The form you fill out is also the company you become
Somewhere around the third or fourth cycle, you notice that your colleagues have stopped sharing information freely. Not dramatically. Nobody slammed a door. It's more that the person two desks over used to flag problems early, and now they don't. The quarterly review is coming, and flagging problems doesn't score well on the rubric.
This is the part most guides on performance management skip entirely. The stated purpose of a review system is to measure performance. The actual effect is to define what performance means inside that organisation, and then to select, over time, for the humans who are best at optimising for that definition. Those two things are not always the same. Frequently they are in direct opposition, and the gap between them is where culture quietly rots.
Any measurement system, once people know they're being measured, changes the behaviour it was meant to observe. This is not a flaw in the humans. It is a near-physical law of organisations.
The mechanism: what gets measured gets gamed, what gets gamed gets normalised
Consider a mid-sized software company that introduces a review framework built around five competencies, each rated one to five. The designers are thoughtful people. They include collaboration, communication, and initiative alongside technical output, run calibration sessions, and feel good about it.
By cycle three, something specific has happened. Engineers have learned that "initiative" is scored by visibility: presenting at all-hands meetings, volunteering for cross-team projects, writing internal blog posts. The quieter engineer who caught a critical data pipeline fault at eleven on a Friday night, fixed it silently, and went to bed gets a three. The engineer who proposed a new onboarding wiki, got two colleagues to co-author it, and presented it at the quarterly review gets a four-point-five. The wiki is fine. The silent fix prevented six figures of lost revenue. Neither the system nor its designers intended this outcome. It happened anyway.
Now multiply that dynamic across two hundred engineers over four years. The people who stayed and got promoted are, on average, people who understood the visibility game. The people who found that game distasteful, or were simply bad at it, left or stagnated. The culture that results is not the culture anyone designed. It's the culture the measurement system selected for, the way a river selects for whatever shape lets water move fastest.
Not through malice. Through iteration.
Forced ranking and the colleague you used to trust
Forced ranking deserves its own reckoning. The logic is seductive: require managers to distribute ratings across a bell curve, prevent grade inflation, force honest differentiation. General Electric made this framework famous under Jack Welch, culling roughly the bottom ten percent of performers annually. The business press celebrated it for years.
The cultural consequence that was consistently underreported is what happens to peer relationships when colleagues are structurally in competition with each other for a fixed number of high ratings. If only fifteen percent of a team can receive the top grade regardless of absolute performance, then your teammate's success is, mathematically, your problem. Helping a colleague debug their work becomes, in a forced-ranking system, a mild act of self-sabotage. The designers of these systems wanted rigour. What they often got was a slow erosion of the cooperative behaviour that makes teams function.
Take two product managers, call them Priya and Marcus, who joined the same company in the same cohort. Same team, same manager, same review cycle. Priya is meticulous and shares her research generously. Marcus is equally talented and has learned that in a forced-ranking environment, differentiation matters more than collaboration. By year two, Marcus has cultivated a habit of solving problems visibly and crediting himself precisely. Priya has not. They are both good at their jobs. One of them has understood the actual incentive structure, and it isn't the one written in the employee handbook.
This is not a story about bad people. It is a story about rational adaptation to a bad system, and forced ranking is a bad system.
What people get wrong: confusing the map with the territory
The most persistent mistake in thinking about review systems is assuming that a well-designed rubric will produce the culture described in the rubric. It won't. The rubric describes the territory the designers wanted. The system, as actually experienced by employees, is the map people navigate by. And maps get annotated. Shortcuts get discovered. Certain routes get marked dangerous.
There is also a subtler error: assuming that the damage only happens when systems are badly designed. Even thoughtful, nuanced systems produce unintended cultural drift. Any formal evaluation creates a distinction between what is measured and what is not measured, and over time, in organisations under pressure, the unmeasured things atrophy. Mentoring a junior colleague rarely appears cleanly on a scorecard. Neither does the institutional knowledge held by the person who has been there twelve years and knows which vendor will actually pick up the phone. Neither does the team member who absorbs conflict quietly, keeping morale intact during a difficult quarter. These contributions are real. Review systems tend to make them invisible, and invisible things get deprioritised, and deprioritised things disappear.
Ask yourself: how many genuinely irreplaceable people has your organisation quietly selected out because they were excellent at things nobody thought to measure?
The folk remedy that needs to die is the belief that you can fix this by adding more categories to the rubric. Organisations have tried. The rubric grows to fourteen competencies, each with a three-paragraph description. Employees spend more time managing their review documentation than they do on any of the fourteen competencies. The problem isn't the number of boxes. It's that the act of boxing things up has consequences that live outside the box.
The slower, quieter damage
The deepest cultural effects of a performance review system take years to appear, which is part of why they go unnoticed until they're severe. Turnover in the first year looks like a recruiting problem. Declining cross-team collaboration in year two looks like a communication problem. The gradual promotion of a particular personality type into senior roles looks, for a long time, like nothing at all. It only becomes visible in retrospect, when someone asks why the leadership team has become so homogeneous, or why the organisation struggles to retain people who are excellent but not self-promotional, or why nobody raised the alarm on a problem that was obviously visible for months before it became a crisis.
Review systems are, in this sense, the tyres of the organisation rather than the engine. Nobody talks about them much when they're working. The catch: they are the thing in constant contact with the road, and they determine, more than most leaders appreciate, both the direction and the traction.
The honest answer is not that performance reviews should be abolished. It's that anyone designing or inheriting one should hold a genuinely uncomfortable question, not "does this system measure what we value?" but "what kind of person will this system, over five years, select for?" Those are different questions. The second one is the one that keeps the good people in the building.