NDA negotiation is supposed to be the fast part. In practice, a mutual NDA between two commercial parties can generate more redline cycles than the underlying commercial agreement. We looked at the patterns across 1,200 NDAs that have come through Clausebeam reviews, and the friction points are not random.
A few clauses generate the overwhelming majority of pushback. Some of those clauses are genuinely high-risk. Others are reflexively redlined by junior associates following firm playbook without fully understanding why. The distinction matters because not all redlines are equal, and knowing which clauses are structurally contested versus stylistically contested is useful intelligence going into a negotiation.
Mutual vs. One-Way Structure: The Starting Fight
The structural framing of an NDA, whether it's mutual or one-way, generates the first contested point in a significant portion of negotiations. The party receiving information typically wants a mutual structure to preserve optionality and reduce the optics of being in a subordinate position. The party providing the information often prefers a one-way structure to limit their own obligations.
We see this go two ways. In early commercial conversations, particularly where both parties are evaluating each other (joint ventures, acquisition discussions, partnership due diligence), the mutual structure is the norm and the one-way version is treated as an overreach. In vendor-to-customer contexts, the one-way structure is more defensible, but vendors still receive pushback when they send one-directional NDAs to companies with active legal teams.
The specific language that triggers redlines in the mutual structure: definitions of "Confidential Information" that are asymmetric even when the protective obligations appear identical. A mutual NDA that defines the receiving party's obligations identically but carves out a broader class of information from Confidentiality for one party is structurally one-sided regardless of the mutual framing. Clausebeam flags this by comparing the definition scope against the protection scope on both sides of the agreement.
Residuals Clauses: The Most Contested Boilerplate in NDAs
The residuals clause is, in our experience, the single most heavily contested provision in technology-sector NDAs. It reads something like: "Notwithstanding the foregoing, either party's personnel who have had access to the Disclosing Party's Confidential Information may use any ideas, concepts, know-how, or techniques retained in their unaided memory..."
The practical effect: if an employee absorbs confidential technical details during a partnership discussion and later uses that knowledge without reference to any written materials, the residuals clause says that's permitted. For companies with mature R&D programs, this provision effectively nullifies a significant portion of the NDA's protective value.
We are not saying the residuals clause is always a deal-breaker or should always be removed. For companies sharing general business strategy or market information, the clause may be a reasonable commercial concession. The problem is that it's often included as boilerplate in templates used for technically-sensitive disclosures where its implications are significant. Law firms advising technology clients consistently flag it. In-house legal teams at companies with proprietary technology consistently strike it.
The residuals clause shows up in approximately 34% of the NDAs we review, and it generates a redline in almost every instance where the receiving party has in-house counsel review.
Return-of-Information Provisions: Where the Details Matter
Post-disclosure obligations, specifically the requirement to return or destroy confidential information at the end of the relationship, generate significant negotiating friction around three sub-issues: timing, scope, and verification.
Timing disputes are common. NDAs that require "prompt" return or destruction are contested by receiving parties who want a defined period (30 or 60 days) to systematically locate and remove information from their systems. "Prompt" creates indefinite obligation.
Scope disputes arise around the question of backup copies and routine system archives. A receiving party's IT infrastructure may retain data in backup systems that are not subject to routine deletion. An obligation to certify complete destruction that doesn't account for this technical reality is either unenforceable or creates a compliance problem the receiving party didn't intend to create.
Verification is the third issue. Some NDAs require written certification of destruction. The party receiving this obligation may not have a mechanism to produce such certification across an enterprise-scale data environment. We see these provisions redlined heavily in NDAs between larger commercial parties, and almost never in NDAs between individual consultants and clients.
Non-Solicitation Overlap: When an NDA Becomes a Restrictive Covenant
A specific category of NDA redline comes from non-solicitation provisions embedded in confidentiality agreements. The logic for including them is understandable: if you're sharing sensitive information about your team and clients during a partnership discussion, you don't want the counterparty to poach your people or approach your clients using that information. The non-solicitation clause is meant to prevent information misuse in that specific direction.
The problem is that non-solicitation clauses in NDAs are often drafted broadly enough to function as standalone restrictive covenants that apply regardless of whether confidential information was actually used. A clause that prohibits soliciting the other party's employees "during the term of this Agreement and for two years thereafter" with no nexus requirement to confidential information is not really an information-protection clause. It's a hiring restriction wearing an NDA's clothing.
Jurisdictional enforceability compounds the issue. Non-solicitation clauses are subject to varying enforcement standards across states, and an NDA that includes them without governing law selection creates uncertainty about the scope of the obligation. California-based companies in particular push back hard on any solicitation restriction in NDAs because of the state's public policy against employee non-competes, which courts have interpreted to limit non-solicitation enforcement as well.
When Clausebeam surfaces a non-solicitation provision in an NDA, we flag it with a note about whether a confidential-information nexus is present and whether the governing law creates enforcement risk. Those two factors together usually determine whether the provision is a commercial concession worth making or a structural overreach worth contesting.
Term and Survival: The Underrated Source of Redlines
The term of an NDA's confidentiality obligation, meaning how long the receiving party must keep information confidential, generates less dramatic redline fights than residuals or non-solicitation provisions, but it's a source of persistent friction. The commercial norm for most bilateral commercial NDAs sits at two to three years from the date of disclosure. Terms significantly longer than that, particularly perpetual confidentiality obligations, are regularly redlined.
The legal basis for contesting perpetual terms is practical: courts in several jurisdictions have declined to enforce perpetual confidentiality obligations on the grounds that indefinite restriction on competition and information use runs against public policy. Whether or not any specific perpetual NDA would survive enforcement challenge, the receiving party's lawyer is likely to flag it.
Survival provisions create a related question: does the confidentiality obligation survive the termination or expiration of a commercial relationship built on the NDA? In many cases, the answer should be yes, but the interaction between an NDA's survival clause and the survival clause in a later master services agreement or licensing agreement can create duplicate obligations that are inconsistently drafted. We flag situations where an NDA survival period overlaps with a later agreement's survival language in a way that could create conflicting obligations, because that inconsistency tends to surface during dispute resolution at the worst possible time.
What This Means for NDA Review Workflows
If you're reviewing NDAs at volume, the practical insight from this analysis is that most NDA redlines are predictable. Residuals clauses, overly broad non-solicitation provisions, return-of-information mechanics, and term length account for the majority of negotiating friction we observe. A first-pass review workflow that systematically identifies these four categories reduces the time an attorney spends scanning for issues and increases the time spent on the ones that are actually contested.
That's what we built Clausebeam to do for NDA review: surface the structurally predictable friction points before an attorney opens the document, so the attorney's time goes to judgment calls rather than identification work. The identification step is rule-based enough to be automated. The negotiation judgment is not.
