Performance Management Has a Pendulum Problem
I’ve worked in and around performance management for nearly two decades. In that time, I’ve watched the conversation about “what good looks like” shift more than once.
About fifteen years ago, much of the energy was focused on making performance management more developmental and forward-looking, with less emphasis on formal reviews and ratings. The concepts I heard most were coaching, growth, trust, and better conversations between managers and employees.
More recently, I’ve observed the center of gravity shift. The discussions and directives I hear now are more often about accountability, differentiation, and results.
In fairness, not every company has followed this trajectory. But I do think performance management has a bit of a pendulum problem.
HR and business leaders become concerned that traditional performance processes are too rigid, bureaucratic, or damaging to trust—so they invest time, money, and organizational mindshare to change them: Remove ratings. De-emphasize formal reviews. Train managers on psychological safety.
Then, the concern pivots. Company performance isn’t strong enough. Managers aren’t addressing underperformance. Employees don’t know where they stand. We need more accountability. More differentiation. Now the solutions become: Increase the frequency and candor of performance conversations. Train managers on delivering difficult performance feedback. Reimplement ratings to drive accountability and add tension back into the system.
These pivots often consume an enormous amount of time for employees, managers, leaders, and HR. Sometimes that investment is exactly what the organization needs. But often, organizations move too quickly from What isn’t working? to What should we change?
Should we have ratings or not?
Should we have three levels or five?
Should reviews be annual, twice a year, quarterly, or abandoned entirely?
Should compensation be tied directly to ratings?
Should we calibrate?
These are important questions. But they’re design questions—not diagnostic questions.
Let’s start with everyone’s favorite debate topic, ratings.
Removing ratings from performance processes can absolutely be the right decision. If managers are spending hours debating distinctions that don’t meaningfully affect outcomes or employees are fixated on the label over the feedback, the rating may not be earning its cost.
In a different organization where performance expectations aren’t clear, managers are reluctant to tell employees when they’re falling short, and compensation decisions are made using inconsistent criteria, taking ratings away could weaken the process even more. Not because ratings solve those problems, but because removing one of the few forcing mechanisms for explicit performance judgments may make those judgments even easier to avoid.
I’ve also seen companies that abandoned ratings later conclude that they’d lost too much accountability. Bringing ratings back seemed like an obvious way to reintroduce tension into the system.
The same mechanism can solve a problem in one organization and make it worse in another.
In the throes of this very debate a few years ago, a team member put it this way: “Ratings or no ratings—that’s not the question.”
She was right. The question is what purpose the ratings need to serve, and whether the conditions around them make that possible.
This is where the pendulum can get us into trouble.
When the prevailing concern is trust, we can start treating anything that creates tension—ratings, differentiation, calibration—as inherently suspect.
When the prevailing concern is accountability, we can make the opposite mistake: assuming that more structure and rigor will create better performance.
Neither is necessarily true. An organization can have a highly growth-oriented performance philosophy and still avoid hard conversations. It can have rigorous ratings and differentiation expectations and still make poor performance judgments. It can simplify the process dramatically and improve it, or simplify it so much that it no longer serves its purpose.
Instead of searching for the perfect balance between “conversational” and “robust” or “soft” and “hard” performance management, a better starting point is understanding what the organization actually needs the system to do.
Before redesigning performance management, companies should ask questions like:
Are performance expectations clear today?
Are managers having useful conversations with their employees throughout the year?
Do managers make sound judgments based on evidence, or do expectations vary widely depending on who you work for?
Does the organization meaningfully distinguish different levels of contribution?
Do performance decisions connect coherently to things like compensation and promotion?
Does the system make sense for this organization—its current strategy, culture, and context?
These questions may be less satisfying to answer than “Should we have ratings?” because there isn’t an obvious, universally “right” answer.
But that’s the point. There may not be any universally right answers when it comes to performance management.
Better performance management starts with a better diagnosis.
