Margie Zyble, CHRO at UC Health Cincinnati, heard one word in the scenario louder than the rest: refuses. She joined host Rebecca Taylor on HR Voices to work through a case the show calls the forced ranking fallout. A company runs a forced distribution system. Every year, each manager must assign 10 percent of employees to the lowest performance tier, regardless of absolute performance. One manager will not put anyone on her team there. She answers in writing instead, certifying that every direct report exceeds expectations.
Then several employees whom other managers placed in the bottom tier file discrimination complaints, arguing that the system produces racially disparate outcomes. The research HR pulls supports them, yet leadership intends to defend the system. Margie’s first move is to pull the defiance and the disparity apart. In her experience, the refusal usually points to a manager’s skill, and the larger risk sits in the data.
What a Manager’s Refusal Usually Signals
A written refusal to follow a mandated process may well meet the test for insubordination. Margie hears a different problem in it. A manager who refuses, she said, “screams leadership capability versus anything else.”
On another HR Voices episode, Kandi Gongora of The Car Group read the same refusal as evidence against the system. Margie allows that it depends on how new the system is and how experienced the manager is. Still, the system is not where she looks first.
“In my experience, when you have outliers, there are managers that are having a hard time reading their team,” she said. The gap is confidence, the skill “to rate their team and assess performance,” and the ability to give feedback. “Conflict is hard to take on,” she added, and for a manager who never built conflict management skills, honest feedback is frightening. That is when “the easiest out is my team is great.”
Research supports reading a rating, in large part, as a measure of the rater. A Journal of Applied Psychology study of two large data sets found that idiosyncratic rater effects accounted for more than half of the variance in performance ratings, 62 percent in one and 53 percent in the other.
Why the Disparate Outcomes Are a Separate Problem
Margie sees “two separate scenarios.” A manager who rates her whole team highly sits “in a little box.” The complaints point to something broader, and she could not yet say whether it was systemic or “in pockets.”
The complaints need the same sorting. Some managers may be treating employees differently, overtly or through bias they do not recognize, which is the ground disparate treatment claims cover. Or the process itself may be generating the skew, so that managers who follow it appear to discriminate.
That second pattern is a disparate impact problem, and good intent does not redeem it. In Griggs v. Duke Power, the Supreme Court wrote in 1971 that Title VII reaches practices that are fair in form but discriminatory in operation.
Margie does “not love the language of forced ranking.” In her current world, managers get guidance, and nothing is forced. Guidance still matters, she argues, because without calibration “everybody thinks their team is the best.”
What Actually Works
Start With the Facts Nobody Has to Tell You
Before any interview, Margie collects “the lay of the land.” She wants to know where the team sits, how it came to be, what its performance history shows, and whether the manager is new.
Then she talks to the manager, because the reason for digging in “could be more complicated than it might appear on paper.” The crux sits with the leader one level up, the “one-over-one,” who sees several teams and can provide the workplace coaching a junior manager needs.
Run Both Lines of Inquiry Before Choosing a Side
Margie describes “two paths of information” to chase down first. One is each employee’s experience and the facts behind it. The other is the data in the system, which shows whether the skew traces to one, two, or three managers or runs across the board.
She would not write the system off as broken, and she would not dismiss employees, or a manager, with real concerns. She warned against letting “emotion help you choose a side” before the facts on both paths are in.
Put Enablement Before Expectations
Margie’s order for managers is fixed. First comes “an enablement phase,” and only then does HR “get to move to expectations.” She calls leading people “probably the hardest job in the organization,” done by people who are “mostly trying to get it right.”
According to Gallup, manager engagement fell from 30 percent to 27 percent in 2024, and even rudimentary training shows benefits. When several managers at one level struggle with the same process, Margie examines the management training they received before she examines them.
Enablement is not leniency. Margie would “not want to come across as a bleeding heart.” When someone is not aligned with the organization’s values, “you should act quickly.” HR can offer the coaching. In her words, “you only control your own behavior.”
Where Employee Relations Fits
By the time this scenario reaches HR, there are investigations to run. Margie’s rule for the complainants is plain. “Any time an employee raises a concern, you should be taking the time to respond to them,” she said. That response is the daily work of an employee relations team.
HR also holds evidence no single manager has. It watches many managers rate and plan compensation, Margie said, and learns who differentiates well and who struggles. She counts that as “valid qualitative data,” provided HR keeps its own bias out.
Rebecca added that performance is where much workplace conflict happens, which is why we built a workflow for performance improvement plans. When an employee does not come through a plan, the record shows why.
Frequently Asked Questions About Manager Defiance and Forced Ranking
Often because rating is hard. Margie finds that outlier managers are usually struggling to read their team, lack confidence in assessing performance, or want to avoid the conflict that honest feedback brings.
Start with facts that need no interview, such as where the team sits, its performance history, and the manager’s tenure. Then talk to the manager and to her manager, where Margie says the crux usually lies.
It does. Margie expects some distribution on a team of ten or more, so people know where they stand. A team of two strong performers is a different discussion.
The data shows the difference. Check whether skewed ratings trace to one, two, or three similarly situated managers or run across the board. Pockets point to people, and a pattern everywhere points to the process.
It means giving managers the tools and skills before holding them to account. Margie starts with how to help a manager get better, then moves to expectations.
The Bottom Line for HR Leaders
Margie Zyble’s framing gives HR an order of operations. The manager gets help first and accountability second, and the disparity stays on its own track, because coaching one manager will not explain a pattern across many.
She closed on the assumption she wants challenged, that HR is not “data-driven” and not “a true business partner.” This scenario tests it. HR teams that separate the manager from the math, and bring evidence on both, are doing the work she describes.
See how AllVoices helps HR teams keep performance decisions and discrimination complaints on a clear, documented track.
