The GCC Talent Playbook: Hiring Your First 50, Attrition Benchmarks and Compensation Design

Hiring your first 50 people in India is not a sourcing problem. India has the candidates. It is a sequencing and arithmetic problem, and the two variables that break most plans are the notice period and the offer acceptance rate. Neither appears on a recruiter’s dashboard, and both are set by the market rather than by you.

TL;DR: Build the hiring plan backwards from the joining date, not forwards from the requisition. Indian notice periods run 30 days at junior level to 90 at senior, and offer acceptance for senior roles is commonly cited at 55 to 65 percent, so a plan for 50 joiners needs materially more than 50 offers and a calendar that starts a quarter earlier than feels necessary. On attrition, Aon’s Annual Salary Increase and Turnover Survey 2025-26 puts overall India attrition at 16.2 percent in 2025, down from 17.7 percent in 2024 and 18.7 percent in 2023. On pay, the same survey projects a 9.3 percent salary increase for Global Capability Centres in 2026, marginally above the 9.1 percent all-India projection. Design the compensation structure before the first offer goes out, because the second hire’s package is set by the first one.

The Arithmetic That Breaks First-50 Plans

Two numbers govern the calendar, and both work against you.

Notice periods. Indian notice periods are among the longest in any major economy. Market practice runs to roughly 30 days for entry and junior roles, 60 days for mid-level engineers and designers, and 90 days for senior engineers, architects and leadership. Research by Analytics India Magazine on notice-period practice found that close to one in three IT roles carries the full 90-day term. Buyouts exist, are common in the IT sector, and are frequently funded by the hiring employer as a joining bonus, but they are a negotiation rather than a right.

Offer acceptance. Offer acceptance for senior roles in India is commonly reported at around 55 to 65 percent, well below comparable US figures. The reason is structural rather than cultural. A 90-day gap between acceptance and joining is 90 days in which a counter-offer can arrive, a competing process can conclude, or a candidate can simply change their mind.

Put those together and the planning consequence is uncomfortable but simple. A cohort of 50 joiners is not 50 offers. It is closer to 75 to 85 offers, made across a window that starts roughly one quarter before you want people at desks, with a deliberate over-offer allowance at senior levels where both the notice period and the drop-out rate are worst.

The mistake I see repeatedly is a plan built forwards. Requisition approved in January, recruiter assigned in February, offers in March, and a leadership team surprised in June that the team is half-built. Build it backwards from the date the work has to start, and the January conversation becomes a different one.

Sequencing the First 50

The order matters more than the pace. A GCC that hires 50 people in the wrong sequence gets a team that cannot absorb work; one that hires 20 in the right sequence can.

Hires 1 to 5: the anchors. The site leader, the first engineering or functional lead, and an HR business partner who can actually recruit rather than only administer. These are the hardest, slowest and most consequential hires, and they should be closed before volume hiring opens, because everyone who follows is recruited by them and evaluated against them. Expect a 90-day notice on all of them and start six months out.

Hires 6 to 20: the spine. Senior individual contributors who will set the technical or process standard, plus the finance and compliance capability the entity legally requires. This cohort defines what “good” looks like in the centre. Hire below standard here and you spend two years correcting it.

Hires 21 to 50: the build. Mid-level and early-career hiring at volume, now possible because there are people to interview, onboard and manage them. This is where cost advantage is actually realised, and where a structured campus or early-career channel starts to pay.

Two disciplines hold this together. First, do not open volume hiring before the spine is in place, however much pressure there is to show headcount. Second, hold the standard on hires 6 to 20 even when the calendar slips, because a compromise hire in that band is visible in the centre’s output for years.

Where the build is smaller than 50, the same sequence applies with the same anchors and a compressed middle. Our explainer on nano GCCs sets out where that floor sits, and the Hyderabad GCC we set up for a US-listed logistics company followed this shape in practice.

If you are running hiring through a partner rather than an internal team, the operating question is whether you are buying requisition throughput or a hiring engine, which we have written about in RPO 2.0.

Attrition Benchmarks: What Good Actually Looks Like

Here the data needs handling with some care, because the numbers circulating on this topic disagree.

The reliable anchor is Aon’s Annual Salary Increase and Turnover Survey 2025-26 India, now in its 32nd edition and drawing on more than 1,400 organisations across 45 industries. It puts overall India attrition at 16.2 percent in 2025, following 17.7 percent in 2024 and 18.7 percent in 2023, which Aon describes as returning close to pre-COVID levels.

That trend line is the useful part. Attrition in India has been falling steadily for three years, and a retention plan built on 2022 assumptions is calibrated to a market that no longer exists.

On GCC-specific attrition, I would be straight with you: there is no authoritative public figure. Estimates published by recruitment and advisory firms cluster below the all-India average, with several putting GCC attrition in the 12 to 14 percent range, while at least one vendor source puts GCC voluntary attrition at 16 to 22 percent. Those are not reconcilable, and both are commercially interested. What the sources do agree on is directional: GCC attrition sits at or below the national average, and specialist roles in AI, machine learning and senior engineering run materially hotter than the centre average.

So use the all-India 16.2 percent as your benchmark and adjust for your own role mix, rather than quoting a GCC-specific number you cannot defend.

Three practical points on measurement, because most attrition reporting is not useful.

Separate voluntary from involuntary. They are different problems with different owners, and blending them hides both. Separate regrettable from non-regrettable, which requires someone to make a judgement per exit and record it. And report attrition by tenure band, because a centre losing people at month nine has an onboarding and expectations problem, while one losing them at year three has a career-path problem. The interventions have nothing in common.

Compensation Design: Four Decisions Before the First Offer

The second hire’s package is set by the first one. That is the whole argument for doing this before you start rather than after.

Decide the band architecture first. How many levels, what the salary range is at each, and how wide the overlap is between adjacent bands. Narrow bands force promotions to deliver pay movement, which creates title inflation. Very wide bands create defensibility problems when people compare. This is a structural choice and it is expensive to change once fifty people sit inside it.

Decide the fixed and variable mix, by level. Junior populations in India generally value certainty; senior populations will trade fixed for meaningful upside. A single mix applied across all levels satisfies nobody. Whatever you choose has to survive the statutory wage definition under the Code on Wages, which treats excluded allowances above one-half of total remuneration as wages for provident fund and gratuity purposes, so the structure is a compliance decision as much as a market one. Aon’s own commentary on the survey noted that the standardised wage definition and expanded social security provisions are prompting many employers to reassess and restructure compensation.

Decide the long-term incentive position early. This is the question GCCs get asked and are least prepared for, because the parent’s equity plan was designed for a different market. Whether you extend long-term incentives beyond leadership to scarce-skill individual contributors is a genuine strategic choice, not an administrative one, and it is materially easier to design at the start than to retrofit once a competitor has forced the issue.

Decide the benchmarking cadence and the counter-offer policy, in writing. Benchmark annually against a defined peer set, not against whatever a candidate claims. Then write down what you will do when a valued person resigns with a competing offer. Organisations without a written position make it up under pressure, pay above band, and teach the whole centre that resigning is the fastest route to a raise. That lesson travels through a 200-person site in about a fortnight.

On the pay-movement question, Aon projects the following for 2026.

The gap at the bottom of that table is worth sitting with. GCCs are projected to raise pay at 9.3 percent while technology consulting and services sits at 6.6 percent. That differential is a large part of why GCCs recruit successfully from IT services, and why IT services firms increasingly compete on other things.

For the employer-cost side of the equation, statutory on-costs and the city-by-city picture are covered in our GCC cost benchmark, which is the number to model against rather than gross salary.

Three Metrics Worth Reporting to the Board

Most GCC talent reporting is a headcount chart. These three are more useful.

1. Offer-to-join ratio, by level. It is the earliest available signal that your package or your process is not competitive, and it moves months before attrition does.

2. Regrettable attrition, by tenure band. Total attrition tells you very little. Regrettable attrition at nine months and at three years are two distinct diagnoses.

3. Time from requisition to desk, including notice. Not time to offer. The number that matters to the business is when the work starts, and the notice period is the largest component of it.

A Closing Note

Every organisation I have worked with treats the first fifty hires as a recruitment exercise and discovers eighteen months later that it was an organisation design exercise.

The people you hire first set the standard, the pay bands, the working norms and the reputation of the centre in a market where engineers talk to each other constantly. Get the sequence right and the next two hundred hires get easier. Get it wrong and you spend the following two years correcting a culture that was set by whoever happened to be available in month three.

People strategy is business strategy in execution. In a GCC that is not an abstraction, because the centre has no product, no customers and no revenue of its own. It has capability, and capability is entirely a function of who you hired and whether they stayed.

If you are planning a first cohort or reviewing a compensation structure that has drifted, Astravise Services works on organisation design, compensation architecture and HR governance for centres scaling in India, including through our fractional CHRO advisory. Start with the band structure. It is much cheaper to design than to unwind.

^{[1][2][3][4][5]}