How to Value a Patent: 4 Methods + IP-Lending Context

TL;DR. Four canonical methods value a patent: cost, market, income/DCF, and relief from royalty (RFR). The choice is governed by purpose. Cost suits early-stage R&D recovery; market needs comparable transactions that rarely exist; income/DCF and MPEEM suit revenue-generating patents with clean cash-flow attribution; RFR is the workhorse for licensed or licensable patents and the method most UK IP-backed lenders prefer. The right answer is usually a triangulation — two methods reconciled, with the third as a sanity check.

The first question a founder, CFO, or investor asks about a patent is "what is it worth?" The honest answer is "to whom, and for what purpose?" A patent valued for a tax-relief claim under the UK Patent Box is a different number from the same patent valued as loan collateral by NatWest, which is a different number again from the same patent valued in an IFRS 3 purchase price allocation. Same asset, three legitimate numbers.

This guide sets out the four canonical valuation methods, when each one fits, what inputs each requires, and what defensibility each produces. It closes with the UK IP-backed lending angle — what NatWest, HSBC, and RBS need to see to underwrite a patent as collateral, and which method survives that lens.

What you are actually valuing

A patent is not the invention. A patent is a 20-year legal monopoly granted by the UK Intellectual Property Office (or USPTO, or EPO) over a defined claim set, under sections 1-6 of the Patents Act 1977. The thing being valued is the cash-flow advantage that monopoly confers — the difference between what the patent holder can charge or save versus a non-patented competitor, multiplied by the remaining enforceable life, discounted for risk.

That framing matters because it filters which methods produce defensible numbers. A method that cannot articulate the monopoly cash-flow advantage cannot defend the valuation under scrutiny.

The four methods at a glance

Before the deep dives, a method-selection decision tree. Pick the column that matches your purpose; the row tells you which method to lead with, what inputs you need, and what defensibility the output will carry.

Three observations from the matrix. First, RFR appears in five of six rows — it is the most universally applicable method, which is why it has become the workhorse of practitioner valuation. Second, the cost approach is almost never the lead method; it is a floor, not a value. Third, no method stands alone — the auditable answer is always a triangulation.

Method 1: The cost approach

The cost approach values a patent at what it would cost to recreate the same asset today, less an obsolescence adjustment. In its simplest form: sum the R&D spend that produced the patent, add a profit margin a third party would have demanded to do the work, subtract depreciation for technical obsolescence.

When it fits. Early-stage assets with no revenue, no licence history, and no observable market. Tax-relief claims where HMRC accepts a reproduction cost basis. Insurance valuations. Floor values in negotiations.

Inputs. The R&D ledger broken down by project, capitalisable costs under IAS 38 (development phase only, not research), a profit-margin assumption (typically 10-25% for technical work), an obsolescence factor (technology age, claim narrowness, competitor activity).

Worked example. A UK life-sciences SME spent £1.8m over three years developing a diagnostic assay. £0.4m was research (expensed under IAS 38) and £1.4m was capitalisable development. A third party would have charged a 20% margin to perform the same work. Technical obsolescence over the development period is estimated at 15%.

Capitalisable development spend:    £1,400,000
+ Third-party margin (20%):         £   280,000
= Reproduction cost (gross):        £1,680,000
- Obsolescence adjustment (15%):    £  (252,000)
= Cost-approach value:              £1,428,000

That number is what the patent is worth as a floor — the price below which the patent holder would rationally refuse to sell, because for any less they could simply replicate the asset themselves. It is not what the patent is worth to a buyer who values the monopoly. The buyer's number is almost always higher, and the gap is the whole point of doing a proper valuation.

Defensibility. The cost approach is auditable but rarely the right answer for a commercially active patent. It under-values successful patents (which generate cash-flow advantages far above their build cost) and over-values failed ones (which generate none). HMRC accepts it for some Patent Box working papers and R&D tax-credit reconciliations. Lenders treat it as a sanity-check floor only.

Note. Under IAS 38, only the development phase of internal R&D is capitalisable, and only when six specific criteria are met (technical feasibility, intent to complete, ability to use or sell, future economic benefit, adequate resources, reliable measurement). Research spend is expensed as incurred and does not enter the cost-approach calculation.

Method 2: The market approach

The market approach values a patent by reference to recent transaction prices for comparable patents. The logic is identical to property valuation — find three sales of similar assets, adjust for differences, take the mean.

When it fits. Patent portfolios traded at auction (the Ocean Tomo and IPwe marketplaces, occasional patent-pool transactions). Standard-essential patents (SEPs) where licensing rates are public. Sector-specific transactions where Companies House filings or SEC 8-K disclosures reveal the price.

Inputs. Comparable transaction prices, normalisation factors (remaining patent life, geographic coverage, claim breadth, enforceability history), market timing adjustments.

Note. The market approach for patents has a structural data problem. Most patent sales happen inside M&A bundles where the patent-specific price is never broken out. Public databases capture perhaps 5-10% of actual transactions, and even those rarely disclose enough context to normalise against your specific claim set.

Why it rarely works. Patents are not fungible. Two semiconductor patents in the same sub-class can have wildly different economic profiles depending on claim scope, prior-art density, and family geography. Public transaction data is also sparse — most patent sales happen inside M&A bundles where the patent-specific price is never broken out. The Ocean Tomo Patent Index data, while useful as a sector signal, is statistically thin for any specific valuation question.

The market approach is most defensible when used as a corroborating method: "the RFR analysis produces £4.2m; the three most-comparable transactions in the sector imply a £3.5m-£5.0m range; the RFR result is within the market band." That's a strong triangulation. Leading with the market method on a unique patent is not.

Defensibility. Medium to low for unique patents. High when comparable transactions exist and are properly normalised. Auditors challenge the comparability assumption hard.

Method 3: The income approach (DCF and MPEEM)

The income approach values a patent at the present value of the economic benefits it produces, discounted at a rate reflecting the asset's specific risk. Two variants matter in practice.

Discounted cash flow (DCF). Forecast the incremental cash flows the patent generates over its remaining enforceable life, discount them to present value. The hard part is isolating the patent's contribution from the contribution of the rest of the business — the product, the brand, the workforce, the customer relationships. Without that isolation the DCF over-states the patent's value.

Multi-Period Excess Earnings Method (MPEEM). The technical solution to the isolation problem. MPEEM forecasts the cash flows of the patent-protected product, then subtracts contributory asset charges (CACs) — notional rents for the working capital, fixed assets, workforce, and other intangibles that the product also needs to generate those cash flows. What remains is the excess earnings attributable to the patent itself. See MPEEM for the full mechanic.

When it fits. Patents embedded in revenue-generating products with clean attribution. IFRS 3 / ASC 805 purchase price allocations where MPEEM is the AICPA-preferred method for the most-important intangible asset. Litigation damages cases.

Inputs. A defensible revenue forecast for the patent-protected product, gross-margin path, contributory asset charges (typically 8-15% of revenue for working capital and fixed assets, plus a workforce charge based on assembled-workforce values), a patent-specific discount rate (usually base WACC plus 2-5% for the asset-specific risk).

Worked example (MPEEM). A UK medical-device SME holds a patent that gives its sole product a 35% gross margin versus an estimated 18% without the patent. Forecast revenue rises from £8m in Year 1 to £18m in Year 7, then plateaus to patent expiry in Year 11.

Year 1 patent-product revenue:           £8,000,000
× Patent-attributable margin (17pts):    × 17%
= Year 1 excess earnings (pre-CAC):      £1,360,000
- Contributory asset charges (~6% rev):  £ (480,000)
= Year 1 excess earnings (post-CAC):     £  880,000

(Repeat for Years 2-11, discount at 14%)

Sum of discounted excess earnings:       £6,200,000 (illustrative)

The £6.2m represents the cash-flow advantage the patent confers, isolated from the contributions of all the other assets the business uses to deliver the product. That is what an acquirer should pay for the patent specifically — and what an auditor under IFRS 3 expects to see in the PPA working papers.

Defensibility. Highest of any method when the inputs are well-evidenced. The discount rate, the contributory asset charges, and the revenue forecast each carry meaningful judgment, so the working papers matter as much as the answer.

Example. When the same medical-device patent is valued for an IP-backed loan, the MPEEM number is the starting point but not the answer. The lender applies a recovery-scenario haircut — "if the borrower defaults and we have to sell this patent in 90 days, what does it fetch?" — which typically lops 40-60% off the MPEEM result. The £6.2m becomes a £2.5m-£3.7m collateral value.

Method 4: Relief from Royalty (RFR)

Relief from royalty values a patent by asking a simple counterfactual: if the patent holder did not own the patent, what royalty would they have to pay a third party to licence it? The present value of those avoided royalty payments is the patent's value to the holder. Mechanically, RFR multiplies forecast patent-product revenue by a defensible royalty rate, then discounts the resulting cash flows over the patent's remaining life. See RFR for the full method statement.

When it fits. Almost every commercial patent valuation. It is the RFR method's combination of conceptual clarity, input availability (royalty databases like RoyaltyStat and ktMINE), and audit acceptance that has made it the workhorse method. UK IP-backed lenders prefer it. IFRS 3 auditors accept it for technology-class intangibles. HMRC accepts it for Patent Box working papers.

Inputs. A forecast revenue stream for the patent-protected product, an arm's-length royalty rate (sector-specific, typically 3-8% for software, 4-12% for medical devices, 1-3% for industrial machinery), a tax-rate adjustment (the royalty saving is pre-tax), a discount rate, the remaining patent life.

Worked example (RFR).

Year 1 patent-product revenue:           £ 8,000,000
× Arm's-length royalty rate (6%):        × 6%
= Pre-tax royalty saving:                £   480,000
× (1 − UK corporation tax 25%):          × 75%
= After-tax royalty saving:              £   360,000

(Repeat Years 2-11 with revenue growth,
discount at 14% for asset-specific risk)

Sum of discounted royalty savings:       £4,400,000 (illustrative)

Note the £4.4m RFR result is materially below the £6.2m MPEEM result for the same patent. That gap is the difference between "what the patent contributes to the business" (MPEEM) and "what the holder saves by owning rather than licensing the patent" (RFR). RFR is typically the more conservative of the two, which is exactly why lenders and auditors lean on it.

Defensibility. High. The royalty-rate evidence is the contentious input, but commercial royalty databases provide arm's-length comparables that auditors and HMRC accept. The method aligns naturally with the OECD Transfer Pricing Guidelines, which simplifies cross-border patent valuations.

Triangulation: how to actually report the number

A defensible patent valuation almost never relies on a single method. The institutional pattern is:

1. Lead method. Pick the method that best fits the purpose (use the matrix at the top of this article).
2. Primary cross-check. Run a second method as a corroborating analysis. RFR and MPEEM cross-check each other particularly well — they should produce numbers within 20-30% of each other for a healthy commercial patent.
3. Sanity check. Run the cost approach as a floor. The lead method's result should comfortably exceed the cost floor; if it doesn't, the patent is probably commercially weak and the working papers need to explain why.
4. Reconciliation. Document the gap between the methods and explain it. The gap is not a problem — it is information. An MPEEM result materially above the RFR result means the patent contributes value beyond pure licensing economics, which is itself a defensible position.

The UK IP-backed lending lens

The brief promise of this guide is to close on the IP-backed lending context, because that is where most UK founders meet patent valuation in commercial practice — when they ask NatWest, HSBC, or RBS to lend against the patent.

UK IP-backed lenders have a specific lens. They are not asking "what is the patent worth in an orderly transaction?" They are asking "what is the patent worth in a forced-sale scenario, in 90 days, with us as the seller?" That lens collapses the four methods into a hierarchy that looks different from the auditor's or the acquirer's.

Warning. A patent that produces a £6m MPEEM result and a £4.4m RFR result will typically support £1.5m-£2.5m of senior debt under UK bank haircut models. Founders who quote their MPEEM number as borrowing capacity are signalling unfamiliarity with the lending lens — and lenders notice.

What lenders actually want to see.

1. An RFR valuation as the headline. The methodology aligns with their underwriting models. The royalty-rate evidence is auditable. The tax-adjusted, discounted result is the basis they will haircut.
2. A cost-approach floor. They want to know the floor. Patents that have not yet generated revenue trade at or near reproduction cost in distressed sales; the cost approach tells them where the realisable floor sits.
3. A recovery-scenario adjustment. They will apply their own haircut — typically 40-60% of the RFR result — to model what the patent fetches in a 90-day liquidation. The borrower's job is to provide the inputs, not to perform the haircut.
4. Evidence the patent is enforceable. A patent that cannot be enforced is worthless as collateral. Validity opinions, freedom-to-operate analysis, prior-art reviews, and litigation history all matter. The valuation number is meaningless if the patent fails the enforceability test.
5. Evidence of marketability. Lenders need to know there is a credible secondary buyer in a default scenario. A patent in a sector with active M&A and patent-pool activity is more valuable as collateral than a patent in a thin or proprietary market — even if the RFR result is identical.

For the full UK lender-specific criteria, see IP-backed lending criteria and the bank-specific notes for NatWest. The IP lending calculator lets you model the borrowing capacity an RFR-valued patent supports under typical UK haircut assumptions.

US and EU notes

Three brief notes on jurisdictional variation, since patents valued in the UK frequently support transactions in the US or EU.

US. The AICPA's "In Process Research and Development" guide and ASC 350/805 align tightly with the IFRS 3 / MPEEM-and-RFR pattern. US Patent Box-equivalent reliefs (the Foreign-Derived Intangible Income regime under §250) require similar working papers. USPTO patent terms run from filing under 35 U.S.C. §154, which can produce different remaining-life inputs from UK Patents Act calculations.

EU. Most member states accept the same four-method framework but with national Patent Box variations — the Netherlands, Belgium, and Ireland each have specific qualifying-IP rules that affect the income-method inputs. The European Patent Convention provides a centralised filing route but national enforceability still drives the valuation cash flows.

The methods are universal; the inputs and the haircuts are jurisdictional.

Where to start

If you are valuing a patent for the first time, the practical sequence is:

1. Define the purpose. Tax, lending, M&A, litigation, internal management? The purpose drives the method.
2. Pick the lead method using the matrix at the top of this article.
3. Gather the inputs — revenue forecast, royalty-rate evidence, R&D ledger, contributory asset estimates, discount rate.
4. Run the lead method, then the cross-check, then the floor.
5. Reconcile the three results and document the gap.
6. Apply the purpose-specific lens — auditor, lender, acquirer, HMRC — and adjust presentation accordingly.

The Opagio Intangible Asset Valuator automates steps 3-5 for the relief-from-royalty and MPEEM methods, with royalty-rate evidence prepopulated by sector and discount rates calibrated to UK SME norms. The output is a working-paper-grade reconciliation that maps directly onto the lender, auditor, or HMRC working papers you are about to be asked for.

Related articles. Brand valuation formula: Relief from Royalty in practice applies the same RFR mechanic to brand assets — useful for IP portfolios that combine patents with trademarks. R&D capitalisation under IFRS (IAS 38) covers the development-cost capitalisation rules that determine what's already on the balance sheet before a patent valuation runs.

Further reading. For the underlying mechanic of the workhorse method see Relief from Royalty Method and Multi-Period Excess Earnings Method. For the broader frame on patents as a class see Patents. For the discounted cash flow mechanic shared by both income methods see DCF. For the lending application see IP-backed lending.

Next step. Run the Opagio Intangible Asset Valuator — pick the patent class, supply the revenue forecast, and receive a triangulated RFR + MPEEM result with the cost floor and the lender haircut already modelled.

Ivan Gowan is Founder & CEO of Opagio. Twenty-five years in financial technology, ex-IG Group. Opagio builds the intangible asset evidence platform that institutional investors, lenders, and acquirers expect. About the team →