The bell curve — forced distribution, forced ranking, whatever your organisation calls it — keeps coming up in Indian HR discussions, usually in the same form: our appraisal ratings are all clustered at the top, should we force a distribution to fix it?
It is a fair question, and the honest answer depends almost entirely on the size of your organisation and on what is actually causing the clustering.
What the bell curve was designed to solve
Forced distribution exists to counter rating inflation. Managers avoid difficult conversations, so ratings drift upward until almost everyone is rated good or better and the data becomes useless for any decision. Requiring a fixed share in each band forces differentiation to reappear.
It does address that problem. The question is what it costs you to solve it that way, and whether the problem it solves is the one you actually have.
Why it fits large organisations and not small ones
Forced distribution rests on a statistical assumption: that in a large enough population, performance approximates a normal distribution. That assumption is defensible across ten thousand people. It is not defensible across a department of nine.
In a hundred-person company, a department might have six people, and three of them might genuinely be excellent — small teams are often deliberately selected to be strong. Forcing one of those three into a lower band to satisfy a distribution is not correcting a bias. It is introducing a falsehood, and everyone in the room knows it.
The damage in a small firm is also more concentrated. In a large organisation an unfair rating is absorbed. In a hundred-person company, the person is visible, their colleagues know the rating was arbitrary, and the credibility of the whole system goes with it. Small organisations run on trust and proximity, which are precisely what forced distribution spends.
What is usually causing the clustering
Before reaching for a distribution, it is worth asking why all the ratings are high. In growing Indian companies the cause is usually one of three things, and none of them is fixed by forcing a curve.
There was no standard to rate against. If result areas were never defined at the start of the period, a manager has nothing objective to point at, so the safe rating is a good one. Forcing a distribution here means managers must now allocate low ratings arbitrarily, because they still have no standard. You have replaced uninformative generosity with uninformative unfairness.
Managers were never taught to give developmental feedback. Clustering is often avoidance rather than genuine assessment. The fix is reviewer capability, not a quota.
Nothing follows from the rating anyway. Where the outcome changes nothing, managers reasonably treat the form as a formality and rate kindly.
What to do instead in a company of that size
Define result areas and measures before the period starts. Most clustering disappears on its own once there is something concrete to assess against. This is the fix that actually addresses the cause.
Calibrate rather than force. Get managers in a room to discuss ratings against agreed evidence before they are finalised. Calibration produces differentiation through argument and evidence; forced distribution produces it through arithmetic. The first is defensible to the person being rated. The second is not.
Use guidance, not quotas. Telling managers that a distribution heavily weighted to the top band usually signals a definition problem, and asking them to justify it, achieves most of what a quota achieves without the arbitrariness.
Differentiate where it matters. Most organisations do not need fine-grained distribution across the whole population. They need to identify a small group performing exceptionally and a small group genuinely struggling. That is achievable without imposing a curve on everyone in between.
If you still want a distribution
Some organisations have reasons to keep one — group policy, or a parent company standard. If so, two conditions make it survivable: apply it at the whole-organisation level rather than within small departments, where the statistics are meaningless; and treat it as a guideline that a manager can depart from with a documented reason, rather than an arithmetic rule.
What consistently fails is a rigid curve imposed on teams of five or six with no route to challenge it.
Further reading
The underlying causes of clustering are covered in why performance appraisals fail in Indian MSMEs. For building the standards that make forced distribution unnecessary, see KRA and KPI design for growing Indian businesses.
By Dr. Babu Balakrishnan — management consultant and HR practitioner, adjunct faculty in management. Full profile.