Appraisal forms in a shared drive: what performance software changes and what it cannot

An HR manager was asked, in June, what a particular employee's rating had been last year and what objectives had been set. She found the folder, then the sub-folder for the department, then a document named with the employee's initials, and opened a form completed in March of the previous year. It had a rating and three objectives. Nobody had looked at it since, the employee's manager had changed, and the new manager had never seen it. The appraisal had happened. Nothing about it had persisted.

That is the state performance software is bought to fix, and it does fix it, provided the business has already decided what its performance process is. Where it has not, the software becomes a more expensive shared drive.

The value of a performance system is not the form. It is that for each person there is one place holding their objectives, the manager's notes through the year, the mid-year check, the rating, the calibration outcome and the development plan — and that all of it survives a change of manager, a move of team or a year passing. A new manager opening that record in June sees what was agreed and what was observed, and the first conversation starts from evidence rather than from nothing.

Continuity also protects the employee. A rating given by a manager who left in April, with no notes behind it, cannot be explained to the person it was given to, and cannot be defended if it is challenged. The record is what makes an appraisal fair as well as useful.

JAN   objectives set, agreed, visible to both
FEB-  manager notes: dated observations, good and bad,
NOV   two or three a quarter
JUL   mid-year check: on track / at risk, per objective
DEC   self-assessment, manager rating with evidence
JAN   calibration outcome, final rating, next objectives

The notes column is the one that changes the
conversation. It is also the one nobody keeps in a
shared drive.

The dated notes are the feature that pays for the system. A manager who has recorded three observations a quarter arrives at the year-end with evidence; one relying on recall is weighted to the last six weeks and to whoever was most visible. Software does not make managers keep notes, but it gives them one place to put them and a reminder to do so, which is more than the drive ever did.

Two things a business cannot see from a folder of forms become obvious in a system: the distribution of ratings, and the distribution by manager. When thirty-four of forty-one people are rated above expectations, that is visible on one screen rather than discovered by someone counting. When one manager's team is rated a full level higher than every other, that is a calibration finding rather than a suspicion. A first cycle run properly depends on being able to see this before ratings are final, which is why the performance management process puts calibration ahead of the review conversations.

Visibility is not the same as action. The system shows the inflation; managers in a room, with the distributions in front of them, correct it. A business that buys the software and skips the calibration meeting has bought a very clear picture of a problem it is not fixing.

1. 1

   What the cycle is for

   Development, differentiation, or pay. The system needs to know whether ratings feed increments, because that changes who can see what and when. This is the first question configuration asks, and most businesses have not answered it.

2. 2

   How many levels, described how

   Three to five, each in words. A system configured with 'meets expectations' as the only description of level three will produce the same argument in software that it produced on paper.

3. 3

   Who sees whose rating

   Managers see their team; skip-levels see the distribution; HR sees everything; the employee sees their own after calibration, not before. Get this wrong and provisional ratings leak, which is worse than the drive.

4. 4

   What the objectives connect to

   Team KPIs, or nothing. The system can cascade objectives from a framework if one exists; where it does not, it cannot, and configuring it as though one does produces objectives nobody set.

Each of these is a performance management process decision. The software will ask for them in the first week of implementation, and a business that has not made them will either delay the project or let the implementer guess. Guessing is how the first cycle produces a rating distribution nobody can defend.

Calibration is where the shared-drive model fails completely: forty-one forms cannot be compared on a screen, so the meeting either does not happen or happens with a summary spreadsheet somebody built by hand the night before. A performance system holds provisional ratings in a state that managers can compare — by team, by level, against the distribution — and lets the calibrated outcome be recorded separately from the manager's first rating. That separation matters. It keeps an honest record of what the manager thought and what the room decided, which is what makes the following year's calibration a conversation about drift rather than an argument about memory.

Configure that state before the first cycle: provisional, calibrated, final. Systems set up with a single rating field skip straight from the manager's number to the employee's screen, and the calibration meeting becomes a discussion of ratings people have already seen. Once that has happened there is no walking it back, and the performance management system has recreated the drive's worst property in software.

Managers experience performance software as admin unless it saves them something. What it saves them is the year-end scramble: no reconstruction of a year from memory, no chasing the previous form, no arguing about what was agreed in January. A manager shown that in the first cycle uses the notes feature in the second. One who experiences it as a form with more fields does not, and the system's record for that team stays as empty as the drive's.

The other thing managers need is the difficult-rating conversation, which no system provides. That remains a skill, and where several managers are having their first such conversations in the same fortnight it is the case for leadership and people management development timed before the cycle rather than after it.

• After the employee record is clean. Objectives attach to people in roles with managers. Where reporting lines are wrong, objectives go to the wrong person and the whole cycle is undermined.
• After the first paper cycle, not instead of it. A cycle run once on paper tells you what to configure. A cycle run first in software tells the implementer what to guess.
• Module over standalone, for most growing businesses. Ratings connect to pay, and pay lives in the HRMS; a separate tool re-creates the handover the drive already had.
• Notes from day one. Configure the observation notes and the quarterly reminder before anything else. It is the feature that changes the year-end, and the one most implementations leave off.

Sequenced that way, performance management software does the one thing the folder could not: it makes an appraisal something that persists, accumulates and can be seen across the business, rather than a document in a sub-folder that the next manager never opens.