Compensation Benchmarking: How to Do It Right (and How Most Companies Get It Wrong)

Every compensation decision rests on an assumption about the market. Benchmarking is simply the discipline of making that assumption explicit and testing it against data. Done well, it gives boards confidence and HR teams authority. Done badly — which is common — it launders guesswork into false precision.

The four failure modes

Before the method, the mistakes. Nearly every flawed benchmarking exercise we review fails in one of four ways:

• Title matching instead of job matching. A "Head of Marketing" in a 50-person D2C startup and one in a multinational FMCG are different jobs with different market values. Surveys match on job content, scope and organisational size — not business cards.
• Survey shopping. Running the data through multiple sources and quoting whichever supports the desired answer. The tell: a benchmarking deck that cites different surveys for different roles without explaining why.
• Percentile drift. Declaring a P50 pay philosophy, then approving offers at P75 "for this exceptional candidate" — every time. Within two years the actual position is P70 while the stated philosophy remains P50, and nobody can explain the gap to the board.
• Stale or mismatched cuts. Using all-industry national data for a Bengaluru-based product company, or two-year-old data in a hot talent market. The cut matters as much as the source.

A rigorous benchmarking process

• Define the talent market per job family. Ask where you actually lose people to, and hire from. Engineering may compete nationally with tech companies; the plant workforce competes within a 30-kilometre radius. One company, several markets.
• Select surveys deliberately. Reputable Indian market surveys differ in their participant base — some skew to multinationals, others to domestic companies or startups. Choose the survey whose participants look like your competitors for talent, and stay consistent year over year so trends are real.
• Match jobs honestly. Use job descriptions, not titles. Where your role straddles two survey benchmarks, document the blend. An honest 80% match noted as such beats a confident mismatch.
• Decide your positioning — by element. Companies often target the median on fixed pay but design total compensation to reach upper quartile through variable pay and LTI for strong performers. Stating positioning element by element ("P50 fixed, P60 total cash, P75 total compensation for top performers") is far more useful than one blunt percentile.
• Triangulate, then judge. Survey data is the backbone; offer and decline data, interview feedback and attrition exit data are the reality check. When they conflict, investigate rather than averaging.

Benchmarking in the Indian context

Two India-specific considerations deserve attention. First, wage code readiness: as the labour codes redefine "wages," comparisons of fixed-pay structures across companies will get noisier for a few cycles, since organisations will restructure allowances at different speeds. Benchmark total fixed cost, not just basic salary. Second, the formal–informal premium: in many sectors, survey participants are disproportionately organised-sector employers, so data may overstate the market for roles you actually fill from less formal competitors.

From data to decisions

Benchmarking is diagnosis, not prescription. The output should feed three decisions: where to set or move pay ranges, which roles or individuals are critically off-market and need correction, and what the total cost of correction is over one to three cycles. A benchmarking report that ends with data tables but no funded action plan changes nothing.

For executive roles, benchmarking deserves extra care: samples are thinner, packages more bespoke, and equity harder to compare. This is one reason boards pair rewards advice with hiring decisions — and why we built our executive hiring cost calculator to make total offer costs explicit. Companies without senior rewards capability in-house often run their first rigorous cycle with fractional CHRO support, building the muscle internally thereafter.

If your last benchmarking exercise produced numbers nobody fully trusted, we should talk.