Proving and Attacking Patent Damages: A Practical Guide

Summary. The sequence that produces a damages number a court will let a jury hear.

The single most important thing to understand

A patent damages case is won or lost on methodology, not on the size of the number.

35 U.S.C. § 284 guarantees a patentee at least a reasonable royalty, so a plaintiff who proves infringement will get something. What determines whether that something is $80,000 or $80 million is whether the expert's analysis rests on a defensible base, a rate tied to case-specific evidence, and an apportionment that isolates the patented contribution. Federal Rule of Evidence 702 — as amended in 2023 to require the proponent to establish by a preponderance that the opinion reflects a reliable application of the methodology — is the gate, and it is the most consequential motion in most patent cases after claim construction.

So the work described below is not preparation for the damages case. It is the damages case.

PART ONE: BEFORE FILING

Step 1 — Value the case before you file it

For the patentee, this analysis determines whether the case is worth bringing. For the defendant receiving a demand letter, the same analysis determines what the threat is actually worth.

Identify the accused revenue. Products, units, and revenue over the period. Public sources will get you within an order of magnitude: financial statements, product pages, teardown reports, industry analyst data, and the defendant's own marketing.

Apply the two hard cutoffs before anything else.

• The six-year window. 35 U.S.C. § 286 bars recovery for infringement more than six years before the complaint. Draw the line and delete everything before it.
• Marking. Under 35 U.S.C. § 287, a patentee who sells products practicing an apparatus claim and fails to mark them recovers only from actual notice. Audit your own products and your licensees' before filing. This single issue eliminates more damages than any other, and a patentee who discovers a marking failure during expert discovery has lost years of exposure it was counting on.

Choose asserted claims with damages in mind. Method claims carry no marking obligation. Apparatus claims do. A portfolio with both gives the patentee a choice, and the choice has multi-year damages consequences.

Build the apportionment story now, not later. What is the smallest salable unit that practices the invention? Is there any credible entire-market-value theory? What does the invention add over the closest prior art? If you cannot answer these before filing, you do not know what the case is worth.

Inventory the comparables. Every license to the patent or its family, every license the defendant has taken in the field, every industry rate you can substantiate. Comparable licenses are the most persuasive royalty evidence there is. Their absence is the most common reason a damages case is weak.

Assess lost profits availability. Are you in the market? Do acceptable non-infringing substitutes exist? If substitutes exist, this is a royalty case and it is worth a fraction of what a lost profits case is worth.

Assess willfulness. Pre-suit notice, copying evidence, internal documents. A trebling multiplier under Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016) transforms settlement posture.

Assess fee exposure both directions. 35 U.S.C. § 285 and Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014) make a weak case expensive to bring.

Then discount for the gauntlet. Multiply the damages model by the probability of surviving claim construction, summary judgment, an inter partes review under 35 U.S.C. § 311, trial, and appeal. This is the number.

Step 2 — Send a notice letter that actually gives notice

If marking is imperfect, actual notice is the fallback, and it must be specific: identify the patent by number, identify the accused product with particularity, and state that you contend it infringes. A letter offering a "license to our portfolio" does not start the damages clock.

Two cautions. A notice letter can create declaratory judgment jurisdiction under 28 U.S.C. § 2201 and the framework of MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007), handing the recipient the choice of forum. And a letter that overstates the case is exhibit A in an exceptional-case motion. Draft it as though it will be read aloud at trial, because it will be.

PART TWO: DISCOVERY

Step 3 — Get the financial data in a usable form

This is the step that most often goes wrong, and it goes wrong in a boring way: the data arrives as tens of thousands of PDF pages, and rebuilding it costs six figures.

Negotiate format in the ESI protocol. Transactional sales data should be produced natively — as database exports or structured spreadsheets with field definitions — not as images. Get this in writing at the Rule 26(f) conference under Federal Rule of Civil Procedure 26, not later.

Ask for the right granularity. Unit sales and revenue by product, by month, by customer segment for the full damages period. Aggregated annual totals are useless for apportionment and for market share analysis.

Ask for cost data that supports a margin calculation. Standard cost build-ups, bills of material, and the accounting policies that define which costs are variable. Incremental margin, not gross margin, drives lost profits.

Do not forget the qualitative documents. Marketing materials that tout the accused feature, win/loss analyses, customer requests, and internal product roadmaps are the evidence that makes or breaks apportionment. Financial data alone tells you the size of the pie; these documents tell you how much of it the invention baked.

Step 4 — Get the licenses

From the defendant: every license it has taken or granted involving comparable technology. Defendants often hold licenses at rates far below what the patentee will propose, and those are the most effective damages evidence available.

From the patentee: every license to the asserted patent or its family, every negotiation that did not result in a license, every valuation of the patent for any purpose (acquisition, financing, impairment testing, tax), and the full litigation and settlement history.

Handle third-party confidentiality early. Most licenses contain notice obligations. Identify them, send the notices, and get the protective order in place before the production deadline, not after.

Distinguish settlement licenses. A license entered to end litigation reflects the cost of defense as much as the value of the technology. Courts admit them with caution and sometimes exclude them. Both sides should know which category each license falls into before building an opinion on it.

Step 5 — Build the non-infringing alternatives record (defense)

This is the Panduit element that defeats lost profits, and the evidence decays. Engineers leave. Design files get archived. Competitor products get discontinued.

Collect: competitor products available during the damages period and their technical specifications; internal design-around studies, including projects considered and abandoned; prior art embodiments that were commercially available; customer testimony about what they would have purchased; and evidence about whether the patented feature appeared in procurement specifications.

The standard is acceptable substitutes, not equivalent ones. A cheaper product lacking the patented feature is an acceptable substitute for the customers who did not care about the feature. Identify those customers by name.

Step 6 — Build the willfulness record (both sides)

For the patentee: documents showing the defendant knew of the patent, evidence of copying (design documents referencing the patentee's product, teardowns, hiring of the patentee's engineers), and the defendant's response to notice.

For the defendant: the good-faith story. Independent development records, prior art the engineers were working from, and — if you choose to rely on it — an opinion of counsel. Understand the trade: relying on the opinion waives privilege over the subject matter, potentially including communications with trial counsel depending on the court. Under 35 U.S.C. § 298, the absence of an opinion cannot be used to prove willfulness or inducement, so the decision is purely about whether the affirmative evidence is worth the waiver.

PART THREE: THE EXPERT

Step 7 — Select the expert

Damages experts fall into recognizable types, and the fit matters.

The economist. Strong on hypothetical negotiation theory, elasticity, and market definition. Excellent when the case turns on price erosion or market share. Sometimes weak on the accounting details of margin computation.

The accountant or valuation professional. Strong on financial data, margins, and license comparability. Sometimes weak on the economic theory an appellate court will scrutinize.

The industry expert. Not usually the damages expert, but essential as a technical companion — the person who explains what the smallest salable unit is and why, and whether the patented feature drives demand.

Selection criteria that matter: prior Daubert history (search it, and read the opinions); ability to explain a regression or a conjoint study to a jury without condescension; willingness to say no to a theory that will not survive; and availability for the deposition and trial dates.

Selection criteria that do not matter as much as people think: credentials alone, and prior work for the client. A well-credentialed expert whose methodology has been excluded three times is a liability.

Step 8 — Manage the apportionment analysis

This is the heart of the report and the target of the motion.

Establish the royalty base. Begin with the smallest salable patent-practicing unit. If the expert wants to use a larger base, the entire market value rule requires proof that the patented feature drives customer demand for the entire product — a demanding standard that survey evidence alone rarely satisfies.

Apportion within the base. Even the smallest salable unit usually contains unpatented features. The expert must explain what fraction of that unit's value the invention contributes, and the explanation must rest on evidence: technical testimony about the contribution over the prior art, customer evidence about what they valued, or a properly conducted conjoint or survey analysis.

Derive the rate from the evidence. Comparable licenses, adjusted for differences in scope, technology, timing, and structure — with the adjustments explained. Rules of thumb are inadmissible. A rate that appears without derivation is an invitation to exclusion.

Do the Georgia-Pacific analysis selectively. Marching through all fifteen factors mechanically signals that the expert has no theory. Address the factors that matter, explain why the others do not, and make factor 13 — the portion of profit attributable to the invention — the analytical center.

Check the date. The hypothetical negotiation occurs when infringement began. Using a later date, when the product was more successful, is a common and fatal error.

Sanity-check against the infringer's margin. A royalty exceeding the infringer's profit on the product invites the argument that no rational licensee would have agreed. Have an answer.

Step 9 — Depose the opposing expert

The deposition is where the Daubert motion is built. Do not treat it as discovery of the opinion; treat it as construction of the record for exclusion.

Lines that produce useful admissions:

• Walk the arithmetic. Ask the expert to compute the result from the inputs, on the record. Errors surface.
• Test each adjustment to a comparable license. "What is the basis for adjusting the rate upward by 40%?" repeated for each adjustment.
• Probe the apportionment fraction. "How did you arrive at 45%? What would 40% have produced? What evidence distinguishes 45% from 40%?"
• Establish what the expert did not consider. Non-infringing alternatives, the defendant's other licenses, the prior art.
• Establish the boundaries of expertise. A damages expert who relies on a technical expert for the smallest salable unit cannot defend that choice independently.
• Ask about the entire market value rule directly. "Do you contend the patented feature drives demand for the entire product? What is your evidence?"

Keep the questions open where you want explanation and closed where you want admissions. Resist the urge to argue; the deposition is for building, not winning.

Step 10 — File the Rule 702 motion, or defend it

As movant, organize around the specific defect, not around general unreliability. Courts respond to "the base violates apportionment because X," "the rate is untethered to any comparable because Y," and "the negotiation date is wrong because Z." They do not respond well to a scattershot attack on credibility, which is cross-examination material.

Ask for the right relief. Full exclusion is appropriate where the methodology is fundamentally unsound. Partial exclusion — barring the entire market value theory while allowing the apportioned analysis — is more commonly granted and is often enough.

As respondent, the strongest defense is that the criticisms go to weight rather than admissibility, and that argument works best where the expert's method is orthodox and the disputes are about inputs. It works poorly where the method itself is the problem. If your expert's report has a genuine methodological gap, the time to fix it is before the motion, by supplementation if permitted, not in opposition briefing.

PART FOUR: TRIAL AND AFTER

Step 11 — Present damages to a jury

Lead with the story, not the spreadsheet. Why is this technology valuable? What did the defendant get from it? Juries follow narratives and tolerate numbers.

Show the apportionment visually. A diagram of the product, the smallest salable unit highlighted inside it, and the patented contribution highlighted inside that, does more work than twenty minutes of testimony.

Anticipate the anchor problem. If the whole-product revenue is coming into evidence for another purpose, address it directly rather than hoping the jury forgets it.

Prepare the expert for the "what if you're wrong" question. Damages experts should have defensible answers at multiple assumptions rather than a single point estimate delivered with false precision.

Keep willfulness clean. If willfulness is being tried, the evidence should be about the defendant's state of mind, not a second infringement argument.

Step 12 — Post-trial

• Move for or against enhancement. Willfulness is the jury's; enhancement is the judge's, applying the Read factors after Halo.
• Move for or against fees under 35 U.S.C. § 285.
• Address prejudgment interest. It is the norm under General Motors Corp. v. Devex Corp., 461 U.S. 648 (1983); the fight is about rate and compounding.
• Address the injunction under 35 U.S.C. § 283 and the four-factor test of eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006).
• If the injunction is denied, negotiate or brief an ongoing royalty, which is typically set above the jury's rate.
• Preserve everything. Renew Rule 50 motions on damages; a damages verdict unsupported by substantial evidence is one of the more reversible things in patent law.

A worked example, from both sides

The case. Thistlewood Robotics, Inc. v. Caldera Automation LLC, District of Minnesota. Thistlewood asserts a patent on a force-feedback control loop used in robotic palletizing arms. Caldera's arms allegedly infringe. The arms sell for $118,000 each; Caldera sold 610 of them over the relevant period. The control board that implements the accused loop costs $2,400 as a spare part.

Thistlewood's team: lead counsel Adaeze Kwarteng; damages expert Dr. Samuel Feiner, an economist; technical expert Dr. Lena Vasilyeva. Caldera's team: lead counsel Roscoe Tam; damages expert Marguerite Ilunga, CPA/ABV.

Month 1 — the pre-filing audit that changes the case

Adaeze runs the marking audit before filing. Thistlewood sells its own palletizing arms and has licensed the patent to one small manufacturer, Kestrel Handling. Thistlewood marks its arms. Kestrel does not — its products carry no patent notice at all, and the license contains a marking clause nobody has ever enforced.

This matters enormously. Under 35 U.S.C. § 287, a licensee's failure to mark can defeat constructive notice for the whole patent. Thistlewood's pre-suit damages are therefore at risk unless it can show actual notice or bring Kestrel into compliance.

Adaeze does two things. She has Thistlewood enforce the marking clause immediately, so that constructive notice attaches from the date Kestrel begins marking. And she reviews the correspondence file, where she finds a 2022 letter from Thistlewood's VP of engineering to Caldera identifying the patent by number and the accused arm by model. That letter is now the most valuable document in the case: it establishes actual notice and it is the foundation of the willfulness claim.

Damages start date: the 2022 letter, not the 2019 first sale. Roughly 40% of the accused units fall outside the recoverable period. Adaeze tells the client this before filing rather than after.

Months 3–6 — the discovery fight that decides apportionment

Adaeze serves discovery for monthly unit and revenue data by product and customer, cost build-ups, all licenses, and the marketing and win/loss files.

Roscoe produces the financial data — as 41,000 pages of PDF. Adaeze moves to compel native production, citing the ESI protocol she insisted on at the Rule 26(f) conference under Federal Rule of Civil Procedure 26. The court orders native production. This saves an estimated $180,000 in expert time and, more importantly, makes a customer-level analysis possible.

The customer-level data turns out to matter. It shows that 96 of the 610 arms went to three food-processing customers whose RFPs specified force-feedback capability by name. Those 96 units become the core of the lost profits case.

Meanwhile Roscoe collects the non-infringing alternatives record. He finds that two competitors, Orenda Systems and Bexley Handling, sold palletizing arms throughout the period using open-loop position control that does not infringe, at prices 8% below Caldera's. He deposes purchasing managers at four Caldera customers; three testify they would have bought the Orenda arm if Caldera's had been unavailable.

Months 7–9 — the licenses

Roscoe's most valuable discovery is not from Caldera. It is from Thistlewood. The Kestrel license carries a 1.8% royalty on the arm price. Thistlewood also produced a 2021 board deck valuing the patent portfolio at $4.2 million for a financing round.

Adaeze knew both documents were coming and has an answer: Kestrel is a tiny manufacturer that took the license as part of a settlement of a separate dispute, and the board deck valued twelve patents, not one. But she also knows that a 1.8% number in front of a jury is gravity, and it shapes her settlement analysis from that point forward.

Months 10–12 — the reports

Dr. Feiner's opening report offers three theories:

1. Lost profits on 96 units where the RFP specified the capability, at $34,200 incremental margin: $3.28 million.
2. Market share lost profits on a portion of the remaining units. Thistlewood holds 22% of the domestic market. Applied to the 274 remaining in-period units: about 60 units, another $2.05 million.
3. Reasonable royalty on the balance. Here Dr. Feiner initially proposes 2.4% of the $118,000 arm price — $2,832 per unit — reasoning upward from the Kestrel license because Caldera is a larger, better-capitalized licensee.

Ms. Ilunga's rebuttal attacks all three. On lost profits she relies on the substitutes testimony. On market share she argues Thistlewood's capacity was constrained. On the royalty she makes the argument that matters: the base violates apportionment. The accused control loop is implemented on a $2,400 board inside a $118,000 machine with a frame, servos, an end effector, a vision system, and safety controls, none of which is patented. Nothing in the record shows the force-feedback loop drives demand for the arm.

Month 13 — the Rule 702 motion

Roscoe moves to exclude Dr. Feiner's royalty opinion under Federal Rule of Evidence 702.

The court grants the motion in part. It holds that the entire market value rule is not satisfied: evidence that three customers specified the capability shows it was valued, not that it drove demand for the machine as a whole. Dr. Feiner may not use the $118,000 arm price as the royalty base.

The court denies the motion as to lost profits, holding that the substitutes dispute goes to weight — a genuine factual question about whether Orenda's open-loop arms were acceptable to these customers.

Dr. Feiner's supplemental opinion, permitted because trial is still five months out, starts from the $2,400 board, apportions 60% of the board's value to the patented loop based on Dr. Vasilyeva's testimony about the contribution over prior art, and applies an 11% rate derived from the Kestrel license adjusted for the narrower base. Result: about $158 per unit — $34,600 across 219 royalty units.

The royalty component has gone from $620,000 to $34,600. The case is now a lost profits case.

Month 18 — trial and after

The jury awards $2.9 million in lost profits (crediting Caldera's substitutes evidence in part), $34,600 in royalty, and finds willfulness based on the 2022 letter and an internal Caldera email reading "Thistlewood's patent is a problem but they've never sued anyone."

The court enhances by 1.75× under Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016), applying the Read factors: deliberate continuation after notice weighs heavily; the closeness of the infringement question and Caldera's professional litigation conduct weigh against trebling. It awards prejudgment interest at the prime rate, compounded annually, following General Motors Corp. v. Devex Corp., 461 U.S. 648 (1983).

It denies a permanent injunction under eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) — Thistlewood had licensed the patent to Kestrel, undercutting its irreparable-harm showing — and sets an ongoing royalty at 15%, above the jury's rate, reflecting adjudicated validity and infringement.

What decided the case. Not the trial. The marking audit in month one, the motion to compel native data in month four, and the apportionment ruling in month thirteen.

Dividing the work

The financial data owner. Someone must own the sales and cost data end to end: what was requested, what was produced, in what format, with what gaps. In a large case this is a dedicated associate plus a consulting analyst. The failure mode is diffuse ownership, where four people each assume someone else validated the numbers.

The license owner. Every license in the case — the patentee's, the defendant's, and the industry's — with a comparability memo for each. This person should be able to state, for any license in the record, why it does or does not inform the hypothetical negotiation.

The technical bridge. The person who connects the technical expert's contribution-over-prior-art analysis to the damages expert's apportionment fraction. This link is where reports most often fail: the damages expert asserts 45% and the technical expert never supplied a basis for it.

The willfulness owner. The knowledge file, the notice correspondence, the internal documents, and the privilege decisions around opinions of counsel.

Lead counsel owns the number. Specifically: what the case is worth at each stage, what it will cost to get there, and what the settlement recommendation is. This number should be written down and revised at every milestone, not carried around in someone's head.

Valuing the case at each stage

Settlement discussions go badly when the parties are valuing different cases. A disciplined approach re-values at each milestone using the same framework.

The practical instruction: never let a settlement conversation happen without a current written valuation. The most common cause of failed mediations in patent cases is that one side is valuing the case as of the complaint and the other as of the claim construction order.

What to do when the damages evidence is thin

Sometimes there is no comparable license, no lost profits story, and no clean smallest salable unit. This happens more often than treatises suggest. Options, roughly in order of preference:

Build a comparable from adjacent transactions. Licenses to related technology, licenses by the same parties in different fields, industry royalty surveys with a documented methodology. Each requires an explained adjustment, but a chain of reasoning beats an assertion.

Use the cost of the next-best alternative. What would the defendant have spent to design around? A design-around cost analysis, supported by engineering testimony about time and expense, is a legitimate ceiling on what a willing licensee would have paid, and courts accept it.

Use conjoint or survey evidence to isolate the feature's value to purchasers. These are expensive and heavily attacked, but a properly designed study is admissible and can supply the apportionment fraction that nothing else does.

Use the infringer's own internal valuations. Business cases, product justifications, and ROI analyses prepared before litigation are powerful because they were not made for litigation.

Accept a small number. A defensible small award is worth more than a large one that gets excluded or reversed. Clients rarely want to hear this, and it is usually right.

A working calendar

Mistakes that recur

Discovering a marking failure after expert reports. Audit before filing.

Producing financial data as PDFs. Negotiate format at the outset.

An expert who uses the whole product as the base without an entire-market-value showing. The most common ground for exclusion.

A rate with no derivation. "In my experience, 5% is typical" is not admissible.

Ignoring the defendant's existing licenses. They are usually the best evidence in the case and they usually favor the defense.

Failing to update the model after claim construction. A narrower construction often shrinks the royalty base or removes accused products.

Treating willfulness as an afterthought. The evidence must be collected in ordinary discovery; it cannot be assembled at the pretrial conference.

Bringing a case whose expected value does not exceed its cost. Run the gauntlet discount honestly.

Working with the damages expert: a management guide

A damages expert is not a witness you hire and then meet again at the deposition. The relationship is a project, and the projects that go well share a rhythm.

The education phase. Before the expert forms any opinion, they need to understand the technology, the patent, the claim constructions, the accused products, and the market. Budget real time for this — a half-day technical tutorial with the technical expert, a walkthrough of the accused products, and a reading list. Experts who skip this write reports that are technically illiterate in ways opposing counsel will make visible.

The data phase. Give the expert the raw data, not your summary of it. Experts who rely on counsel-prepared summaries get destroyed in deposition on the question "did you verify this yourself?" Provide the native files, the field definitions, and access to a company witness who can explain the accounting.

The theory phase. Ask the expert what theories the evidence supports, and listen to the answer. The failure mode here is counsel arriving with a number and asking the expert to justify it. Good experts refuse; the ones who agree produce reports that get excluded.

The stress-test phase. Before the report is served, run an internal Daubert exercise. Have a lawyer who has not worked on the damages case attack the opinion for two hours. Every weakness found internally is a weakness fixed for free.

The deposition phase. Prepare the expert on the arithmetic, on every adjustment, and on the sentences in the report that are least well supported. The most damaging deposition testimony in damages cases is not an admission about the law; it is an inability to reproduce a calculation.

The trial phase. Simplify without distorting. A jury will follow one clear chain of reasoning with three or four numbers in it. It will not follow a regression output.

Reading the other side's report

A disciplined intake protocol for an opposing damages report, done in the first week after service:

1. Reproduce every calculation. Have an analyst rebuild the model from the inputs. Arithmetic errors in damages reports are more common than one would expect, and finding one changes the tenor of the whole engagement.
2. Trace every citation. For each factual assertion, find the underlying document. Assertions that cite deposition testimony are often characterizations rather than quotations.
3. List every assumption. Write them out. Then determine, for each, what evidence supports it and what the result would be if the assumption changed.
4. Identify the base and test it. Is it the smallest salable patent-practicing unit? If larger, what is the entire-market-value showing?
5. Identify the apportionment fraction and its source. Is it derived from evidence, or asserted?
6. Check the negotiation date. Compare it to the first accused sale and to the notice date.
7. Check the comparables. For each license relied on: what are the differences, and did the expert adjust for them with an explanation?
8. Check for double counting. Lost profits and royalty on the same units is a recurring error.
9. Check the damages period against 35 U.S.C. § 286 and the marking analysis.
10. Write the exclusion memo. Two pages: the defects, ranked, with the relief you would seek for each.

This protocol takes about a week of analyst and associate time and it is the highest-return week in the damages phase.

Frequently asked questions

Can a patentee recover damages for products it never accused? No. Damages run to accused products, and a patentee that discovers additional infringing models late must amend its contentions — which courts permit or refuse based on diligence and prejudice. Audit the defendant's full product line early.

What if the accused product is sold to a customer who then infringes? Direct infringement by the customer, and induced or contributory infringement by the supplier, are separate claims with separate damages. Induced infringement requires knowledge under Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754 (2011), and a good-faith belief in invalidity is no defense after Commil USA, LLC v. Cisco Systems, Inc., 575 U.S. 632 (2015). Suing the supplier is usually more efficient, but the damages base is the supplier's sales, not the customers' downstream revenue.

How early should a damages expert be retained? Earlier than most cases do it. An expert who participates in framing discovery requests gets the data they need; an expert retained after the close of fact discovery works with whatever happens to have been produced.

Can a plaintiff amend a damages theory after a Rule 702 exclusion? Sometimes. Courts occasionally permit a supplemental report, particularly where the defect is curable and no trial date is imperiled. Do not count on it.

Are settlement licenses admissible? Sometimes, with caution. Courts weigh whether the license reflects technology value or litigation avoidance. Expect a fight either way.

What if the defendant has no profits? A reasonable royalty is still owed — § 284 sets a floor. But the absence of profit is powerful evidence about what a willing licensee would have paid.

Does an IPR affect damages? Substantially. Claims cancelled in an inter partes review yield no damages, and surviving claims are often narrower, which changes apportionment. 35 U.S.C. § 315 governs the interaction.

How are damages handled for products sold abroad? Generally not recoverable, with the important exception recognized in WesternGeco LLC v. ION Geophysical Corp., 585 U.S. 407 (2018) for foreign lost profits flowing from domestic infringement under § 271(f).

Should a defendant put on its own damages expert? Almost always. A defendant who attacks the plaintiff's expert without offering an alternative leaves the jury with one number to work from.

Related documents

• Patent Damages: Reasonable Royalties, Lost Profits, Apportionment, and Enhancement
• Patent Damages Checklist: A Practical Checklist
• Patent Damages Toolkit: Royalty Models, Expert Reports, and Daubert Motions
• Preparing for a Markman Hearing: A Practical Guide
• How to License Your Patent: From Valuation to Term Sheet
• Expert Witnesses After the 2023 Amendment to Rule 702
• Proving Damages in Civil Litigation: Certainty, Causation, and the Numbers That Survive
• Patent Marking Compliance: A Practical Checklist
• Discovery Toolkit: A Roadmap and Resource Guide for Federal Discovery