Winning the Culture War: Attrition Is The Real Risk

In Article 3 we worked through the operational runbook for building a GCC — the legal, real estate, buildout, talent, vendor, knowledge transfer, and change management steps that take a center from idea to launch. In this final article we turn to what happens after the doors open. Building the center is the hard part, most leaders say. They are wrong. Building it is the easy part. Keeping it running well — maintaining quality, productivity, and a workforce that chooses to stay — is the real challenge, and the one most organizations underestimate until they are already losing. The single greatest threat to an operating GCC is attrition. Not technology. Not geopolitics. Not the quality of your real estate. Attrition. Everything else in this article connects back to that fact.

Why Attrition Destroys GCC Value

Many outsourcers run annual attrition rates of 20 percent or higher. I have consulted on a GCC project where a large multinational with 6,500 contractors in India was establishing its own center not to save money, but because 70 percent annual attrition had made stable staffing functionally impossible. They were perpetually training the next wave of people who would leave for a competitor across the street. The math is brutal. At 20 percent attrition in a 500-person center, you are replacing 100 people every year. Each replacement requires recruiting, hiring, onboarding, and a knowledge transfer period before the new hire reaches full productivity. The cost and distraction of that churn is enormous — and it compounds. Your most experienced people, the ones who train the new hires, spend a disproportionate share of their time doing so rather than executing the work they were hired for. At the three GCCs I created in India and Romania within a large financial services company (2,800 FTEs), we sustained attrition rates of 6 to 8 percent/year over 5 years — roughly one-third of industry averages. Approximately 30 percent of that advantage came from the fact that both centers were growing rapidly throughout their existence, which meant new hires from two years prior were finding promotion opportunities without having to leave. But the remaining 70 percent came directly from the culture we built, and that culture was deliberately designed before we hired the first employee.

Culture Is Not a Program. It Is a Decision.

The most important insight I can share about GCC culture is this: what takes months to build takes one or two bad policy decisions to destroy. Culture is fragile, especially in the early years, and must be guarded carefully. The instinct of many US-based executives is to treat the GCC as a cost center and manage it accordingly — minimizing benefits, densifying the office, capping salaries at the market median. Every one of those decisions sends a signal to your staff about how the organization views them. Those signals accumulate, and when they accumulate negatively, people leave. The alternative is to make an explicit, early decision to invest in your people — not lavishly, but genuinely. Here is what that looked like in practice in my 3 GCCs:

Compensation above market. We paid 5 to 20 percent above market rates, funded by part of the margin we were capturing from outsourcers. The cost of this premium was a small fraction of what we would have spent replacing the staff who would otherwise have left for slightly better offers. Understand market rates in your target city and stay ahead of them. Salary surprises are a primary driver of attrition, and they are entirely preventable. Meaningful benefits. In India, transportation to and from the office is not a perk — it is an expected benefit of employment. Organizations that do not provide it are immediately at a competitive disadvantage in the talent market. Beyond transportation, work with local specialists to understand which benefits are government-mandated, which are expected to be competitive, and which are meaningful extras. Many add little to overhead but generate significant goodwill. Recognition and spot bonuses. We made sure that every department meeting called out exceptional work, regardless of scale. A spot bonus given publicly in a team meeting costs relatively little and sends a clear signal that the organization notices and values contribution. This simple practice, repeated consistently, has a compounding effect on morale and retention. Employee-led working groups. At our India GCC, staff self-selected into working groups around topics they cared about — climate change, charitable giving, sports, office infrastructure, and others. These groups presented ideas to senior management monthly. The climate change group, for example, identified that the cafeteria was disposing of 1.5 to 2 tons of food waste per month. They analyzed the root cause — fixed meal trays that included items many staff did not want — and proposed a tasting station model that allowed staff to opt out of courses they would not eat. Food waste dropped to 300 pounds per month. The group was recognized in an all-employee meeting. That recognition mattered more to the team than the solution itself. The cumulative effect of these practices is a family-like environment. People leave companies. People leave managers. People generally do not leave their families. That is the standard worth aiming for, and the GCCs that achieve it create a durable competitive advantage that is genuinely difficult for outsourcers to replicate.

The Centralized vs. Decentralized Management Debate

Every GCC will eventually face a fundamental operating model question: where does command and control live? There are two models, and the debate between them is one of the most persistent in GCC management. In the centralized model, the GCC is managed as a unified resource pool under local leadership. When a resource has spare capacity, the local manager assigns additional tasks from across the organization. Advocates of this model argue — correctly — that it optimizes resource utilization. If a developer in Hyderabad finishes a sprint two days early, centralized management means that capacity is visible and deployable to whoever needs it. I have spent my career pushing back against this model, and for a reason that matters more to me than utilization rates: the disconnect between accountability and outcome. If I am the US-based leader of a function and my GCC team drops the ball, I have taken the accountability hit for a delivery failure that I did not have the management authority to prevent. That disconnect is fundamentally unhealthy and creates friction between US and GCC leadership that compounds over time. My strong preference is decentralization: GCC resources report into the functional group leader, wherever that leader is based. Local GCC management handles HR, government compliance, facilities, and the other locally-administered functions, but the work program and performance management sit with the group that owns the function. This model preserves clear accountability, reduces change management friction, and means that the leader who is accountable for an outcome has the authority to manage the people delivering it. In practice, many mature GCCs use a hybrid. Decentralization is the default operating model across the majority of functions. Centralization is applied selectively to functions with high workload variability — QA, analytical reporting, and similar roles where dedicated allocation to a single function would leave significant capacity unused. If a team regularly has variable demand, a shared service model under central management makes sense. If a team has predictable, full-time work, decentralization is almost always preferable. In my example, approximately 20% of the projects used the centers of excellence exclusively, 50% had dedicated resources (e.g. QA) augmented with center of excellence resources and 30% used only dedicated resources. The right answer is not ideological. It depends on the nature of the work. What matters most is that the right people weigh the trade-offs honestly for each function rather than applying a blanket policy that optimizes for one factor at the expense of others.

Metrics, KPIs, and the Discipline of Measurement

You cannot improve what you do not measure, and a GCC without a strong measurement framework will drift. We collected dozens of metrics across all of the GCCs I created and grew. But the discipline of measurement is not about collecting more data — it is about collecting the right data. Every metric and KPI should come with two things: a clear rationale for why it matters, and an action plan for what happens when it deviates from its baseline. A measurement that does not carry actionable information is not worth tracking. I have seen organizations waste enormous amounts of time monitoring metrics that, when they moved, triggered no meaningful response. If you cannot describe what you would do differently when a metric moves, remove it from the dashboard. The metrics that matter most in an operating GCC fall into a few categories:

• Attrition: monthly and trailing twelve-month, by level and function. Segment by voluntary vs. involuntary. Track time-to-fill for open roles.
• Productivity: output per FTE against defined baselines, with SLA adherence by function.
• Quality: defect rates, rework rates, and client satisfaction scores from the US- based stakeholders the GCC serves.
• Compensation competitiveness: quarterly benchmarking against market rates in your target cities. Do not let salary drift below market without knowing it.
• Culture indicators: participation rates in working groups, utilization of recognition programs, absenteeism trends. Review the full dashboard in a monthly operating meeting that includes both GCC leadership and the US-based functional leaders. Make the data visible and the conversation honest. The worst thing you can do is optimize your metrics for appearance rather than reality — it delays the identification of problems until they are significantly harder to fix.

Optimization Is Never Finished

A GCC is not a program with a completion date. It is an operating entity that requires continuous attention, refinement, and course correction. The organizations that treat GCC establishment as a one-time project and then shift their attention elsewhere are the same organizations that find themselves with high attrition, bloated teams, and eroding ROI three years later. Ongoing optimization means periodically reviewing the functions in your GCC against the salary arbitrage math. As your center matures and your GCC staff develop deep domain expertise, you will find new functions that are now candidates for transition offshore that were not viable candidates at launch. It also means reviewing team compositions and management spans with the same rigor you applied when setting up the center. It means staying current on the local talent market. The cities that are the right choice today may not be the right choice in five years as workforce dynamics evolve. In our Romania footprint, Cluj was not on our radar in our original plan. We originally only planned on Bucharest. It became our third center because we stayed close to where specialized talent was actually concentrating. And it means staying ahead of the AI-driven changes that will reshape the GCC landscape over the next several years. As I described in an earlier article in this series, AI code tools are expected to materially reduce the number of IT roles required to execute software development programs over the next two to five years. Those reductions will happen gradually, but organizations with GCCs should be planning for them now — negotiating lease flexibility, modeling headcount scenarios, and thinking carefully about which roles they are growing vs. which they should be managing toward natural attrition.

A Final Word on the GCC Opportunity

Across seven GCCs in three organizations, I have seen this model work at every scale — from 80 FTEs to 2,800. I have seen it deliver not just the cost savings that justified the investment, but things more durable: organizational capability and innovation. The GCCs I am most proud of did not just reduce a budget line. They gave their organizations access to talent, scale, and execution speed that would have been difficult to achieve any other way. The leaders who hesitate — who hire consultants to plan what an experienced team could execute directly, who spend years in analysis while their competitors build — pay an opportunity cost that is rarely calculated but always real. The salary arbitrage opportunity exists today. The talent markets are mature today. The playbook is proven. If you have more than 150 offshore resources and you do not own them, you are leaving money, control, and organizational capability on the table. The question is not whether a GCC makes sense. The question is how quickly you can move.