Most Indian MSMEs that run an annual appraisal are not getting much from it. The forms come back late, the ratings cluster in a narrow band, a few people are upset for a fortnight, and the business carries on exactly as before. The exercise consumes real management time and changes almost nothing.
The reasons are consistent enough to be predictable. Below are the five that account for most of it, and what actually fixes each one.
1. Nothing was defined before the year began
This is the root cause, and the other four largely follow from it. If a role was never given clear result areas at the start of the period, the appraisal form asks a manager to rate performance against a standard that was never agreed.
The conversation then becomes personal, because there is nothing external to point at. Both parties are arguing about character rather than about work, which is why these meetings are dreaded on both sides of the table.
The fix: define result areas and their measures before the period starts, role by role, and have the person acknowledge them. This is unglamorous work and it is the whole game. Everything downstream depends on it.
2. Recency decides the rating
Where there is no review rhythm through the year, managers rate from memory, and memory is dominated by the last six to eight weeks. A strong performer who had a difficult final quarter gets marked down. Someone who was invisible for ten months and busy in the eleventh gets marked up.
People notice this quickly, and the rational response is to look busy near appraisal time rather than to perform consistently.
The fix: a quarterly review that is documented, however briefly. It does not need to be elaborate — a short recorded conversation against the agreed measures, four times a year, removes almost all recency distortion because the annual rating becomes a summary of four data points rather than one impression.
3. Individual goals are not connected to business goals
In a surprising number of organisations, everyone scores well and the company still misses its numbers. That is a cascade failure: departmental targets were never translated into individual accountability, so it is entirely possible to succeed on every line of your form while the business underperforms.
The fix: a deliberate cascade. Business goals for the year, stated in numbers leadership genuinely tracks. Then what each function must deliver for those numbers to be met. Then role-level result areas drawn from that, limited to what the role controls. If you cannot trace a line from a person’s result area back to a business goal, one of the two is wrong.
4. Managers were never taught to hold the conversation
Most managers in growing firms were promoted for technical competence. Nobody taught them how to give specific developmental feedback, and the natural response to discomfort is avoidance — which shows up as uniformly good ratings.
This is the step most implementations skip, and it is the main reason systems that launch well decay within two cycles. A well-designed format handed to untrained reviewers produces inflated, uninformative data.
The fix: treat reviewer capability as part of the system, not an optional extra. Managers need a documented conversation structure, practice, and calibration with their peers so that a four in one department means roughly what a four means in another.
5. Nothing follows from the rating
If the outcome changes nothing — no development plan, no role change, no reward differentiation, no consequence for sustained underperformance — people correctly conclude the exercise is ceremonial and invest accordingly.
The fix: connect outcomes to something real. That does not have to mean money in year one, and probably should not. Development planning, stretch assignments, and honest role conversations are all consequences. But something must visibly follow, or the credibility never arrives.
The sequence matters more than the format
Organisations often approach this by looking for a better form, or by buying software. Neither addresses the causes above. A template applied to undefined roles fails in exactly the same way as the template it replaced, and software laid over an undefined process simply automates the confusion at greater expense.
The order that works is: define result areas, choose measures that can be collected honestly, build a review rhythm, train the reviewers, run one cycle without money attached, correct what the data exposes, and only then connect outcomes to reward. Software, if you want it, comes after the design has survived a real cycle.
Further reading
For the mechanics of defining result areas and choosing indicators, see KRA and KPI design for growing Indian businesses, and how to set KRAs for factory roles if you run a manufacturing operation. For how a full implementation is scoped and run, see performance management system consulting.
By Dr. Babu Balakrishnan — management consultant and HR practitioner. Full profile.