New operating models: how ER&D centres are restructuring the leadership bench

We mapped how the ER&D centres we work with are restructuring, and the through-line is a migration from staff-augmentation to product ownership. In the old model, the centre supplied capacity against the parent’s roadmap. In the new one, the centre owns pieces of the roadmap outright: the architecture, the trade-offs, the accountability for whether the thing actually ships and works.

That migration happens through who you put in the senior seats. A staff-augmentation centre is led by people who are superb at throughput and utilisation. A product-owning one needs people who can hold a technical vision and defend it against the parent when it matters.

The scale gives the shift its weight. India’s GCC sector already runs past 1,700 centres employing roughly 1.9 million people, and ER&D is where the value is climbing fastest, centres are being chartered to own engineering. When a mandate changes from “deliver against a spec” to “own the spec,” the person you needed to lead it changes with it, and most of the restructuring pain traces back to that one substitution not being made.

The visible change is titles: more principal engineers, product directors, architecture leads. The real change is what those titles have to be able to do. Product ownership pulls the whole bench upward: the centre needs people who can make roadmap calls, own technical debt as a strategic choice, and translate between engineering reality and business intent without a translator from the parent.

These leaders are scarce, and they come from a different search entirely, often from product organisations rather than services ones. Promoting the best delivery manager rarely produces one. ER&D centres that understand this are hiring for the operating model they’re becoming. The ones that don’t are decorating an augmentation bench with product titles and wondering why ownership never quite lands.

Getting ahead of that is its own discipline. A product-owning centre has to hire for a charter it’s still growing into, which is why the sharper GCCs treat the leadership bench as something to build before the mandate arrives rather than after. That logic, hiring ahead of the charter, is what separates a centre that absorbs a bigger mandate from one that stalls under it.

It helps to be concrete about the gap, because “more senior” isn’t the same as “different.” A delivery leader is measured on whether the committed work landed on time and on quality. A product-owning leader is measured on whether the right work got chosen in the first place, and that’s a harder thing to hire for, because it rarely shows up cleanly on a services CV.

Three capacities separate the two. The first is roadmap judgment: the willingness to say a parent’s request is the wrong build and to hold that line with the reasoning to back it. The second is owning technical debt as a strategic instrument rather than a hygiene metric, knowing when to take it on deliberately and when it’s become a liability. The third is a P&L instinct: reading the cost of an engineering choice against the value it returns, without the parent’s finance function pre-digesting the trade-off. A leader who has only ever run to an SLA has rarely had to build any of these, which is why the search for them looks so different.

This is also where compensation stops tracking the old benchmark. You’re now paying for people who could hold a product line at the parent’s headquarters and have chosen to hold it here instead. The band for that person sits well above the delivery-lead band the centre is used to, and the centres that win these hires have already reset the number rather than discovering the gap at offer stage.

Product ownership is one new operating model. There’s a second, further along the same road, and it changes the hire even more sharply: the centre that starts running charters for peers as well as the parent. A captive centre has exactly one customer. The moment it shares a scarce capability outward, running work for peer organisations, it acquires a second, and everything about the talent model shifts. This is the shape people mean by GCC-as-a-service, and it’s showing up first in the centres whose capability is genuinely rare.

Suddenly the centre needs people who can hold a client relationship, scope work that isn’t handed to them, and own a commercial outcome rather than a delivery one. That’s a leader the captive model never had to hire. In a pharma centre, the regulatory-and-quality depth the parent always demanded stays non-negotiable, but now it has to sit alongside a commercial instinct the centre used to outsource entirely to the parent.

The judgment moves too. Inside a captive, a lot of judgment lives upstream: the parent decides what’s worth doing and the centre executes cleanly. Selling capability outward repatriates that judgment. The centre has to price its own work, decide which peers to serve, and stand behind outcomes for organisations that owe it no loyalty. That judgment isn’t free, and it doesn’t come from the existing bench: you’re, in effect, hiring people who could run a services business, because that’s what the centre has quietly become.

The failure mode is rarely dramatic. A centre announces product ownership, keeps its delivery leaders in place, and rebadges them with product titles. For a while nothing looks wrong. Then the first hard roadmap call arrives, the parent wants a build the centre knows is the wrong one, and the leader defaults to the only muscle they have, which is to execute cleanly and on time. Ownership never lands, because ownership was never staffed.

The same trap catches the GCC-as-a-service move. A centre wins a peer charter and asks a delivery lead to hold the client relationship. The regulatory depth is there, the quality is there, and the commercial judgment simply isn’t, because it was never part of the job. In pharma, where the trust bar is unforgiving, that gap is the difference between a service line and a liability, and it shows up in a client that quietly stops extending the charter, never in a status report.

The parent organisation is often the last to see it. It reads a single headcount line and a single delivery metric and assumes the operating model changed because the memo said it did. The org chart is the wrong place to look. The restructuring is real only where the senior bench actually changed, and that’s the part a reorg deck tends to skip.

The centres making these transitions cleanly do one thing early: they name the operating model they’re moving to and hire the leadership bench for that model before the mandate fully arrives. If the charter is product ownership, the senior search runs into product organisations and looks for roadmap judgment and P&L reading. If the charter is running work for peers, the search adds client-facing and commercial ownership to the regulatory depth that was always required.

Practically, that means two moves. Reset the compensation band before you open the search, because the leader you now need sits above the delivery band the centre is used to, and pretending otherwise loses the candidate at offer stage. And read each senior requisition for the model it belongs to, a delivery lead and a product owner can read almost identically on paper, and the interview is where you find out which one the role actually needs. Do that and the restructuring becomes the thing it was always meant to be, which is a change in who the centre trusts to decide.