What Makes a Patent Valuable, and Why Detectability Matters

A valuable patent often covers what customers actually pay for, holds up to scrutiny of its validity, and is hard to design around. Another practical consideration is easy to overlook when the application is drafted: detectability. If it is difficult to tell from the outside whether a competitor’s product uses what your claims describe, licensing, enforcement, and valuation can all become harder and more expensive.

Detectability is shaped largely when the claims are written. This article explains what it means, how it relates to other value factors, and how to draft with it in mind.

The factors that make a patent valuable

Assessments of patent value commonly consider questions such as these, along with licensing potential, defensive leverage, standards-related coverage, transferability, and the strategic uses of the patent:

Coverage: do the claims cover the reason customers buy the product, or a product containing it?
Detectability: can infringement be identified from evidence available outside the infringer’s walls?
Design-around difficulty: can a competitor deliver the same benefit without practicing the claims?
Strength: are the claims well supported by the specification and likely to hold up against prior art?
Ownership and term: is ownership clean, and how much of the patent term remains?
Room to grow: is a related application still pending, so new claims can follow the market?

Detectability interacts with the other factors. Claims that cover exactly the right feature can be harder to license or enforce if it is difficult to show that a competitor uses that feature. It is one consideration among several, and its weight depends on the technology and on how the patent will be used.

What detectability means

A claim is detectable when someone can determine whether a product infringes it using information available without access to the other company’s internal systems: the product itself, its behavior, its documentation, its marketing, or public technical sources.

Before filing a patent lawsuit, a patent owner generally needs a reasonable basis, from a pre-suit investigation, to believe the claims are infringed. Licensing discussions and investor diligence work similarly in practice: the patent owner usually needs to point to evidence. A claim whose elements can be matched to public evidence supports those conversations. A claim that depends on facts hidden inside a competitor’s servers or factory is harder to support at those stages.

Detectability is not, however, a requirement for validity or infringement. A claim that is hard to investigate before litigation is not impossible to prove, since evidence can be obtained through discovery once a case is filed, and patents directed to nonpublic processes can still have significant commercial value. The importance of detectability should be weighed alongside claim scope, validity, enforceability, available remedies, market relevance, and the strategic uses of the patent.

Where evidence of use comes from

Depending on the technology, infringement can often be observed through:

The product itself: physical teardown of hardware, inspection of circuit boards and chips, and measurement of signals and performance
Observable behavior: what a device or application does, what it displays, and how it responds to particular inputs
Data that crosses a boundary: messages exchanged between a client and a server, wireless transmissions, and API requests and responses
Documentation: user manuals, data sheets, developer documentation, support articles, and published specifications
Marketing and public statements: product pages, white papers, conference talks, and job postings that describe how a product works
Industry standards: when a product complies with a standard and the claims track the standard’s requirements
Public code: open-source components and published software, where available

Some of these sources come with legal limits. For example, license agreements or terms of use may restrict reverse engineering, and other laws can apply as well. Whether such restrictions are enforceable depends on the circumstances, so the investigation itself needs care.

What is hard to detect

Some of the most valuable technical ideas are also the hardest to observe:

Algorithms that run only on a company’s own servers
Training methods for machine learning models, as opposed to how the trained model behaves
Internal data structures and storage arrangements
Manufacturing and test processes inside a factory
Steps that happen in a different country from where the product is sold

None of these make a patent worthless, and some can be proven through discovery once litigation begins. But claims built entirely around hidden steps can make the stages before that, from licensing to diligence to the decision whether to sue, harder and more expensive.

Drafting claims you can prove

Detectability is largely a claim drafting question, and it is best considered before the application is filed. Several techniques can help, where the invention and disclosure support them.

Claim from the observable side. If the inventive idea involves a client and a server, consider which side’s behavior can be seen. Claims directed to what the client device sends, receives, and displays are often easier to prove than claims to the server’s internal processing. Where possible, include claims from each actor’s perspective.

Claim results and interfaces, not only internals. A claim that requires a particular output, message format, signal characteristic, or user interface behavior can often be tested directly. A claim that requires a particular internal sequence of operations may not be.

Keep each claim to a single actor. When a method claim’s steps are performed by different parties, such as a user and a service provider, proving infringement becomes more complicated, because direct infringement of a method generally requires that all steps be performed by, or attributable to, a single entity. Framing claims around steps performed by one party can reduce that problem, though whether a particular entity infringes still depends on the facts.

For AI inventions, include claims on how the model is used. Training methods are rarely visible from outside. How a trained model is applied, what inputs it takes, and what the system does with its output often are. Claims directed to the application of the model can sit alongside claims on the training method.

Use measurable parameters. Where a claim recites a threshold, timing, or range, choose terms that can be measured from outside the product and define them clearly in the specification.

Track standards where they apply. For communications and other standardized technologies, claims that map to required features of a standard can be shown through compliance documentation and testing.

Each of these needs to be balanced against breadth and validity. A detectable claim that is easy to design around, or that is anticipated by prior art, helps little. In practice, many claim sets include a mix: broad claims on the core idea and narrower claims written specifically to be provable.

When a patent may not be the right tool

If an innovation is truly invisible from outside, such as a back-end process that cannot be observed or reverse-engineered, it is worth asking whether a patent is the best way to protect it. A patent application is generally published, which discloses the method to competitors. For some hidden processes, keeping them as trade secrets may offer more practical protection, provided the company takes reasonable steps to keep them confidential.

That is a business decision as much as a legal one, and it is worth making deliberately before the application is filed rather than after.

Questions to ask before filing

When reviewing a claim set, it helps to ask these questions of each independent claim:

If a competitor used this, how would we find out?
Which public sources would show each element of the claim?
Does any element require knowledge of a competitor’s internal systems?
Are all steps performed by one party?
Is there a claim written from the perspective of the product the customer actually buys?

Your patent attorney or agent can work through these questions with you and adjust the claims before filing.

A practical takeaway

A patent’s value depends on more than whether it is granted. Before filing, it is worth asking whether the claims cover what customers pay for, and whether at least some of them could be supported with evidence obtainable from the market. Detectability is far easier to consider while the application is being drafted than to add afterward.