What is the experimental use exception to the on-sale bar?
The experimental use exception to the on-sale bar preserves a patentee's rights when pre-filing transactions serve primarily to validate or perfect an invention rather than to exploit it commercially. Without this doctrine, a closely supervised field test involving a device sale could forfeit the right to patent protection, discouraging inventors from conducting the real-world testing that many inventions require.
Origins of the doctrine
The exception traces to City of Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126 (1877), in which the Supreme Court held that publicly testing an invention does not constitute a public use barring patentability when done in good faith to test its qualities, remedy its defects, and bring it to perfection. Inventor Samuel Nicholson had placed his wooden-block pavement on a Boston turnpike for roughly six years before filing for a patent. The Court declined to invalidate the patent because Nicholson supervised the experiment himself, retained authority over the pavement, did not license it to others, and filed for a patent as soon as testing confirmed the invention worked.
How the exception fits the Pfaff framework
The Pfaff v. Wells Electronics two-part test requires that the on-sale bar attach only when the invention is the subject of a commercial offer for sale not primarily for experimental purposes, and the invention is ready for patenting. The phrase "not primarily for experimental purposes" is structural: it embeds the experimental use exception directly into the definition of what qualifies as an on-sale bar trigger. An offer whose primary purpose is experimentation is not a commercial offer for sale in the relevant sense.
What qualifies as experimental
A sale or offer qualifies as experimental when it represents a bona fide effort to perfect the invention or to ascertain whether it will answer its intended purpose. Experimentation must be the primary purpose, and any commercial exploitation must be merely incidental.
One boundary deserves particular attention: market testing falls outside the doctrine. Market testing, where the inventor is attempting to gauge consumer demand for the invention, does not qualify as experimental use; the purpose of such activity is commercial exploitation, not experimentation. Putting a product in the hands of prospective buyers to learn whether the market wants it is a commercial activity, even if the design is still evolving.
Objective evidence requirement
Courts do not accept a bare inventor's assertion that a transaction was for experimental purposes. An inventor's secretly held subjective intent to experiment, even if true, is unavailing without objective evidence of the experimental purpose; when goods are placed outside the inventor's control in an ordinary commercial environment, objective evidence is required.
Among the most important markers of objective evidence: at a minimum, the customer must be made aware that the sale or offer is for experimental purposes; failure to communicate the experimental character of the transaction to the purchaser is fatal to the experimental use defense. An inventor who quietly treats a commercial shipment as a private test, without informing the recipient, cannot later invoke the exception.
The thirteen factors
Courts evaluate the experimental use exception by considering thirteen non-exclusive factors; no single factor is necessarily determinative. Those factors include: necessity of public testing, degree of inventor control, nature of the invention, length of the test period, whether payment was made, whether a secrecy obligation was imposed, whether experimental records were kept, who conducted the testing, degree of commercial exploitation, whether real-use evaluation was needed, whether testing was systematic, whether the inventor monitored the testing, and the nature of contacts made with potential customers.
The table below maps each factor to the practical question it answers:
| Factor | Key question |
|---|---|
| Necessity of public testing | Could the inventor have tested this in a private or lab setting? |
| Degree of inventor control | Did the inventor retain the ability to direct, modify, or terminate the test? |
| Nature of the invention | Does the invention inherently require real-world conditions to evaluate? |
| Length of test period | Was the duration reasonable for the stated experimental objective? |
| Payment made | Did the counterparty pay commercial prices, or was the device provided at cost or no charge? |
| Secrecy obligation | Were the counterparty and third parties restricted from disclosing the arrangement? |
| Records maintained | Did the inventor keep contemporaneous test logs, protocols, or measurement data? |
| Who conducted testing | Did the inventor or a supervised agent run the experiment, or did the counterparty test independently? |
| Degree of commercial exploitation | Was revenue generated, or was the primary return data and feedback? |
| Need for real-use evaluation | Did the inventor need field conditions to answer an unresolved technical question? |
| Systematic nature of testing | Was there a written test plan with defined success criteria? |
| Inventor monitoring | Did the inventor actively track results throughout the test period? |
| Nature of customer contacts | Were customer interactions framed around evaluation, or around procurement? |
When experimental use ends
The experimental use exception ends with an actual reduction to practice; once inventors determine that the invention works as intended, further activity cannot be characterized as experimental for purposes of the exception. A completed and validated invention that then ships to a second customer is not in experimental use. The inventor has already answered the relevant experimental questions, and further distribution is commercial regardless of how the parties label it.
Personal nature of the exception
The experimental use doctrine is personal to the inventor; activity by independent third parties without inventor supervision and control cannot support the exception. A licensee's internal testing program or a customer's self-directed evaluation run independently of the inventor will not protect a patent from an on-sale bar challenge based on that activity. The inventor must remain in the loop.
Recent Federal Circuit guidance
In Sunoco Partners Marketing & Supply LP v. U.S. Venture, Inc. (Fed. Cir. 2022), the Federal Circuit confirmed that the experimental use determination is a question of law to be analyzed based on the totality of the surrounding circumstances. The court observed that explicit sales language and profit-oriented consideration in an agreement weigh against the experimental use defense, and that acceptance tests alone do not render a use experimental. A commercial supply agreement containing a buyer-acceptance clause will generally not support the experimental use defense simply because testing is built into the delivery process.
AIA applications
Under AIA 35 U.S.C. § 102(a)(1), a claimed invention is not patentable if it was "on sale" before the effective filing date. The experimental use exception developed primarily under pre-AIA case law, and no Federal Circuit decision has definitively resolved whether the exception applies in the same form under post-AIA § 102; the MPEP references pre-AIA case law for post-AIA guidance on this issue. Practitioners advising on AIA applications should treat the exception as available as a defense but consult current case law before relying on it as the primary argument, given that interpretive questions remain open.
Practical notes for practitioners
| Scenario | Experimental use risk |
|---|---|
| Written test protocol, inventor monitoring, counterparty informed | Strong experimental use case |
| Product shipped without disclosing test purpose to recipient | Fatal: customer awareness missing |
| Structured beta test with defined evaluation criteria | Potentially protected if inventor retains control |
| Supply agreement with buyer-acceptance clause | Weighs commercial; acceptance testing alone is insufficient |
| Invention already works as intended; second shipment follows | Exception unavailable; invention is reduced to practice |
| Revenue-generating transaction labeled internally as "pilot" | Market exploitation; label does not transform purpose |
- Document before you test, not after. Draw up a written test protocol with defined objectives and log results as testing proceeds. After-the-fact characterization of a transaction as experimental is difficult without contemporaneous records showing the experimental purpose existed at the time.
- Tell the counterparty it is a test. Silence is not a defense. If the counterparty does not know the transaction is experimental, a court will not accept the characterization. A clause in the test agreement stating the device is under evaluation and subject to modification is minimum practice.
- Retain meaningful control. Shipping a product to a customer who uses it freely, without inventor monitoring, check-ins, or the ability to modify the design, is not experimental use. Maintain control through written check-in obligations, test plan revisions, or inventor site visits.
- Distinguish market testing from technical testing. If the goal is to learn whether a market exists rather than whether the invention works, the experimental use exception does not apply. Consumer preference studies and pilot sales programs are commercial activities.
- Monitor reduction-to-practice timing. Once the inventor concludes the invention works for its intended purpose, the exception is no longer available. Any subsequent transactions, even if labeled a test, are analyzed as commercial activity.
- Keep experimental use records separate from commercial records. The absence of normal commercial paperwork (invoices, revenue recognition, standard sales terms) supports the experimental character of a transaction; their presence weighs against it.