Performance systems. We start with the governance, not the form.

Nearly every company that calls us about performance calls in October and asks for a new form. We ask two questions instead. Who decided last year's ratings, finally? And when somebody disagreed, who broke the tie and on what stated basis? In about eight out of ten conversations there is a pause, and then a version of "it depends" or "it went to the founder". That is the engagement. The form is three afternoons.

What we work on

Rating governance, written down

Two paragraphs, agreed by the leadership team before anything else is touched. Who proposes a rating. Who moderates. Who has final say and in what forum. What evidence a manager must be able to produce. Whether a skip-level can overturn a direct manager, and if so what the direct manager is told. Whether the founder can overrule calibration — the honest answer is usually yes, and it is far better to write that down than to pretend a committee is sovereign when it is not.

Calibration that is a meeting, not a spreadsheet

We design the calibration session itself: who is in the room, how long each person is discussed, what a manager must bring, and what a chair does when the discussion turns into trading. We sit in on the first live cycle and chair or co-chair it if asked. On a bell curve: we are not doctrinaire about forced distribution, but we insist a company decide explicitly whether it has one, because a company that denies having a curve while quietly enforcing budget-driven ratios is teaching its managers that the process lies.

Competency framework rewrites

Almost always a subtraction. We test the existing framework by reading it aloud to the managers expected to apply it and asking for one real example per behaviour, from the last quarter, with a name attached. Nineteen behaviours routinely become four or five. Those five get written in the words your managers actually use — not "demonstrates ownership orientation" but the sentence a delivery head in Kharadi would say about the person who noticed the client escalation before the client did.

Cycle mechanics and calendar

Where goals live and who writes them. Whether you need mid-year at all — many companies at this size do not, and a badly run mid-year does more damage than none. Self-appraisal: keep it, but cap it. The dates: an appraisal cycle that overlaps January closure, the run-up to March year-end and a client escalation season will lose to all three. We publish the whole calendar, including the letter date, before the cycle opens.

The paperwork that survives a dispute

Appraisal comment boxes, PIP templates with defined periods and named skip-level reviewers, promotion notes, increment communication, and the difference between a rating and a compensation decision — which should be separated by at least three weeks and usually are not. We work with your existing HRMS rather than against it. Darwinbox, Keka, Zoho People, greytHR and SuccessFactors all constrain what a cycle can look like, and a design that ignores those constraints becomes a parallel set of spreadsheets by week two.

Career paths and bands, only if you have the problem

Levelling frameworks are seductive and are the single most common way a 120-person company spends four months producing a document that exists to be argued with. We build them when promotion decisions are genuinely arbitrary and people are leaving over it — not because a competitor in Magarpatta published theirs.

Design the form first and you will discover mid-calibration that nobody agreed who breaks a tie. The form is the easy part.

How the eight to twelve weeks run

• Weeks 1–2: read the last two cycles in full, including overturns, appeals and everything the HRMS did not capture. Interview eight to ten managers and everyone who chaired a moderation discussion.
• Weeks 3–4: governance drafted and argued out with the leadership team. This is where the engagement is won or lost, and it is usually two long, tense sessions.
• Weeks 5–8: framework rewrite, forms, calendar, templates, HRMS configuration notes for your people-ops lead.
• Weeks 8–12: manager briefings by cohort, a dry-run calibration on last year's real data, and the communication that goes to everybody — written by us, sent by your founder, never the reverse.
• We sit through the first live cycle end to end at no extra fee if it falls within four months of handover. A system nobody has watched run once is a hypothesis.

What we will not do

We do not sell or resell performance software, and we take no referral fee from any HRMS vendor. We do not benchmark your rating distribution against an industry dataset, because the datasets are self-reported and the comparison ends arguments that should be had. And we will not design a system whose real purpose is to make terminations easier to defend. If that is the brief, say so early and we will decline early.