What the 75th percentile is actually buying you

Start with what the number actually is. The 75th percentile of a salary band means that, in the surveyed sample, three-quarters of the people in that job code are paid at or below this figure and a quarter are paid above it. It's a position on a distribution of other employers' decisions. That's all it is. It's a photograph of what a market did, sorted from low to high, with a line drawn three-quarters of the way along.

It isn't a valuation, though. It carries no information about the person in front of you: their scarcity, their track record, the counter they can raise, the cost to you of not hiring them. It can't, because it was computed before that person existed as a candidate. Anchoring an offer to the 75th percentile treats a summary statistic as if it were a judgement about an individual. It isn't one, and the gap between the two is where senior hires are lost.

“We target the 75th” is a comfortable sentence. It signals seriousness, it survives a compensation-committee review, and it outsources the hard judgement to a survey. The 50th sounds cheap; the 90th sounds reckless; the 75th sits in the reassuring upper-middle, aggressive enough to feel competitive and moderate enough to feel responsible. It's the percentile you reach for when you want to look like you've made a call without having made one.

But it answers the wrong question. It measures how competitive you want to look. What closes a candidate is a harder question: what's this particular person worth to us, against what they can actually get elsewhere? The percentile ducks that entirely. And the mechanics make the duck worse: the distribution it's drawn from is already lagging, so the 75th of a stale basket can sit below the median of the live one. You feel aggressive and land ordinary, roughly a rung or two behind where the candidate is actually being courted.

The candidates who matter at the leadership and specialist level are, almost by definition, in the top slice of their own distribution. That's what makes them the ones you want. So when you set an offer at the 75th percentile of the whole population, you're, structurally, aiming below the part of the curve the interesting people occupy. The tail is where the scarcity is, and scarcity is exactly what a percentile computed across everybody averages away.

That tail is thickening fastest in the roles GCCs most need to fill. Demand for AI talent in India is projected to exceed 1.25 million by 2027, from a base of roughly 600,000–650,000 in 2022, according to the Deloitte–Nasscom report on bridging the AI talent gap, which also found only about 16% of IT professionals were AI-skilled. When demand roughly doubles and the qualified pool stays thin, the top of the distribution detaches from the middle. You can watch it happen in the postings before you see it in the survey: Naukri's JobSpeak index recorded the senior 20-plus-LPA band up about 16% over the year, a real-time signal a compiled survey can't match. The aggregate backdrop compounds it: India's Net Employment Outlook was +68% for Q2 2026, the strongest since 2008, yet 82% of employers reported difficulty finding the skills they need, per the ManpowerGroup Employment Outlook Survey. Strong intent meeting thin supply is precisely the condition under which the senior tail pulls away from the 75th.

When you finally engage a senior candidate worth having, someone else usually is too, and the incumbent employer definitely is. The counter that lands on the table is built from what that specific person is worth to a team that already knows them: the projects that stall if they leave, the relationships that walk with them, the cost of a 9-month re-hire. That number tends to sit well above wherever the survey put the 75th, and it arrives with relationship and momentum attached.

The pressure is structural, and the external evidence points the same way. Korn Ferry's global rewards work puts the most common AI-talent pay premium at 10–15% above peer roles, a premium that a whole-population 75th percentile, by definition, doesn't contain. At the very top the mix stretches further still: Deloitte's executive rewards survey puts CXO pay up 7–11%, with a large share of senior earnings performance-linked rather than fixed. A percentile expressed as a single base-salary figure is bringing a description of last year's average to a negotiation that's being fought on this year's exceptions. It loses because it isn't a decision, and a counter-offer is.

Benchmarks still matter here. They're the most defensible floor you have, and the convergence between the major houses is genuinely useful: Aon projects India salary increases of 9.1% for 2026 and Mercer projects around 9%, two independent gold-standard surveys landing in the same place. Use that consensus to bound the conversation. Then take an actual position on the person: what they'd cost, what they'd build or save, what walking away costs you when the mandate stays open another quarter. That's a decision a counter-offer has to beat.

Pricing on the person also means pricing the whole number, base, variable and equity together. A percentile is a base-salary figure; the offer that closes a senior candidate rarely is. At leadership level variable pay and equity carry a growing share of the package, and the re-pricing runs hotter in GCCs specifically: Zinnov puts average GCC increments near 9.9%, above the national number, while Mercer reports the wider market shifting toward skills-based, more transparent pay frameworks. So the position you take runs on two axes. Level: how far above the published band the live market has moved for this scarcity. And mix: how base, variable and equity combine to make the offer competitive at this seniority. A committee that adjusts the base but ignores the mix has corrected half the problem.

For a compensation committee, this resolves into one procedural change: treat the percentile as one input the decision is built from. The survey enters the room as evidence. The verdict is a priced position on the specific hire: the band as a floor, aged forward for the scarcity of this role, with the mix specified for the level, and a defined cost-of-loss for leaving the seat empty.

In practice that means three standing questions before a senior band is signed off. First: is this segment re-pricing, and if it's AI, data or a scarce IT skill, it almost certainly is, so the whole-population percentile is understating the tail? Second: what number will the incumbent's counter be built from, and is our position designed to beat that or merely to look competitive against a survey? Third: is the figure a base number or a total-package number, mix included? A committee that asks those three is pricing on the person. A committee that reads “the 75th” off a table is pricing on a policy, and, as the counters keep landing nearer the 90th, will lose the hire it most wanted to make, slowly, one great candidate at a time.