Trade-secret law protects commercially valuable information only when its holder has taken reasonable measures to keep it secret. The requirement is sometimes described as proof that the owner genuinely valued confidentiality: a firm that leaves a formula on a public website should not later call it secret. Yet courts that equate reasonable measures with maximal restriction misunderstand both secrecy and modern production. Useful information must be shared with employees, suppliers, and sometimes customers. Absolute secrecy would preserve information by preventing the activity that gives it value.

The best account of the requirement is boundary maintenance. Passwords, limited access, confidentiality clauses, and exit procedures identify which information a firm treats as restricted and communicate that boundary to those who receive it. These measures reduce accidental disclosure and make deliberate appropriation harder to excuse as misunderstanding. Their legal significance lies less in making leakage impossible than in separating protected knowledge from ordinary workplace skill and publicly available information.

That account also explains why reasonableness must depend on context. A two-person design studio cannot deploy the monitoring systems of a pharmaceutical company, while a pharmaceutical company cannot defend leaving clinical formulas in an unrestricted shared folder merely because a smaller firm might do so. Courts should ask whether the precautions fit the information's value, the foreseeable routes of disclosure, and the holder's resources.

More protection is not always better. Restrictions drafted so broadly that they label an employee's general experience secret can suppress lawful mobility and obscure the very boundary the measures are supposed to mark. A coherent doctrine therefore rewards precautions that are specific and proportionate, not precautions that simply burden the greatest number of people.

The boundary account changes how courts should treat a breach. If an outsider defeats careful controls through deception, the breach does not retroactively prove that the information was never secret; otherwise the most determined theft would erase the protection it violated. By contrast, repeated uncorrected disclosures to people who receive no notice of restriction may show that no intelligible boundary existed. The relevant evidence is the system of communication and response, not the bare fact that information escaped.

There is, however, a danger in making context an excuse for hindsight. After valuable information is taken, almost any omitted precaution can be described as reasonable to have expected. Courts should focus on risks foreseeable before the dispute and compare the challenged practice with workable alternatives available then. Industry custom is evidence but not a safe harbor: an entire industry may tolerate a cheap, obvious vulnerability. The proportional inquiry is demanding because it asks whether the holder maintained a comprehensible boundary under real operating conditions, not whether counsel can compile an impressive list of restrictions after litigation begins.