Patent Infringement Contention Checklist: Element-by-Element Charting, Equivalents, and Indirect Theories

By Casey Scott McKay · 2025-04-09

Contentions are the most consequential documents in a patent case after the claims themselves, because they bind and they are served before the evidence exists. This checklist runs them in twelve phases: verify standing and the cheap complete defenses, identify the contested element, build the evidence plan, chart element by element with real citations, plead equivalents only after the file wrapper, test every method claim for divided performance, address indirect theories and the territorial provisions, run the marking analysis, then amend, supplement, and carry the contentions into expert reports and summary judgment. Each box gives the reason, the authority, and the trap. It covers both chairs, because non-infringement contentions follow the same discipline in reverse. A worked matter runs throughout.

IP and Technology > Patent Litigation | Checklist | Published 9 April 2025 - Updated 10 October 2025 | Casey Scott McKay - marksy.us

Summary. Contentions are the most consequential documents in a patent case after the claims themselves, because they bind and they are served before the evidence exists. This checklist runs them in twelve phases: verify standing and the cheap complete defenses, identify the contested element, build the evidence plan, chart element by element with real citations, plead equivalents only after the file wrapper, test every method claim for divided performance, address indirect theories and the territorial provisions, run the marking analysis, then amend, supplement, and carry the contentions into expert reports and summary judgment. Each box gives the reason, the authority, and the trap. It covers both chairs, because non-infringement contentions follow the same discipline in reverse. A worked matter runs throughout.

Keywords: infringement contentions, claim chart, local patent rules, element by element, evidence citation, doctrine of equivalents, prosecution history estoppel, induced infringement, contributory infringement, divided infringement, source code review, product testing, marking analysis, standing, exhaustion, amendment good cause, expert report, summary judgment, accused instrumentality, non-infringement contentions

How to use this checklist

Infringement contentions are served early, on a schedule set by local rules, and long before the discovery that would actually establish infringement. That timing is the entire problem. A patentee must commit to a theory using public information, and then live inside that commitment for two years. A defendant must respond to a theory before it has read the plaintiff's evidence. Both sides are asked to be precise at the moment they know least.

The discipline that follows is the one that survives that problem: identify the element that will actually decide the case, chart it first, cite evidence rather than assertion, and keep the theory narrow enough that it does not need amendment and specific enough that it does not need supplementation. Everything else in this checklist supports those four habits.

| Phase | What it covers | When | |---|---|---| | 1 | Standing and the cheap complete defenses | Week one | | 2 | Identify the contested element | Week one | | 3 | Build the evidence plan | Weeks one to two | | 4 | Identify accused instrumentalities precisely | Before service | | 5 | Chart element by element | Before service | | 6 | Equivalents, after the file wrapper | Before service | | 7 | Method claims and divided performance | Before service | | 8 | Indirect and territorial theories | Before service | | 9 | Marking and the damages period | Month one | | 10 | Serve, then supplement | Per schedule | | 11 | Amend on diligence | As facts emerge | | 12 | Carry into experts and motions | Later |

Contentions bind. In local-rule districts they are amended only on a showing of good cause, and a theory not disclosed is generally unavailable. Boxes marked [Gate] should clear before service.

The matter. A seven-limitation claim asserted against a network appliance. Six limitations were plainly met. The seventh recited allocation of a buffer "prior to receipt of the first data packet," and nobody had asked when allocation actually happens. The boxes below follow the case from that question.

Phase 1. Standing and the cheap complete defenses

Nothing in this phase requires an expert, a code review, or a single deposition. It requires an afternoon with the assignment records, the client's own contract files, and the patent's front page. It is skipped constantly, and it is where cases end cheaply when they end cheaply at all.

[ ] Confirm the plaintiff owns the patent, with a complete recorded chain.
- Why. A patent is personal property that moves by written assignment, and every link has to exist.
- Authority. 35 U.S.C. § 261.
- Trap. An unrecorded assignment or a gap from an acquisition three owners back. A defect here is a complete answer and it costs an afternoon.
[ ] Confirm whether an exclusive licensee must join the patentee.
- Why. Only a party holding all substantial rights can sue alone; anyone else needs the patentee.
- Trap. A plaintiff holding less than all substantial rights suing alone.
[ ] Check the term.
- Why. Expired, lapsed for maintenance fees, or shortened by a terminal disclaimer.
- Authority. 35 U.S.C. § 154(a); 35 U.S.C. § 41(b); 37 C.F.R. § 1.321.
- Trap. Term adjustment and terminal disclaimers pulling in opposite directions, so the front page alone does not answer the question.
[ ] [Gate] Read the client's own license portfolio.
- Why. Rights may exist through a corporate parent, an acquisition, a supplier's license with pass-through rights, or a portfolio agreement. The cheapest defense available and routinely skipped.
- Trap. Nobody in the legal department knowing what the company bought in a transaction five years ago.
[ ] Check exhaustion.
- Why. An authorized sale exhausts rights in that article. A complete defense for a reseller or refurbisher.
- Trap. Assuming a contractual restriction on resale preserves patent rights it does not preserve.
[ ] Check implied license, particularly where the patentee sells a component with no substantial non-infringing use.
[ ] Check whether the patent was declared essential to a standard.
- Why. A licensing commitment changes both injunction risk and the royalty range.
- Authority. See Standard-Essential Patents and FRAND Toolkit.
[ ] Check the assignor-estoppel posture where the plaintiff's inventor now works for the defendant.
- Why. An assignor who represented that a claim was valid may be barred from contesting validity, and the doctrine reaches privies. It is a trap that surfaces late and changes the defense strategy entirely.

Phase 2. Identify the contested element

[ ] [Gate] Rank every element of every asserted independent claim by confidence.
- Why. The case turns on the least confident one, and charting the obvious elements first feels like progress while hiding the problem.
- Trap. Six limitations documented in detail and the seventh marked "met."
[ ] Chart the least confident element first, in full.
- Why. If it cannot be charted with evidence, the remaining six charts are wasted work, and knowing that in week one is worth more than knowing it in month twelve.
[ ] Ask an engineer, not a document.
- Why. The person who wrote the routine knows when the buffer is allocated; the datasheet does not address it.
- Trap. A datasheet that does not address the contested behavior at all.
[ ] Chart independent claims only at this stage.
- Why. A product avoiding the independent claim avoids everything depending from it.
[ ] Identify which limitation the examiner treated as the point of novelty.
- Why. The element the applicant fought for is usually the element the accused product was designed around, and the file wrapper says which one that was.
[ ] For the defendant, run the same exercise for absence.
- Why. One element disproved rigorously ends the case; seven contested weakly reads as having nothing.

Phase 3. Build the evidence plan

[ ] Work backward from the contested element to the evidence that would prove it.
- Why. The question is not what discovery is available but what specific document, file, or test result would answer this specific question.
[ ] Rank evidence by strength.
- Why. Source code and internal engineering documents are strong. Design records and specifications are strong. Testing with a documented methodology is medium. Marketing materials are weak and wrong in both directions.
[ ] [Gate] Sequence: documents, then testing, then code, then depositions, then the expert.
- Why. Each stage tells you what to ask for in the next one, and code review without a document review first is an expensive way to read unfamiliar software.
- Trap. Engaging the expert before the evidence exists, which produces an opinion revised later and a client billed twice.
[ ] Identify what may be unobtainable.
- Why. An internal, unobservable, undocumented limitation may be unprovable, and that belongs in the month-two assessment rather than the month-twelve one.
[ ] Budget the three large items: code review, testing, and expert work.
- Why. These dominate the litigation budget, and a case whose contested element requires all three is a different case economically from one that turns on a public specification.

Phase 4. Identify accused instrumentalities precisely

[ ] Name each accused product, model, version, and configuration.
- Trap. "The accused products" collectively, which conceals variation on exactly the contested element.
[ ] Cover the whole damages period, not only the current version.
- Why. The implementation may have changed, and the earlier period is otherwise unproven.
- Trap. Charting the shipping version and assuming three prior releases behaved the same way.
[ ] Address configurable features.
- Why. Where a claimed feature is present but disabled by default, infringement may turn on whether customers enable it, which is an evidentiary question about usage.
[ ] Address services separately from products.
- Why. A hosted service cannot be purchased and tested, and the case depends entirely on discovery.
[ ] Identify components sold separately.
- Why. Kits and components usually implicate indirect rather than direct infringement.
[ ] Identify third-party components inside the accused product.
- Why. A supplier's licensed component may carry rights the defendant did not know it had, and it may shift the real dispute to an indemnity obligation.

Phase 5. Chart element by element

[ ] Break the claim at every limitation, not every comma.
- Trap. Grouping three requirements into one row, which hides which is contested.
[ ] Use three columns minimum: claim element, corresponding structure or step, evidence citation.
- Why. The middle column forces identification of what in the product corresponds to the element, which is the analysis; without it a chart restates the claim next to a product name.
[ ] [Gate] Cite evidence with specificity.
- Why. Document and page, source file and line, test result with methodology, or deposition page and line.
- Trap. "The Accused Product performs this step," which is an assertion and is routinely struck.
[ ] Address every element, including the obvious ones.
- Trap. A chart skipping elements invites an argument that they were not proven.
[ ] Chart the preamble where it may be limiting.
- Why. A preamble that gives life and meaning to the claim, or that supplies antecedent basis for a later term, is a limitation and has to be met like any other.
[ ] Chart the claim as construed once a construction exists, and re-run earlier charts.
- Authority. See Claim Construction Checklist.
[ ] For means-plus-function limitations, chart the disclosed structure and its equivalents.
- Why. For a computer-implemented function the structure is the algorithm, not the processor.
- Authority. 35 U.S.C. § 112(f).
- Trap. Charting the functional language as though it were the limitation.
[ ] Chart per product and per asserted claim.
- Why. One chart covering a family of products is the fastest route to a motion to strike.

Phase 6. Equivalents, after the file wrapper

[ ] [Gate] Read the prosecution history before pleading equivalents at all.
- Why. Estoppel defeats most theories and the file wrapper is public.
- Trap. Committing to a theory the defendant dismantles with a document.
[ ] Identify every narrowing amendment and the reason for it.
- Why. A narrowing amendment for patentability surrenders the territory between the original and amended claim, and a presumption of surrender attaches.
[ ] Determine whether the accused equivalent lies in the surrendered territory.
- Why. If it does, the analysis is essentially over regardless of how similar the accused element looks.
[ ] Assess the rebuttals honestly.
- Why. Unforeseeability at the time of amendment, a rationale bearing only a tangential relation, or another reason the patentee could not reasonably have described the equivalent. The theories that survive usually involve after-arising technology.
[ ] Check argument-based estoppel separately.
- Why. A clear surrender in argument limits equivalents without any amendment.
[ ] Check the disclosure-dedication rule.
- Why. Subject matter described in the specification and not claimed is dedicated to the public.
[ ] Check the prior art ceiling.
- Why. An equivalent cannot be broad enough to encompass the prior art.
[ ] Check claim vitiation.
- Why. An equivalents theory that reads a limitation out of the claim entirely fails as a matter of law, and the usual sign is a theory that treats an express structural requirement as optional.
[ ] If pleaded, chart function, way, and result for the accused element.
- Trap. Equivalents pleaded as a placeholder, which a defendant will move to strike.

Phase 7. Method claims and divided performance

[ ] [Gate] Identify who performs each step of every asserted method claim.
- Why. Direct infringement requires all steps performed by or attributable to a single entity.
- Trap. A claim reciting steps by a server operator and an end user, asserted as though one party performed both.
[ ] Where more than one entity performs, plead the attribution facts.
- Why. Direction or control, or a joint enterprise. Conditioning participation or a benefit on performance of a step and establishing the manner or timing can support attribution.
[ ] Assert a system claim instead where one exists.
- Why. A system claim can be infringed by making or using the system, which avoids the actor problem entirely.
[ ] Determine whether the claim requires the recited order of steps.
- Trap. Assuming a listed sequence is required, or assuming it is not.
[ ] For the defendant, raise divided performance early.
- Why. Frequently the strongest available argument and available at the pleading stage.

Phase 8. Indirect and territorial theories

[ ] Plead the underlying direct infringement for any indirect theory.
- Why. There is no indirect liability without someone, somewhere, directly infringing.
- Authority. 35 U.S.C. § 271(b); 35 U.S.C. § 271(c).
[ ] Establish the knowledge date for inducement.
- Why. Knowledge of the patent and knowledge that the induced acts infringe. Willful blindness suffices; negligence does not.
- Trap. Failing to note that pre-notice conduct is generally outside indirect liability, which splits the damages period.
[ ] Identify the inducement evidence.
- Why. Instructions, manuals, marketing, technical support communications, and the demand letter itself.
[ ] For contributory infringement, address the staple-article defense.
- Why. A component with substantial non-infringing uses cannot support it, and the defense turns on actual usage evidence.
[ ] Run the territorial analysis on every accused act.
- Authority. 35 U.S.C. § 271(a); components supplied abroad under 35 U.S.C. § 271(f); products made abroad by a patented process under 35 U.S.C. § 271(g).
- Trap. Manufacture abroad assumed to be outside reach, or export of components assumed to be safe.
[ ] Check the regulatory safe harbor where applicable.
- Authority. 35 U.S.C. § 271(e).

Phase 9. Marking and the damages period

[ ] [Gate] Run the marking analysis in month one.
- Why. A patentee making or selling a patented article must mark, or damages accrue only from actual notice.
- Authority. 35 U.S.C. § 287(a).
- Trap. Discovering it at expert reports, when it has already eliminated years of damages.
[ ] Check licensee marking, not only the patentee's.
- Trap. An unmarked licensee product limiting the patentee's recovery.
[ ] Note that method-only assertions generally escape the marking requirement.
- Why. A real strategic consideration in claim selection.
[ ] Establish the actual notice date with specificity.
- Why. A communication identifying the patent and the accused product. A general suggestion that a license may be needed is not notice.
[ ] Apply the six-year limitation.
- Authority. 35 U.S.C. § 286.
[ ] Check provisional rights.
- Authority. 35 U.S.C. § 154(d).
[ ] Track willfulness evidence from the notice date.
- Authority. 35 U.S.C. § 284; note 35 U.S.C. § 298. See Patent Damages Checklist.

Phase 10. Serve, then supplement

[ ] Serve on the local-rule schedule.
- Trap. A rule district treated as an ordinary scheduling order.
[ ] Plead alternatives expressly where the record does not yet permit a choice.
- Why. Literal and equivalent, direct and indirect, with a stated basis for each.
- Trap. Boilerplate reservations, which do not preserve undisclosed theories.
[ ] Supplement promptly as discovery produces the technical documents.
- Why. Contentions served on public information and never updated are the ones that get struck.
[ ] For the defendant, serve invalidity contentions with the same rigor.
- Why. Charts per reference per claim, combinations identified, and the motivation to combine stated. Listing forty references without charting them is not contentions.
[ ] Serve non-infringement contentions where the rules require them, focused on the contested element rather than every element.

Phase 11. Amend on diligence

[ ] Move within weeks of learning a new fact, not months.
- Why. Good cause generally turns on whether the party acted promptly after learning the information.
[ ] Identify what changed and when you learned it.
- Why. A motion that says the theory evolved will lose; one that names a document, a production date, and a deposition answer usually wins.
[ ] Amend after claim construction where the construction changes the theory.
- Why. Courts frequently permit amendment following a construction that neither party proposed.
[ ] Do not rely on a reservation of rights to add a theory.
- Trap. A patentee that pleaded literal infringement and not equivalents, unable to add equivalents after construction goes badly.

Phase 12. Carry the contentions into experts and motions

[ ] Ensure the expert report tracks the contentions.
- Why. An opinion advancing a theory not in the contentions is vulnerable to exclusion.
- Authority. Fed. R. Civ. P. 26; Fed. R. Evid. 702.
[ ] Ensure the expert identifies corresponding structure rather than reciting claim language.
- Trap. The most common weakness in these reports.
[ ] Move for summary judgment on the element you proved best.
- Why. The all-elements rule means the absence of evidence on any single element supports the defendant, and a genuine dispute on any single element defeats the patentee.
- Authority. Fed. R. Civ. P. 56.
[ ] Reduce the case to one element for trial.
- Why. A fact-finder will follow one clear question and not seven.
[ ] Keep the chart, the evidence index, and the estoppel analysis in the trial file.
- Why. These three documents answer most questions that arise at trial, and reassembling them under time pressure is how errors enter the record.

Outcome. The seventh limitation was the case. Source code showed allocation on receipt rather than before it, a design specification confirmed the architecture, and the engineer who wrote the routine testified to it. The defendant moved for summary judgment on that element alone, and did not contest the other six. The motion was granted.

What the schedule actually looks like

Contentions do not exist on their own timeline. In most local-rule districts they sit between the scheduling conference and claim construction, which means the patentee is charting while the defendant is searching prior art, and both sides are doing it before substantial document production. Understanding the surrounding dates prevents the two most common scheduling errors: treating contentions as a formality to be revised later, and deferring the marking and standing work until it can be paired with something else.

The practical shape is this. Infringement contentions land first, on public information. Invalidity contentions follow, usually within a month or two, and they set the prior art universe for the case and often for a parallel proceeding at the Board. Exchange of proposed terms and constructions comes next, and it is driven directly by what the contentions revealed to be contested. The construction briefing and hearing follow, and the order that issues from it either confirms the charts or requires them to be redone. Expert reports come last, and they are bounded by everything that came before.

Every deadline in that sequence is easier when the contested element was identified in week one. A party that knows which limitation decides the case proposes the right terms for construction, takes the right depositions, and asks its expert the right question. A party that has not done that work proposes twelve terms, deposes everyone, and receives an expert report that addresses the wrong dispute at considerable expense.

Phase 13. The defendant's chair, in reverse

Non-infringement contentions are the same discipline run backward, and they are done badly for the opposite reason. A patentee's chart fails by claiming too much; a defendant's response fails by claiming everything.

[ ] Pick the elements you can win and say so.
- Why. A response contesting all seven limitations reads as a party with no argument. A response contesting one, with a document and a code citation behind it, reads as a party that has looked.
- Trap. A boilerplate denial of every element, served because it feels safer, which forfeits the credibility that would have carried the summary judgment motion.
[ ] Identify the construction your non-infringement position requires, and check that you can win it.
- Why. A non-infringement argument that depends on a construction you will lose is not an argument. Sequence the two analyses together rather than treating construction as a separate proceeding.
- Authority. See Litigating Claim Construction.
[ ] Separate "does not literally infringe" from "does not infringe under any theory."
- Why. A defendant that wins literal non-infringement and has not addressed equivalents has won half a case, and the estoppel analysis is usually the other half.
[ ] Preserve the design-around option and price it.
- Why. If the accused feature can be changed for a modest engineering cost, that number caps the settlement value of the case going forward, and it should be known before the first mediation rather than after the expert reports.
[ ] Document the non-infringing configuration and when it shipped.
- Why. A change that predates the complaint shortens the damages period; one that postdates it stops accrual. Either way the date is evidence, and engineering release records establish it better than testimony a year later.

Phase 14. Source code review, in practice

Code review is where infringement cases are usually decided and where the largest share of the technical budget goes. It is also the stage most often run without a plan.

[ ] Negotiate the protective order's source code provisions before the review, not during it.
- Why. Standalone machine, no network, printing limits, page caps, and inspection location are all negotiated terms, and a review conducted under terms nobody read is a review that produces unusable notes.
[ ] Go in with a list of questions, not a list of files.
- Why. The contested element states the question. The reviewer's job is to answer it and record where the answer lives, not to read the codebase.
[ ] Record file path, function name, and line numbers for every relevant excerpt.
- Why. These become the chart citations and the expert report's foundation, and a note saying "allocation happens in the receive path" without a location cannot be used in either.
[ ] Trace the execution path rather than reading by directory.
- Why. The claimed behavior is a sequence, and the sequence crosses files. Following the call chain from the entry point answers timing questions that reading files individually never will.
[ ] Check version control history where it is produced.
- Why. Commit messages and dates establish when the accused behavior appeared, which is a damages-period question and occasionally a willfulness question.
[ ] Bring the questions the code cannot answer to the engineer's deposition.
- Why. Code shows what the software does; it rarely shows why, and intent-adjacent questions such as the purpose of a configuration default belong to a witness.

Phase 15. A quality pass before service

Run this list against the finished chart, out loud, with someone who has not worked on it.

[ ] Every row cites a document, a file and line, a test, or testimony. No row says only that the product performs the step.
[ ] Every asserted claim has its own chart, and every accused product has its own chart.
[ ] The contested element's row is the longest row in the chart. If it is the shortest, the analysis has not been done.
[ ] Equivalents appear only where the file wrapper permits them, and each is charted function, way, and result.
[ ] Every method claim names the actor for every step.
[ ] Every indirect theory states a knowledge date and identifies the evidence of knowledge.
[ ] The marking analysis is in the file, with the licensee products checked and the actual notice date fixed.
[ ] No row depends on a construction that has not been proposed.
[ ] The chart is readable by a person who has not read the patent. Charts are read by law clerks and jurors, and a chart that only its author can follow is a chart that will be characterized by the other side.

Key Authorities at a Glance

| Authority | Proposition | Phase | |---|---|---| | 35 U.S.C. § 271(a) | The infringing acts | 5, 8 | | 35 U.S.C. § 271(b) | Inducement; requires knowledge | 8 | | 35 U.S.C. § 271(c) | Contributory infringement; staple article | 8 | | 35 U.S.C. § 271(e) | Regulatory safe harbor | 8 | | 35 U.S.C. § 271(f) | Components supplied for combination abroad | 8 | | 35 U.S.C. § 271(g) | Products made abroad by a patented process | 8 | | 35 U.S.C. § 112(f) | Means-plus-function; what to chart | 5 | | 35 U.S.C. § 154(a) | Term | 1 | | 35 U.S.C. § 154(d) | Provisional rights | 9 | | 35 U.S.C. § 261 | Assignment; standing | 1 | | 35 U.S.C. § 282 | Presumption of validity; defenses | 1 | | 35 U.S.C. § 284 | Damages and enhancement | 9 | | 35 U.S.C. § 285 | Fees in exceptional cases | 12 | | 35 U.S.C. § 286 | Six-year limitation | 9 | | 35 U.S.C. § 287(a) | Marking and actual notice | 9 | | 35 U.S.C. § 298 | Failure to obtain advice not evidence of willfulness | 9 | | 35 U.S.C. § 41(b) | Maintenance fees; lapse | 1 | | 37 C.F.R. § 1.321 | Terminal disclaimers | 1 | | Fed. R. Civ. P. 26 | Expert disclosures | 12 | | Fed. R. Civ. P. 56 | Summary judgment | 12 | | Fed. R. Evid. 702 | Expert testimony | 12 |

The five things people get wrong

One: they chart the easy elements first. Six obvious limitations documented in detail and the seventh marked "met" is the ordinary shape of a losing case, and it is prevented by ranking elements by confidence in week one. The instinct is understandable, because charting what you can prove produces visible progress and charting what you cannot produce silence. But the silence is the information.

Two: they cite marketing materials. Products do less than the marketing claims and more than the specification sheet says, and both directions produce charts that collapse in discovery. A brochure describing a feature is evidence that someone in marketing believed it, and nothing more.

Three: they plead equivalents without reading the file wrapper. Estoppel is public, it defeats most theories, and pleading first commits to a position the defendant dismantles with a document. The file wrapper takes an afternoon to read and it decides whether the theory exists at all.

Four: they never ask who performs each step. Divided infringement is the strongest defensive argument in a large share of software and platform cases, it is available at the pleading stage, and it is invisible to anyone who has not walked the claim actor by actor. The exercise takes an hour with a whiteboard.

Five: they run the marking analysis late. Licensee marking can eliminate years of damages, it is answerable in days, and it is routinely discovered at expert reports instead of in month one. 35 U.S.C. § 287(a). See Building or Defeating a Patent Infringement Case.