that retain talent are the ones that offer genuine career paths: into product leadership, into architecture, into roles that shape strategy rather than execute it. This requires the home organisation to accept that its most senior technologists and domain experts may sit in Pune rather than London — and that is a cultural concession many organisations are not prepared to make.
Whether the work is designed or transferred. Activity transfer by spreadsheet — mapping existing processes onto new teams without redesigning them for the new operating model — is the original sin of outsourcing, and GCCs inherit it wholesale. The centre receives a process designed for a different context, a different proximity to the business, a different set of assumptions about collaboration. It then operates that process at a structural disadvantage and is held to the standards of the original. The organisations that get this right invest as much in work design as in facilities — rethinking what the centre should own end-to-end, what interfaces need to change, and what the home organisation needs to stop doing.
The design test is not whether the centre has good people — it almost certainly does. The test is whether the organisation has designed a model in which good people can do good work, or one in which they are structurally prevented from doing so.
The Culture That No One Names
Underneath these operational realities sits a cultural one that is harder to name and harder to fix. Most GCCs operate within a two-tier culture — and the organisation’s leadership would deny it if asked.
The first tier is the home office: the place where strategy is set, where relationships with customers and executives are maintained, where the important meetings happen in the right time zone. The second tier is the centre: capable, growing, increasingly essential to delivery — and subtly subordinate in every interaction that matters. The subordination is rarely explicit. It lives in which meetings are scheduled for whose convenience, whose context is assumed in communications, who gets the interesting problems and who gets the overflow. It lives in the architecture reviews where the centre’s lead joins by video to present work that was shaped in a corridor conversation she was not part of. It lives in the annual planning cycle where headcount targets for the centre are set before its leadership is consulted — because the centre is a line item in someone else’s budget, not an institution with its own strategic voice.
We have seen this pattern before. It is precisely the dynamic that made the best outsourcing relationships fragile — the vendor’s team was capable but never quite trusted, never quite inside the circle. The GCC was supposed to solve this by bringing the team in-house. But in-house is a legal status, not a cultural one. An organisation that treats its Hyderabad engineers as a cost-optimised resource rather than a peer capability will produce exactly the engagement, retention, and quality outcomes that framing deserves.
The Repricing Nobody Budgeted For
And then there is the tension that most GCC business cases have not yet confronted.
The economic foundation of the GCC model is labour-cost arbitrage. A software engineer in Bangalore costs a fraction of one in San Francisco or London. The arithmetic is straightforward, the savings are real, and the business case writes itself. But that arithmetic assumed a stable relationship between headcount and output — and AI is breaking that assumption in real time.
The coding assistants and automation tools now entering production use are not replacing engineers wholesale, but they are changing the output curve. A team of forty, well-tooled, is beginning to deliver what once required sixty. The GCC that opened eighteen months ago with a five-year ramp to eight hundred engineers may find, by year three, that its optimal size is four hundred — and that the business case built on eight hundred heads no longer computes. The ratio is shifting quarter by quarter, and the trajectory is clear even if the endpoint is not. The implications for the GCC model are profound and largely unacknowledged.
If the business case is built on labour arbitrage, and AI compresses the labour requirement, the centre’s economic rationale erodes from underneath — not dramatically, not overnight, but steadily. The centres most vulnerable are precisely those designed as outsourcing contracts without the vendor: high-headcount, activity-based, measured on utilisation. The centres most resilient are those designed as capability institutions — because their value is not denominated in seats but in outcomes, and outcomes do not diminish when fewer people deliver more.
This is the conversation that GCC leadership teams need to be having now, while the build programmes are still underway. Not how many seats do we fill this quarter? but what is the centre’s value proposition in a world where AI has halved the labour-to-output ratio? The answer, if the centre is well-designed, is that it becomes more valuable — a concentrated pool of high-calibre practitioners who leverage AI tools to deliver disproportionate impact. The answer, if the centre is a headcount play, is that the business case unravels within three years of the ribbon-cutting.
The Question That Matters
The GCC boom is real, and many of these centres will succeed. But the ones that succeed will be the ones that understood, from the beginning, that they were building an institution — not executing a sourcing strategy by other means.
The test is simple, even if the execution is not. Measure the centre on outcomes it owns. Build career paths that are genuinely competitive with the local market. Migrate decision authority with the work. And be honest — radically honest — about what AI is doing to the labour economics underneath the entire model.
The sourcing story has three acts. The third does not have to end the way the first two did. But it will, unless the organisations writing the business cases stop long enough to ask what, exactly, they are building — and whether the answer is different enough from what came before.