“We hire on merit, so we do not need a diversity programme.” You will hear this in some form in many Indian offices, usually said sincerely. It is also the belief that deserves the most careful look, because it assumes that merit arrives in the hiring process already measured, and that nothing in the process distorts it. Let us take that belief and three of its relatives apart, one at a time.

Belief one: diversity hiring means lowering the bar

It can, if it is done badly. A company that says “we need three women engineers by March” and then waives its technical screen for them has indeed lowered the bar, and everyone, including the three women, will sense it. But that is a description of a poorly run programme, not of the idea.

The better version holds the bar still and widens the pool that reaches it. Suppose an illustrative software firm in Mysuru receives 600 applications for 10 engineering roles, and 9% come from women. If the final ten hires include one woman, the hit rate for women (1 of 54) is about 1.9% against about 1.6% for men (9 of 546). Nothing was wrong at the selection stage. The shortfall is upstream, in who applied. So the useful work is to change where you advertise, which colleges you visit, what the job advert says, not to relax a test.

Here is the arithmetic, checked: 9% of 600 is 54 women and 546 men; 1/54 is 1.85%, 9/546 is 1.65%. Whether those few percentage points differ meaningfully with such small numbers is another matter, and a good analyst would say that ten hires prove almost nothing either way. Which brings me to the point: measure funnel stage by stage before deciding what the problem is.

Belief two: it is only a pipeline problem

Sometimes yes, as in the Mysuru case. But the opposite error is just as common: believing the pipeline explains everything and so doing nothing inside the process. Pipeline and process interact. Job descriptions padded with ten “must-have” requirements put off some candidates more than others, since people differ in how readily they apply when they meet only part of the list. Interview panels made up entirely of one kind of person judge “culture fit” in their own image. Unstructured interviews let personal comfort pose as judgement.

Look at each stage separately: applications, shortlist, interview, offer, acceptance, joining. If women or candidates from smaller towns drop out disproportionately at one specific stage, that is where to dig. It might be the shortlist (screeners over-weighting brand-name institutions), or acceptance (the offer came late, or the manager call was awkward).

Belief three: blind CVs solve it

Removing names, photos and sometimes colleges from CVs at the screening stage is a reasonable step and some firms use it. But it is a partial fix. It does nothing for bias at the interview, where the candidate is in the room. Details such as a gap in employment, a hometown or the name of a school can still tell a story. And if the screen’s underlying criteria are skewed (only graduates of a few institutions), anonymity just hides who is being filtered out.

More dependable improvements are less glamorous: structured interviews in which every candidate receives the same questions scored against the same anchors; written, job-related criteria agreed before looking at candidates; panels with more than one perspective; and scoring each interviewer’s ratings independently before they discuss, so the loudest voice does not set the tone.

Belief four: targets and quotas are the same thing

They are not, and conflating them causes confusion in both directions. A quota is a fixed requirement, typically set by law, that a certain number or share of positions go to a group. In India, reservation in government jobs and public institutions is a constitutional and statutory matter. Private employers in general are not bound by such a quota in hiring, though specific obligations exist in some areas (for example, the Companies Act requires certain classes of company to have at least one woman director, and disability law places duties on establishments).

A target, or aspiration, is a goal the company sets for itself, such as increasing the share of women in tech roles from 18% to 25% over three years. A target tells you what to monitor and where to look; it does not instruct a recruiter to select anyone. Companies that handle this well pair the target with process changes and report progress openly, including when it is missed. Companies that handle it badly turn the target into a monthly pressure on hiring managers, and hiring managers then game it.

There is a legal edge here worth knowing. Indian wage law bars discrimination on grounds of gender in recruitment and conditions of service for the same work, so any practice that screens out candidates on a protected ground, even with good intentions, can create exposure. Taking advice from counsel before introducing group-based preferences is sensible.

What tends to move the numbers

  • Instrument the funnel by group at each stage, with numbers large enough to mean something.
  • Rewrite job adverts: fewer inflated requirements, clear about flexibility and pay range where possible.
  • Widen sourcing beyond the usual campuses and referral circles, since referrals tend to reproduce the network that makes them.
  • Structure the interview and calibrate the interviewers.
  • Look at what happens after joining. A company that hires diversely but loses those people in eighteen months has fixed the wrong end of the problem.

I am not claiming that any one of these works everywhere. Small firms have small samples and few levers; large ones have scale but also inertia. What I would defend is the habit of looking at the evidence in your own funnel before accepting either “we are fine” or “we need a quota”.

Figures and names in the examples are illustrative.

Leave a Reply

Trending