Eleven cities. Forty indicators. One ranking — for your role mix, not your competitor's.

Most boards agonise over the first city and never model the second. Yet the second site is where the economics of an India center are actually decided. The first site is chosen for depth: the leaders you can hire, the peers you can poach from, the ecosystem that makes a relocation offer credible. Depth is expensive, and it should be. It buys you a site head, a first hundred engineers and a brand in the market.

The second site is chosen for leverage. By the time a center reaches two to three hundred seats, the work has stratified: a core that needs the deepest bench and a much larger layer of application, platform, QA and operations roles that do not. Keeping that layer in a Tier-1 corridor at Tier-1 rents and Tier-1 attrition is the single most common way a center that looked cheap on the board paper becomes expensive by year three.

The index on this page is built to be run twice: once with a role mix weighted to depth, once weighted to cost and retention. If the same city wins both, build one site. If different cities win, plan the second site into the operating model before the first lease is signed — the entity, payroll and compliance spine already stretch across states, and the marginal cost of a spoke is a fraction of what the herd assumes.

Between 41 and 46 percent of new capability centers in the last year chose Hyderabad. That is a remarkable number for a city that held a fifth of the market a few years ago, and it invites the obvious question: is it a signal or a stampede?

The case for signal is strong. Operating costs sit 15–20 percent below Bengaluru, the state government has run a consistent policy for a decade, Grade-A supply kept pace with demand, and the BFSI and pharma clusters give a new center peers to hire from on day one. Half of last year's new BFSI entrants chose it, and for that sector the choice is close to rational by default.

The case for herd is about what 41 percent does to the leadership market. Senior engineering managers, site heads and heads of function are a thin layer in any city, and forty new centers competing for the same two hundred people bids their cost up and their tenure down. Our advice is unglamorous: if your role mix is BFSI or multi-function operations, Hyderabad probably wins and the index will say so. If it is AI-first product engineering, run the index before you follow the crowd — Bengaluru's premium buys a bench nobody else has, and Pune's discount buys a very good one.

A cost index compares a fully loaded seat in one city with the same seat in another. It is a useful number and it is on this page. It is also the number most likely to mislead a board, because it describes a center that already exists and says nothing about the year it takes to build one.

The twelve months between the board's approval and a productive floor carry costs that never appear in a per-seat index: a site head hired in month three whose team arrives in month eight; a lease signed for a floor that fills in three cohorts; consultants, lawyers and recruiters working in parallel on separate statements of work; and, largest of all, the opportunity cost of the work that was supposed to have shipped. Setup-month opportunity cost is the line in the ROI calculator that clients most often ask us to remove, and the one we most often refuse to.

This is the honest argument for Build-Operate-Transfer. It does not make India cheaper per seat — the index does that. It makes the twelve months shorter and the cost of them predictable, because one partner is accountable for the critical path and carries the lease, the leadership guarantee and the hiring-velocity penalty. Model the seat cost, by all means. Then model the year.