Pay & compensation transparency.

Most compensation conversations in India happen in private. We publish band data because the alternative is the loudest voices winning.

Compensation intelligence drawn from live placements across 64 GCCs, not from benchmark surveys. What roles actually pay. Where bands are moving. Why the published band and the closed offer are different numbers — and which one to trust.

21 articles Updated 21 May 2026 Written by Sachith Rai, Christabel, Recruise editorial

Key takeaways

The published compensation band and the closed-offer number are different figures — in senior AI/ML hiring across Bengaluru GCCs this quarter, they differed by an average of 32%.
Bands don’t shift uniformly. They split. A handful of outlier closes signal a new top tier before the median moves; comp committees that miss the split over-pay for the bottom and lose the top.
The 75th percentile is a defensive number, not a strategic one. In senior segments, it loses to the 90th-percentile counter-offer every time.
Pay transparency laws are now live in the EU, parts of the US, Australia, and the UK. India has no statutory pay transparency law yet, but DPDP Act compliance is reshaping how compensation data can be handled internally.
The metrics that tell you whether your compensation decision was right are offer-accept rate, counter-offer rate, and 12-month retention — not the band you published.

Senior AI/ML — Bengaluru — Q1 2026 Source: Recruise live placements

Published band (top)

₹85L

Offer made

₹98L

Offer accepted

₹1.12 Cr

Across 14 senior AI/ML placements at Bengaluru GCCs in Q1 2026, the gap between published band and closed offer averaged 32%. The number that mattered to the candidate was the third one, not the first.

The published band isn’t the real band.

Most published compensation bands are written by people not running active mandates. They’re aggregated from prior-year hires, benchmark surveys, and what HR thinks is competitive.

The actual offer-accept number is something different. It’s shaped by the specific candidate’s current package, their counter-offer, their alternative options, and whether the hiring manager values them enough to break the band quietly.

In senior AI/ML hiring across Bengaluru GCCs this quarter, we’ve closed 14 placements where the accepted offer exceeded the published top of band by an average of 32%. The published band hasn’t moved. The closed offers have.

If your comp committee is calibrating to the published band, you’re calibrating to a number that doesn’t reflect what your competitors are actually paying.

Read the AI/ML salary band analysis

The pattern

A band doesn’t shift first. It splits.

Recruise placement data · Q1 2026

Bands split before they shift.

Watch a senior compensation band over six quarters and you’ll see a predictable pattern. The band doesn’t move. Then a few outliers close at numbers that look like errors. Then more outliers. Then the band re-anchors at a new median — usually 15–20% higher than it was.

The outliers aren’t errors. They’re the leading edge.

In senior AI/ML right now, the band has split. The bottom of the band — people doing traditional ML work, deep learning research, applied ML in production — sits roughly where it was a year ago. The top of the band — people who can stand up GenAI products end-to-end, who understand both the modelling and the productisation — has moved meaningfully.

The mistake most comp committees make is treating this as one band that needs to shift up. It’s two bands now. If you price as one, you over-pay for the bottom and under-pay for the top.

Above: Section visual placeholder. Replace with editorial image, data viz, or typographic statement at handoff.

What the 75th percentile is actually buying you.

“We pay at the 75th percentile” is the most defensive compensation statement in Indian recruitment. It sounds rigorous. It is, in practice, a way to avoid taking a position.

The 75th percentile of what? Last year’s closed offers? Survey data with a 6-month lag? Aggregated benchmarks that don’t distinguish between Bengaluru and Tier 2? The number depends entirely on the basket. Most comp committees never inspect the basket.

What the 75th percentile is actually buying you, in the senior segment where you’re competing with Google, Microsoft, and the well-funded scale-ups: a credible offer that loses to a 90th-percentile counter every time. You’ve done the work. Made the offer. Run the process. And then watched the candidate take a number that wasn’t even in your model.

The senior market doesn’t price at the 75th percentile. It prices at what the best alternative offer is. That’s the only number that matters, and it’s a number you have to source from live mandate data — not from a benchmark deck.

Read: the hidden cost of paying at the 75th percentile

The third lever

Base, variable, ESOP. The third one is moving the deal more than people admit.

Senior placement data · Last 6 quarters

ESOPs are the quiet third lever.

Senior GCC roles increasingly come with meaningful ESOP allocations from the parent company. We’re seeing this most clearly at the VP level and above, where the parent ESOP can represent 30–60% of the package’s long-run economic value if the parent is publicly traded and performing.

The structure of these grants varies wildly. Cliff schedules, vesting terms, performance gates, exercise windows, exchange rates at vest, tax handling in India versus at parent — each of these is a lever that can make two superficially similar offers worth very different things.

The candidates who win these negotiations understand the structure. The candidates who lose them anchor to the headline number and let everything else slide.

For comp committees: if you’re competing with companies that have meaningful parent ESOPs, your cash package alone won’t close it. You need either an equivalent equity story or a cash compensation that genuinely compensates for not having one. Splitting the difference loses the candidate.

The numbers that matter are downstream of comp.

The compensation decision is upstream. The metrics that tell you whether the decision was right are downstream. Most comp committees track the upstream metric — the offer made — and miss the downstream ones entirely.

Three downstream numbers worth tracking, none of which are difficult to capture:

“The offer-accept rate is the single most undervalued metric in Indian recruitment. It tells you, in one number, whether your compensation is actually competitive in the segment you’re hiring in.”

Sachith Rai · MD & Founder, Recruise

Offer-accept rate. What percentage of offers made were accepted. Below 70% in senior segments suggests your comp is below market for the segment. Above 90% suggests you may be over-paying or your process is over-filtering before the offer stage.

Counter-offer rate. What percentage of offers triggered a counter from the candidate’s current employer. Rising counter-offer rates signal that the candidates you’re reaching are well-priced where they are — and you’re likely paying the same band.

12-month retention of placed candidates. What percentage of accepted offers were still in role 12 months later. Under-pricing buys you a candidate who’s already shopping for the next role at 9 months.

Comp is the visible decision. These three numbers are the invisible scoreboard. If you’re only watching the first one, you’re only watching a third of the picture.

How we know what we know

The compensation intelligence on this page is drawn from live placement data, not surveys.

4,000+

Senior placements analysed across the last 18 months. Every number on this page is back-tested against closed offers, not estimates.

64 GCCs

Active mandate relationships covering IT, BFSI, Pharma, and ER&D. Compensation data spans Fortune 500 captives through scaling mid-market GCCs.

94%

12-month retention rate on placed candidates. Compensation reads only matter if they hold — this is how we know whether ours do.

On this page

01 / Definition

What pay transparency actually is.

Pay transparency is the practice of making compensation decisions explainable — to candidates, to current employees, and to regulators. Practised seriously, it is a discipline. Performed for compliance, it is a liability.

There are three distinct levels of pay transparency, often conflated. They have different costs and different consequences.

Range transparency

Publishing the salary band attached to a role — in the job advert, in the application portal, in an offer letter. This is the level most pay transparency laws now require. It is the easiest to implement and the most commonly faked. Many companies publish wide ranges that technically comply but tell the candidate nothing useful.

Process transparency

Explaining how compensation decisions get made. The factors that move the number. The role of performance, market data, internal equity, and negotiation. This is harder to implement because it requires the compensation philosophy to be coherent in the first place. Many companies discover, when trying to document their process, that they don’t have one.

Individual transparency

Making individual pay levels visible inside the organisation — either by publishing them outright or by making them readily inferable. This is the rare end of the spectrum, practised most often by smaller organisations with strong cultural cohesion. It is the most demanding and the most informative.

Most of what gets called “pay transparency” in corporate practice is range transparency with some gestures toward process. The serious work is at the process layer.

02 / Regulation

The pay transparency laws — by region.

Pay transparency moved from voluntary practice to statutory requirement in many markets between 2021 and 2026. The specifics matter — what must be published, when, and to whom — and they vary widely.

This is a working summary, not legal advice. Compensation laws change; always verify against current statute before acting.

European Union In force from June 2026

The EU Pay Transparency Directive (2023/970) requires employers in all 27 member states to disclose pay ranges in job adverts or before the first interview, prohibit asking candidates about salary history, and report gender pay gap data above defined workforce thresholds. Pay gap reporting begins for companies with 250+ employees from June 2027. Companies must take corrective action if the unexplained pay gap exceeds 5%.

Source: EUR-Lex 32023L0970 · National transposition deadlines vary

United States — state by state Live in 10+ states

No federal pay transparency statute. Pay range disclosure in job adverts is now required in California, Colorado, Connecticut, Hawaii, Illinois, Maryland, Minnesota, Nevada, New York, Washington, and the District of Columbia, with similar laws active or pending in further states. Requirements vary — California and Washington require the range, Colorado requires range plus benefits, New York City requires “good faith” ranges. Salary history bans are now active in roughly half of US states.

Source: state department of labour filings · consult employment counsel by state

United Kingdom Gender pay gap reporting in force

The UK requires gender pay gap reporting for employers with 250+ employees and is consulting on broader pay transparency rules likely to take effect 2026–27. No general pay range disclosure requirement yet, but voluntary disclosure has become near-standard in financial services and tech sectors. Salary history questions are increasingly discouraged in guidance, though not banned outright.

Source: gov.uk — Gender pay gap reporting

Australia In force since December 2022

Australia’s Workplace Gender Equality Amendment Act 2022 prohibits pay secrecy clauses in employment contracts and gives employees the right to discuss their pay. Employers with 100+ employees must report gender pay gap data to the Workplace Gender Equality Agency, with company-level data now published publicly since February 2024. No range-in-advert requirement at federal level.

Source: Workplace Gender Equality Agency

India No statutory requirement

India has no statutory pay transparency law. Range disclosure in job adverts remains voluntary. However, the Digital Personal Data Protection Act, 2023 reshapes how compensation data can be collected, processed, and stored internally — particularly within GCCs that handle employee data on behalf of overseas parents. The Code on Wages, 2019 includes equal pay provisions but doesn’t mandate transparency.

Source: Ministry of Electronics & IT · Code on Wages, 2019

Singapore & rest of APAC Mostly voluntary

Singapore, Hong Kong, and most of South-East Asia have no statutory pay transparency requirement. Japan’s 2022 amendment to the Act on the Promotion of Female Participation requires employers with 301+ employees to disclose gender pay gap data. South Korea is consulting on similar measures. Voluntary range disclosure is rising in tech and finance hiring across the region but remains the exception.

Source: regional employment law filings

03 / Trade-offs

The risks and benefits, unvarnished.

Pay transparency’s benefits are real and well-documented. Its costs are real, less well-documented, and the reason most organisations move slowly. Both deserve naming.

What it actually buys you

Faster hiring at the senior end. Candidates self-select; the ones who’d have wasted six weeks of process drop out at the advert.
Lower offer-stage failure rate. When the range is known up front, the awkward “your offer is below my expectation” conversation happens earlier and resolves cleaner.
Sharper internal equity. Publishing ranges forces the organisation to defend why two adjacent roles are priced differently. Many discover they can’t.
Genuine retention signal. Employees who can see the band understand their position in it. The ones who’d have left over a perceived inequity often stay; the ones who’d have stayed under a misunderstanding sometimes leave — usually for good reason.
Regulatory protection. In markets where transparency is now law, voluntary early adoption establishes a defensible track record before the deadline arrives.

What it actually costs you

Existing internal inequities become visible. The two people doing the same job at very different pay points will have a conversation. You should be ready for it before the policy ships.
Competitor poaching gets easier. Your published bands tell competitors exactly where to position counter-offers. Most don’t bother. Some do.
Negotiation leverage shifts to candidates. When the top of the band is known, more candidates ask for the top of the band. Some get it.
The discomfort of having to defend the compensation philosophy. Range transparency without process transparency is incoherent. The work is at the process layer, and it’s harder than publishing a number.
The lawful basis question, in privacy-regulated markets. Publishing or sharing detailed compensation data interacts with the DPDP Act (India), GDPR (EU), and equivalent statutes. The legal review is non-trivial.

The honest read: transparency’s benefits accrue over years; its costs hit in the first quarter. Organisations that move first absorb the cost and earn the benefit. Organisations that delay pay the same cost later, under regulatory pressure, with less control over the rollout.

04 / Practice

The five moves that actually matter.

Most published “best practice” guides for pay transparency list ten things, eight of which are obvious. Five moves do the real work. The other five are noise.

Audit existing internal inequities before you publish anything.

The first thing transparency reveals is the gap between what your organisation says about compensation and what it actually does. Find those gaps in private, name them honestly, and resolve them — before any external policy ships. Publishing a band without doing this audit is a faster route to litigation than to trust.

Decide the level of transparency you can actually defend.

Range transparency is a different commitment from process transparency, which is a different commitment from individual transparency. Pick the level you can defend coherently. Going further than you can defend is worse than not starting; going short of what law requires is worse than going further than you needed to.

Equip your managers to talk about pay before you publish the ranges.

Transparency policies fail at the manager-employee conversation. The day the bands publish, every manager will be asked to explain them. Train the conversation. Script the difficult cases. “Why am I in the bottom half of the range?” should not be the first time the manager has thought about the answer.

Use the live-mandate data, not the survey data.

Benchmark surveys have a six-month lag and a basket bias. The reliable input for setting bands is what your competitors are actually closing offers at right now, which sits in live mandate data — from recruiters running senior searches in your segment. If you set bands from surveys alone, you will be six months behind by the time you publish.

Track the downstream metrics, not the band.

The band is the upstream decision. Offer-accept rate, counter-offer rate, and 12-month retention are the downstream metrics that tell you whether the decision was right. Most pay transparency programmes get measured by whether the policy shipped. The ones that work get measured by whether these three numbers moved.

Get the latest senior compensation benchmark.

The Q2 2026 compensation intelligence drop covers senior bands across IT, BFSI, Pharma, and ER&D GCCs. Drawn from live placement data, not surveys. Updated quarterly.

Get the benchmark