What can the recipient actually do with the information?
Confidentiality is only half the story. An NDA also restricts use — what the recipient may do with the information beyond keeping it secret. A recipient who keeps the information confidential but uses it to develop a competing product has still breached most NDAs.
Standard use restrictions limit the recipient to using the information for the purpose of the engagement — evaluating a deal, performing services, working in a job. Anything beyond that purpose is prohibited. Aggressive variants extend the restrictions far beyond the engagement, sometimes indefinitely, sometimes covering entire fields of work the recipient might want to pursue.
The most heavily negotiated provision in this category is the residual knowledge clause — language addressing what happens to information retained in an employee's memory after exposure to confidential material. This clause is often the difference between an NDA that respects employee mobility and one that effectively prohibits the recipient from working in the same industry.
The residual knowledge debate
When an engineer reviews a company's confidential codebase, then leaves to work elsewhere, they don't forget what they saw. The general approaches, the design patterns, the conceptual structure remain in memory. They will inevitably draw on that experience in future work, often without consciously trying to use the specific confidential material.
Residual knowledge clauses attempt to handle this reality. They permit the recipient to use information retained in unaided memory — general concepts, ideas, know-how, techniques — even after the engagement ends, while still protecting specific confidential material like documents, code, customer lists, and trade secrets.
The clause is controversial because it cuts directly against the interest of the disclosing party. Sophisticated technology companies — Google, Microsoft, Apple, Amazon, and most venture-backed startups — routinely insist on residual knowledge clauses in their NDAs because their engineers move between companies and cannot realistically segment their accumulated knowledge. Sophisticated disclosing parties resist them because they functionally permit competitive use of disclosed information.
The clause is controversial because it cuts directly against the interest of the disclosing party. Whether to grant residual rights is a deal-specific judgment call, but its presence or absence dramatically changes how the NDA functions in practice.
What this looks like in real contracts
The Silicon Valley standard for residual knowledge
Major technology employers — and many of their NDAs with vendors, partners, and investors — include residual knowledge clauses as a matter of standard practice. The reasoning is practical: engineers move between companies frequently, and any NDA that doesn't acknowledge the persistence of general knowledge will be either ignored in practice or used as a tool to threaten departing employees, neither of which serves the industry's interest in mobility.
Outside the tech industry, residual knowledge clauses are less common and often absent entirely. In M&A NDAs, the disclosing party typically fights against any residuals provision because the deal-specific information being disclosed is exactly the kind of material a residuals clause could permit a recipient to use after the deal falls through. The absence of a residuals clause in an M&A NDA is normal; its absence in a vendor or employment NDA in technology is unusual.
What NDASentry flags in this category
7.1 Overbroad use restrictions
Use restrictions that extend beyond the stated purpose of the engagement, sometimes prohibiting any 'use' of the information indefinitely. These can effectively bar the recipient from working in related fields, regardless of whether confidential material is actually involved.
7.2 Residual knowledge clause (present or absent)
Either the presence of a residual knowledge clause (which permits use of information retained in memory, favoring the recipient) or its absence (which arguably restricts everything the recipient remembers, favoring the discloser). The pattern is flagged either way because the choice has major downstream impact on what the recipient can do after the engagement ends.
7.3 No-reverse-engineering clause
Restrictions on reverse engineering, decompilation, or deriving underlying ideas from disclosed materials. These can outlast the underlying agreement, conflict with default rights under some state laws (e.g., California permits reverse engineering of lawfully obtained software), and prevent legitimate analysis activities.
We are scoring a corpus of public NDAs to publish prevalence data for each pattern in this taxonomy. The findings — including what percentage of real NDAs contain the patterns above, broken down by industry and jurisdiction — will appear here when the study is complete.