Value of a patent and what really drives it

A patent has no single intrinsic value. Its worth depends on the purpose of the valuation and on who would hold it. For an operating company that sells a protected product, the patent is worth the extra profit it protects by keeping competitors out. For a licensor, it is worth the royalties that companies practicing the invention would reasonably pay. For a buyer building a portfolio, it may be worth what it adds to a broader licensing or defensive strategy. For a lender, it is collateral, and the question is what it would fetch in a forced sale.

The purpose also changes the standard of value. A valuation for financial reporting, a valuation for a tax transfer between related companies, a damages estimate in litigation and a negotiating position in a sale all use similar tools but answer different questions. Before anyone produces a number, it helps to write down why the number is needed and who the hypothetical buyer or licensee is.

• Operating use: the profit protected by excluding competitors.
• Licensing: the royalties that users of the invention would pay.
• Sale: what a buyer would pay for title, usually in a negotiated or brokered deal.
• Collateral: what a lender could recover if it had to enforce its security interest.
• Litigation: the reasonable royalty or lost profits a court might award.

Valuation professionals generally use three approaches, often in combination, and reconcile the results.

The cost approach estimates what it would cost to recreate the invention or to develop an equivalent non-infringing alternative. It is easy to calculate from research and prosecution costs, but it often bears little relation to value, since an expensive research program can produce a patent nobody needs and a cheap insight can produce one an entire industry has to license. The cost approach is most useful as a floor, or for early stage technology with no market data.

The market approach looks at comparable transactions: sales of similar patents, licenses for similar technology, and royalty rates in the same industry. It is intuitive and persuasive when good comparables exist. The difficulty is that patent deals are often confidential, and no two patents are truly identical, so adjustments are always needed for claim scope, remaining life and the terms of the deal.

The income approach estimates the future cash flows the patent will produce and discounts them to present value at a rate that reflects risk. A common variant is relief from royalty: the value of owning the patent equals the royalties the owner avoids paying because it does not need a license, projected over the remaining term and discounted. The income approach is the most flexible, and also the most sensitive to assumptions about sales, royalty rates, term and discount rate.

When the question is what a licensee would pay, the Georgia-Pacific factors are the most widely used framework. They come from a 1970 decision in Georgia-Pacific Corp. v. United States Plywood Corp. and list 15 considerations, including royalties already received for the patent, rates paid for comparable patents, the nature and scope of the license, the licensor's policy on licensing, the commercial relationship between the parties, the effect of the invention on sales of other products, the duration of the patent, the profitability and commercial success of the patented product, the advantages over older alternatives, the portion of profit attributable to the invention rather than other features, expert opinion, and the result of a hypothetical negotiation between a willing licensor and a willing licensee at the time infringement began.

Even outside litigation, walking through these factors is a useful discipline, because they force the analysis to consider apportionment (how much of a product's value comes from the patented feature) and available alternatives. Our article on patent license agreements explains how royalty structures such as running royalties, lump sums and milestones turn a valuation into contract terms.

Remaining term is the most mechanical driver. A US utility patent generally expires 20 years from its earliest non-provisional US filing date, subject to adjustments and extensions in some cases and to terminal disclaimers that can shorten it. Maintenance fees are due at 3.5, 7.5 and 11.5 years after grant, and a missed fee causes the patent to expire early unless it is reinstated. A patent with three years left is worth less than an otherwise identical one with twelve, simply because there is less time to collect.

Claim scope matters more than the title or the abstract. Broad independent claims that cover a commercially important feature, and that competitors cannot easily design around, carry most of the value. Narrow claims with many limitations may be valid but rarely infringed. Valuations that are serious about the patent read the claims and map them to real products.

Validity risk is the discount for the chance that the patent will not survive a challenge. Since the America Invents Act of 2012, anyone can petition the Patent Trial and Appeal Board for inter partes review on grounds of anticipation or obviousness based on patents and printed publications. A buyer will look for strong prior art, the prosecution history and any past challenges. A thorough prior art search is often part of pricing that risk.

Enforceability and detectability matter too. A patent on a method performed inside a competitor's data center may be valid and infringed but very hard to prove. A patent on a visible feature of a consumer product is easier to enforce, and that shows up in value.

• Remaining term and maintenance fee status.
• Breadth of independent claims and ease of design around.
• Validity risk, including exposure to inter partes review.
• Detectability of infringement and quality of evidence of use.
• Patent family coverage in other countries, since a US patent only covers the United States.

Evidence of use is often the single most important commercial driver. A patent that reads on products already on the market, with clear claim charts, is worth far more than one protecting a technology nobody uses yet. Market size matters in the obvious way: a small royalty on a very large product base can exceed a large royalty on a niche product.

Portfolio context also counts. A patent that is part of a family with continuations, foreign counterparts and related patents is easier to enforce and harder to design around. Patents declared essential to a standard can be valuable because implementers must use the standard, but the FRAND commitment that usually comes with them limits the royalty the owner can demand.

Finally, the counterparty matters. The same patent may be worth much more to a company that needs freedom to operate in that space than to a licensing company that would have to litigate to earn anything. Many buyers run an FTO analysis on their own products, and patents that clear a path for them can command a premium.

A patent with an uncertain owner is worth much less than one with a clean chain of title, because the buyer cannot be sure it is buying what it pays for, and a licensor without clear title may not be able to enforce. Before any number is trusted, someone should confirm who owns the patent according to USPTO assignment records and whether the history supports that.

Under 35 U.S.C. 261, an assignment must be in writing, and an unrecorded assignment is void against a later purchaser or mortgagee for value without notice unless it is recorded within three months of its date or before the later purchase. The USPTO records documents without checking their validity, so the records show what was filed, not a guarantee of title. Gaps appear when an inventor never signed an assignment, when a company changed its name or merged without recording the change, or when an assignment was recorded against the wrong patent number.

Encumbrances reduce value even when title is clean. Licenses already granted travel with the patent, so a buyer cannot collect from existing licensees, and licenses do not have to be recorded, so they may surface only in diligence. Security interests give a lender rights in the patent until a release is recorded. Government rights under the Bayh-Dole Act apply to inventions made with federal funding and include a license to the government. Terminal disclaimers can make a patent enforceable only while it is commonly owned with another patent, so separating them in a sale can damage enforceability.

• Is there a recorded assignment from every inventor to the first owner?
• Does the chain run without gaps to the current owner of record?
• Are recorded security interests released, or will they be released at closing?
• Which licenses, covenants not to sue and settlement agreements already exist?
• Do government rights or terminal disclaimers limit how the patent can be used or sold?

Assignee's patent valuation page explains how ownership data supports this part of the analysis.

In practice, a defensible valuation follows a predictable order. First, confirm the asset: the patent number, its status, remaining term, maintenance fee history and the current owner of record with the full chain of title. Second, assess the legal strength: claim scope, prosecution history, prior art and any past or pending challenges. Third, assess commercial relevance: products that practice the claims, market size and the cost of alternatives. Fourth, choose the methods that fit the purpose and the available data, run them, and reconcile the results into a range rather than a single figure.

Documentation matters as much as the number. Buyers, auditors and courts will ask how each assumption was reached. A clear record of ownership with the recorded documents identified by reel and frame number, combined with claim charts and the valuation model, makes a far stronger case than a number on its own.

• Confirm status, term, maintenance fees and the owner of record.
• Review claims, prosecution history and validity risk.
• Map the claims to products and size the market.
• Apply cost, market and income approaches as fits the purpose.
• Present a range with documented assumptions.