A Rollins Inc. pest-control training video sits on YouTube. Public. Free. Anyone with a browser can watch a technician walk through the methods, step by step. A few miles from the office where that video was filmed, a former Rollins tech opens a cease-and-desist letter. It tells him those same methods are trade secrets worth two years of his working life. The contract was never about a secret. The secret was already on the internet.

"A noncompete protects the company's investment in you." If you have spent time in a senior role or run your own shop, you have heard some form of this. It sounds fair. It is not. It is a fear tool dressed in legal language.

The research on this is worth a close read. Rollins, one of the largest pest-control firms in the country, locked more than 18,000 workers under a blanket two-year, 75-mile noncompete. Every role. Every level. The FTC looked into it and found the case for those contracts hollow. The company's own methods were posted on its website and in YouTube videos for anyone to see. Rollins offered the same training in states where it did not use or enforce noncompetes at all. FTC Chairman Andrew Ferguson and Commissioner Mark Meador put it plainly: "We come up nearly empty" on any real reason for the contracts.

Rollins kept enforcing them anyway. The company sent hundreds of cease-and-desist letters to former workers who had started their own shops. Many of those people shut down or pulled back. Not because a court told them to. Because they could not afford to fight a company that size.

The contract did not need to hold up in court. It needed to scare.

There is a name for this in the literature. Economist Evan Starr, with J.J. Prescott and Norman Bishara, published findings showing that noncompetes cut worker mobility at the same rate in states that enforce the contracts and states that do not. Workers turn down offers from rivals at similar rates in both settings. The law is not the engine. The belief is.

A follow-up study by Prescott and Starr found that workers tend to believe their noncompetes bind them even when the law says the opposite. Worse: firms keep their workers in the dark about this on purpose. The confusion is not a side effect. It is the product.

The noncompete does not fail people. It fails the system they are trying to build.

The flaw is not that companies want to keep good people. The flaw is that a noncompete replaces a real retention plan with a fear tool that adds nothing to information security. A 2024 field experiment by Cowgill, Freiberg, and Starr tested this head-on. Signing a noncompete cuts worker mobility by 57 percent. It cut the sharing of guarded information to zero. A nondisclosure agreement alone did the same job. The noncompete added nothing except a cage.

That is the number worth saying out loud: 57 percent less mobility. Zero percent less leakage.

The better principle: treat a noncompete as a claim to be tested, not a wall to be obeyed. Once that is clear, three moves follow from it.

Fourteen states now set wage floors or job criteria that limit who can be bound by a noncompete. Oregon banned them for hourly workers years ago. When researchers Lipsitz and Starr studied the result, they found wages rose 2 to 3 percent on average, and 14 to 21 percent for workers who had actually been bound. That wage gap is the price tag of fear, measured in real dollars you did not earn. This step takes an hour with an employment attorney in your state, not a web search. The law is specific enough that a general reading will mislead you.

The FTC itself kept non-solicitation agreements in place as valid tools in the Rollins order. A narrowly written NDA or non-solicitation clause protects client ties and guarded data without locking you out of your own field. The split matters because many workers treat all post-employment restrictions as one thing. They are not. One restricts what you can say. The other restricts where you can work. Only one of those survived the FTC's own test.

If your noncompete runs two years in a 75-mile radius, build client ties and skills in the space the contract does not cover. Start before you leave. The restriction has edges. Map them. Most operators treat the noncompete as a blanket. It is a fence with known lines.

Running these three moves over 90 days does something the old advice never did:

You see which parts of the contract are real and which are theater. You find the gap between what the company claims and what the law allows. You learn whether the restriction has teeth or just a loud bark. And you stop making career choices based on a document you never checked against the statute that governs it.

At the end of 90 days, ask three things.

→ What part of the restriction is real, confirmed by an attorney, not by your employer's HR page?

→ What felt like a wall but turned out to be a bluff?

→ What skill or relationship did you build in the open space you found?

That is the difference between advice that sounds right and a system that proves itself.

One-third of workers see their noncompete for the first time after they have already accepted the job. Eighty-six percent are promised nothing extra for signing. Only 10 percent try to negotiate. Prudential Security kept mailing the same noncompete to its guards after a court ruled it void. The companies know the contract runs on fear. Now you know the name of the mechanism.

The training videos are still on YouTube. The methods are still public. The noncompete never protected a secret. It protected a roster.