Urban Meyer won national championships at Florida and Ohio State. In Netflix’s 2023 documentary series Untold: Swamp Kings, about his Florida teams, a former player recalls Meyer’s philosophy in words that have been widely repeated since, roughly: treat your stars like stars, and treat the worst players like, well, a word we will not print here. (Retellings vary on the exact wording.) The player’s point was that every coach does it, but few say it out loud.
Let’s be clear about what this article is not. It is not a defense of that sentence. A company that adopted those words as its written policy, or let a manager say them, would be handing every plaintiff’s lawyer in the state an opening exhibit. You would be sued, and you would deserve the bill.
But the question underneath is a serious one, and most HR departments never ask it: is it actually illegal to treat high performers better than low performers?
The short answer: usually not
Performance is not a protected characteristic. No federal or Texas law requires an employer to treat a top performer and a bottom performer the same way. There is no general American law against being a demanding boss, and Texas has no workplace civility statute. Paying stars more, promoting them faster, giving them better assignments, more freedom, and more attention is not just legal. It is how merit is supposed to work.
The law steps in when different treatment is connected to something it does protect:
- Performance as a proxy or a cover. If the “stars” are disproportionately one race, sex, or age group because of who gets the good assignments, the mentoring, or the benefit of the doubt, the performance label becomes evidence of discrimination, not a defense.
- Harsh treatment that tracks a protected trait. Being tough on low performers is legal. Being tough in ways that are tied to sex, race, age, disability, religion, or national origin can become a hostile work environment.
- Systems with a disparate impact. A neutral-looking rating system that falls harder on a protected group can violate federal law unless it is job-related and consistent with business necessity. Age claims are possible too. Ford paid $10.5 million in 2002 to settle class actions challenging its forced-ranking system as age discriminatory, and Microsoft’s stack ranking was at the center of a gender-bias class action.
- Retaliation. A low performer who complained about discrimination, wages, or safety cannot be pushed harder because of the complaint.
- Disability, pregnancy, and leave. Low performance caused by a disability or pregnancy may trigger an accommodation duty, and FMLA leave cannot be counted against someone.
- Wage laws. Stars who are non-exempt still get overtime, and “pushing” anyone cannot mean unpaid work or skipped breaks.
- Favoritism. Ordinary favoritism, and even favoring a manager’s romantic partner, is generally not sex discrimination under federal law. Widespread favoritism based on sexual relationships can be.
- Labor law and contracts. Treating people differently because of protected concerted activity is unlawful, and union contracts and individual agreements can limit differentiation.
Where HR gets it wrong: equal rules are not equal outcomes
Most HR systems default to treating everyone identically, no matter what. The larger the organization, the more burdensome that becomes. And the rules are almost always written for the bottom quintile: the approval chains, the monitoring, the rigid schedules, the documentation requirements. They exist because of the people who need them. The upper quintile lives under the same rules and is held back by them. Over time, the best people decide they can get more freedom somewhere else.
The law does not require that. What anti-discrimination law requires is consistency among employees who are similarly situated. Employees with documented, very different performance are not similarly situated. The legal goal is not identical treatment for everyone. It is identical treatment for everyone in the same documented position, and a clear, provable reason for every difference. That is the opening a good system uses.
Why the coach’s philosophy works, and why it is harder at work
On a football team, performance is visible. Everyone watches the same film, and playing time follows it. Stars get latitude because they have earned trust, and others are pushed because the standard is clear. It is merit as the forcing function, which is exactly what most workplaces lack. At work, performance is rarely that visible, so differential treatment looks like favoritism unless the system makes the standard visible first.
Designing it defensibly
1. Make the tiers real, objective, and movable
Build performance tiers on output metrics, the results people produce, not hours, likability, or visibility to leadership. Write down the criteria. Calibrate ratings across managers so one generous manager does not create a team of “stars.” Re-tier on a schedule, and make it possible, and common, to move up. A tier is a description of recent results, not a caste.
2. Default rules, earned exemptions
This is the most useful idea in this article. Keep the rules written for the bottom quintile as the default for everyone, and let people earn their way out of them. An “earned autonomy” ladder might include:
- Higher approval and spending limits.
- More schedule flexibility, within wage-and-hour rules.
- Lighter reporting and fewer check-ins.
- First choice of projects, territories, or shifts.
- Training and conference budgets, mentoring from senior leaders, and stretch assignments.
Every exemption has written criteria, applies to everyone who meets them, and can be lost if performance drops. Stars get treated like stars, and the reason is on paper. Be careful with any privilege that might also be a reasonable accommodation, such as remote work, because an accommodation does not depend on tier.
3. Reward stars in ways that compound
Differentiate pay, bonuses, and equity meaningfully, not by a token percentage. See incentive pay done right. A documented merit or production-based system is itself a recognized defense to an equal pay claim.
4. Push low performers with structure, not scorn
“Pushing harder” should mean more structure: specific, measurable improvement goals with a deadline, weekly check-ins, closer review of work product, coaching and training, and less autonomy until results improve. That is demanding, it is fair, and it is defensible. If the goals are not met, the record supports the next step. See modern discipline in an at-will state.
5. A dignity floor for everyone
Whatever someone’s tier, some things never change: respect, safety, lawful pay and breaks, freedom from harassment, and full access to complaint channels. No public humiliation. The tiers decide privileges and pressure, never basic dignity. That floor is what separates a high-performance culture from a lawsuit.
6. Build in the guardrails
- Bias checks: structured criteria, more than one rater, and calibration sessions with written reasons.
- Impact audits: periodic statistical review of tier outcomes by protected group, done through counsel, with job-relatedness documented for every metric.
- Accommodation check: before anyone moves down a tier, ask whether disability, pregnancy, or protected leave is involved.
- Retaliation firewall: extra review for any tier change close in time to a complaint or other protected activity.
- No quotas: avoid forced distributions that require a set percentage to be rated low, the feature most often challenged.
- An appeal route: a way for employees to question a rating, reviewed by someone outside the chain.
- Neutral names: call tiers something like “standard,” “advanced,” and “lead.” Every word you write is discoverable.
In the real world, this gets messy
On paper, a tiered system is clean. In practice, it drifts. Managers rate the people they like. Metrics that were neutral start correlating with protected traits. A privilege granted once becomes an expectation, and an exception becomes the precedent the next employee points to. A star with a disability, a low performer who just filed a complaint, a reorganization that scrambles the metrics: each of these is where a defensible system becomes an exhibit.
That is why the hard part is not the design. It is making sure the system actually does what it says, every period, for everyone. Before you build one, and at least once a year after, have it reviewed. Crawford Law designs performance tiers, earned-autonomy programs, and improvement plans, audits how they are working in practice, and helps fix the drift before it turns into a claim. Merit should be the forcing function in your business. The system should prove that it is.
Sources
- Netflix, Untold: Swamp Kings (2023), documentary series on the Florida Gators under Urban Meyer
- OutKick, Urban Meyer’s Coaching Philosophy At Florida Was Brutally Honest (2023)
- Title VII of the Civil Rights Act, 42 U.S.C. § 2000e-2 (disparate treatment and disparate impact)
- Smith v. City of Jackson, 544 U.S. 228 (2005) (disparate impact under the ADEA)
- Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011)
- Fortune, Microsoft, GE, and the futility of ranking employees (2013) (Ford’s 2002 settlements)
- Moussouris v. Microsoft Corp., No. 2:15-cv-01483 (W.D. Wash.)
- U.S. EEOC, Policy Guidance on Employer Liability under Title VII for Sexual Favoritism (1990)
- Equal Pay Act, 29 U.S.C. § 206(d) (merit and production systems)