PTAB Petition Checklist: Grounds, Prior Art, Expert Declarations, and the Institution Decision

By Casey Scott McKay · 2026-06-30

A petition to the Patent Trial and Appeal Board is the entire case, filed once, under a hard statutory bar, with no discovery to fill gaps and no appeal from a denial of institution. This checklist runs it in ten phases: calendar the bar, order the file wrapper, commission a targeted search, identify every real party in interest, inventory what estoppel will cost, select grounds under the all-or-nothing rule, build the charts and the declaration, file and move to stay, respond as patent owner, and manage the trial phase through the final written decision. Each box states the reason, the authority, and the trap. Boxes marked as gates should clear before filing. A worked matter runs through the phases.

IP and Technology > Patent Litigation | Checklist | Published 30 June 2026 - Updated 16 July 2026 | Casey Scott McKay - marksy.us

Summary. A petition to the Patent Trial and Appeal Board is the entire case, filed once, under a hard statutory bar, with no discovery to fill gaps and no appeal from a denial of institution. This checklist runs it in ten phases: calendar the bar, order the file wrapper, commission a targeted search, identify every real party in interest, inventory what estoppel will cost, select grounds under the all-or-nothing rule, build the charts and the declaration, file and move to stay, respond as patent owner, and manage the trial phase through the final written decision. Each box states the reason, the authority, and the trap. Boxes marked as gates should clear before filing. A worked matter runs through the phases.

Keywords: PTAB petition, inter partes review checklist, one-year bar, real party in interest, prior art search, ground selection, claim charts, motivation to combine, level of ordinary skill, expert declaration, mandatory notices, word limit, preliminary response, discretionary denial, institution decision, stay pending review, estoppel inventory, motion to amend, objective indicia, final written decision

How to use this checklist

| Phase | What it covers | When | |---|---|---| | 1 | Calendar the bar and the real deadline | Day one | | 2 | Read the file wrapper | Weeks two to three | | 3 | Commission the search | Week six | | 4 | Real parties in interest | Before drafting | | 5 | Estoppel inventory | Before deciding | | 6 | Ground selection | Month five | | 7 | Charts and declaration | Months six to seven | | 8 | File, then move to stay | Month seven | | 9 | Patent owner's preliminary response | Three months after filing | | 10 | Trial phase through decision | Twelve months from institution |

Boxes marked [Gate] should clear before the petition is filed. There is no amendment to the petition and no appeal from a denial of institution.

The matter. A patent on a network protocol, asserted against an equipment maker. The examiner allowed the claims after a single amendment adding a timing limitation. Everything below follows from that amendment.

Phase 1. Calendar the bar and the real deadline

[ ] Enter the one-year date from service of the complaint.
- Why. A petition filed after it is barred.
- Authority. 35 U.S.C. § 315(b).
- Trap. Counting from the filing date rather than the service date, or from an amended complaint rather than the original.
[ ] Enter a go or no-go date at roughly month six.
- Why. A competent petition needs a six-week search, an expert briefed, charts built, and a drafting cycle. Working backward, the decision cannot wait.
[ ] Check whether any privy or real party in interest was served earlier.
- Why. Service on a privy starts the clock for the petitioner.
- Trap. A supplier sued a year before the customer, funding the customer's defense.
[ ] Check whether post-grant review is still open.
- Why. Nine months from issuance, and it permits every ground.
- Authority. 35 U.S.C. § 321.
[ ] Note the district court trial date.
- Why. It is the strongest argument against institution and it argues for filing early.

Phase 2. Read the file wrapper

[ ] [Gate] Read the prosecution history before commissioning any search.
- Why. It identifies the point of novelty, which is where the search should aim.
- Trap. A general search of the field, which returns volume and not the reference that matters.
[ ] List every reference the examiner cited and what the office actions said about it.
- Why. Art already considered invites discretionary denial and rarely persuades.
[ ] Identify the amendment that produced allowance.
- Why. The limitation the applicant added is the limitation the petition has to meet.
[ ] Note every statement the applicant made about the prior art.
- Why. Characterizations of what a reference teaches bind the patent owner and are usable at the Board.
[ ] Check for related applications, continuations, and family members.
- Why. A cancelled patent with pending continuations is not the end of the exposure.

Phase 3. Commission the search

[ ] Direct the search at the specific limitation the examiner allowed over.
[ ] Prioritize sources examiners search poorly.
- Why. Non-patent literature, conference proceedings, standards contributions, foreign patent documents, product manuals, and theses.
[ ] Ask for a materiality assessment, not a reference list.
- Why. The question is whether the art is materially better than what the examiner saw, and only the searcher who read both can answer it.
[ ] Verify the public availability date of every candidate reference.
- Why. A printed publication must have been publicly accessible before the critical date, and a conference paper or standards contribution requires proof of when and how it became available.
- Authority. 35 U.S.C. § 102.
- Trap. A reference with an ambiguous date, which the patent owner will attack and which may sink the ground.
[ ] [Gate] Decide whether the art is materially better. If it is not, do not file.

Phase 4. Real parties in interest

[ ] [Gate] Prepare a written real party in interest analysis before drafting.
- Why. Incomplete disclosure terminates petitions on grounds unrelated to the merits.
[ ] Consider parents, subsidiaries, affiliates, indemnitors, indemnitees, suppliers, covered customers, insurers, funders, and joint defense members.
- Why. The test is functional — at whose behest is the petition filed — and formal separation does not resolve it.
[ ] Confirm none of them was served with a complaint more than a year ago.
[ ] List related matters completely.
- Why. Every district court case, every other Office proceeding, and every related application.

Phase 5. Estoppel inventory

[ ] [Gate] List the grounds that will survive a final written decision.
- Authority. 35 U.S.C. § 315(e).
[ ] Confirm eligibility, section 112, on-sale, public use, and inequitable conduct all stay in district court.
- Authority. 35 U.S.C. § 101; 35 U.S.C. § 112.
[ ] Assess whether the documentary case is the whole invalidity case.
- Why. If it is, the petition bets everything. If the on-sale bar is independently strong, the bet is cheap.
[ ] Understand that estoppel reaches art a diligent search would have found.
- Trap. Holding back a fourth reference in the belief it is preserved.

Phase 6. Ground selection

[ ] [Gate] Rank candidate grounds by the strength of the weakest element mapping.
[ ] Cut everything below the line.
- Why. The Board institutes on all grounds or none, so a marginal ground endangers the good ones.
- Authority. 35 U.S.C. § 314.
[ ] Prefer anticipation where a single reference genuinely discloses every limitation.
- Why. It avoids the motivation-to-combine problem entirely.
[ ] Prefer two-reference combinations over three-reference stacks.
[ ] Justify parallel petitions in writing if more than one is filed, ranked in order of preference.
- Trap. Two petitions filed without explanation, reduced to one by the Board.
[ ] Check joinder if a co-defendant has already filed.
- Authority. 35 U.S.C. § 315(c).

Phase 7. Charts and declaration

[ ] Chart element by element with pin citations.
- Trap. A block quotation next to a limitation, which maps nothing.
[ ] Chart dependent claims individually.
[ ] State the level of ordinary skill explicitly, with expert support.
[ ] Construe only the terms that affect the grounds.
- Trap. Eleven constructions, ten of which create disputes without benefit.
[ ] [Gate] Supply a specific motivation to combine.
- Why. Grounded in the references, the knowledge of the skilled artisan, or the problem being solved, with a reasonable expectation of success.
- Trap. Same field of endeavor, which the Board rejects routinely.
[ ] Address teaching away in the petition if the reference contains discouraging language.
[ ] Have the expert write the analysis independently.
- Trap. A declaration tracking the petition verbatim, which is discounted.
[ ] Brief the expert on the file wrapper before the declaration is drafted.
- Trap. Discovering at deposition that the expert never read what the applicant argued.
[ ] Prepare the expert for hindsight questioning.
[ ] Address foreseeable objective indicia and the nexus question.
[ ] Check the word limit and cut background, not charts.

Phase 8. File, then move to stay

[ ] Complete the mandatory notices: real parties in interest, related matters, counsel, service.
[ ] File the certification of standing.
[ ] Move to stay the district court case on filing.
- Why. Usually denied, costs little, and preserves the record.
[ ] Renew the stay motion after institution.
- Why. Courts grant stays far more often once a trial has been instituted.
[ ] Update district court invalidity contentions to reflect what is at the Board.

Phase 9. The patent owner's preliminary response

[ ] Lead with procedure where it is available.
- Why. An unnamed real party in interest or a bar running from a privy is cleaner than the merits.
[ ] Build discretionary denial from the district court record.
- Why. Trial date, investment, overlap, and the petitioner's delay.
[ ] Argue previously presented art where the petition recycles the examiner's references.
[ ] Attack the combination rather than the references.
- Why. Petitions fail on motivation far more often than on what a reference discloses.
[ ] Hold the objective indicia and the detailed technical rebuttal unless decisive on their face.
- Trap. Fronting the whole defense, losing institution, and having previewed everything.

Phase 10. Trial phase through decision

[ ] Prepare declarant depositions as trial testimony.
- Why. In a paper proceeding the transcript is the testimony, and there is no motion in limine to fix a bad answer.
[ ] Keep claim construction positions consistent with the district court case.
- Trap. A construction argued to avoid the art that also defeats infringement, or one argued to capture the product that also captures the prior art.
[ ] Confine the reply to responding.
- Why. New evidence and new theories in reply are excluded.
[ ] Treat any motion to amend as a fallback.
- Authority. 35 U.S.C. § 316(d).
- Why. Substitute claims must be narrower, supported by the original disclosure, and patentable over the art of record.
[ ] Prepare for the hearing as questioning, not presentation.
- Why. Demonstratives may not contain new argument, and the panel has read everything.
[ ] Calendar the final written decision within twelve months of institution.
- Authority. 35 U.S.C. § 316(a)(11).
[ ] Update the district court contentions after the decision to remove estopped grounds.
[ ] Consider appeal, and read the standard honestly.
- Authority. 35 U.S.C. § 141.
- Why. Legal conclusions are reviewed without deference; what a reference teaches is a factual finding reviewed for substantial evidence, and it is hard to disturb.

Phase 11. The petition document, section by section

A petition has a conventional structure. Departing from it costs the panel's patience without buying anything, and a section that is missing reads as an oversight rather than a choice.

[ ] Mandatory notices, complete.
- Why. Real parties in interest, related matters, lead and back-up counsel, and service information. Related matters means every district court action, every other Office proceeding touching the patent or its family, and every related application.
- Trap. A related continuation left off, which reads as carelessness about everything that follows.
[ ] Grounds for standing.
- Why. The certification that the petitioner is neither barred nor estopped. Short, and it is where the one-year analysis surfaces on the face of the document.
[ ] Statement of the precise relief requested.
- Why. The ground table in prose. A reader should see the entire case in half a page: which claims, which statutory basis, which references.
[ ] Technology background, kept short.
- Trap. Five pages of tutorial consuming the word budget the charts need. A page of orientation is generous.
[ ] Level of ordinary skill, with support.
- Why. Every obviousness question is answered from that person's perspective, and leaving it blank leaves the central variable undefined.
[ ] Claim construction, for contested terms only.
- Why. State the construction, cite the intrinsic evidence, and show the ground works under it. If the district court construed the term differently, say so and address it.
[ ] One section per ground, self-contained.
- Trap. Heavy cross-referencing, which forces a panel reading ground three to reconstruct it from ground one.
[ ] Objective indicia addressed where foreseeable.
- Why. A petitioner who knows the product succeeded commercially should meet nexus in the petition rather than in a reply constrained against new argument.
[ ] Exhibit list, numbered and consistent.
- Trap. Citations in the declaration that use different exhibit numbers than the petition.

Phase 12. Ground drafting, in detail

[ ] For anticipation, confirm every limitation appears in one reference, arranged as claimed.
- Why. A reference disclosing all the pieces in a different arrangement does not anticipate.
[ ] For inherency, show the feature is necessarily present, not merely possible.
- Trap. An expert opinion that the feature "would typically" be present, which is not inherency.
[ ] For obviousness, state what each reference teaches against the specific limitations at issue.
- Why. A summary of the reference as a whole is orientation, not analysis.
[ ] State the reason to combine, ranked from strongest source.
- Why. Strongest: the references themselves — one identifies a problem, the other solves it, or one expressly suggests the modification. Next: the state of the art with expert support. Weakest, and routinely rejected: shared technical field.
[ ] State the reasonable expectation of success.
- Trap. Skipping it where the modification is non-trivial, which is exactly where it is required.
[ ] Address any teaching-away language in the references.
- Why. The patent owner will find it, and answering first is cheaper than answering under a reply constraint.
[ ] Confirm the combination does not require redesigning the primary reference.
- Why. A modification that changes the principal operation of the primary reference, or renders it unsatisfactory for its purpose, is a familiar route to rejection of the ground.
[ ] Chart each dependent claim.
- Trap. Claim 1 charted in detail and claims 2 through 12 disposed of in a paragraph, which produces institution on claim 1 alone if at all.

Phase 13. The declaration

[ ] Confirm the expert is of ordinary skill in the actual field.
- Trap. A distinguished academic from an adjacent discipline, easier to impeach than a working engineer.
[ ] Confirm the expert read the patent, the file wrapper, and every reference before drafting.
[ ] Require reasoning for every conclusion.
- Why. Conclusory expert testimony on obviousness receives little weight, and the Board says so in decision after decision.
[ ] Confirm the declaration is not verbatim the petition.
[ ] Confirm every exhibit cited in the declaration is filed.
[ ] List the expert's prior testimony and engagements.
- Why. It will be asked at deposition, and a surprise there is a credibility problem the record keeps.
[ ] Prepare the expert on the four questions that matter.
- Why. How were the references selected; what in the reference supports this conclusion; what did the file wrapper say; what would have led a skilled artisan to this combination without knowing the patent.

Phase 14. Common defects, collected

Run this list against the finished petition before filing, with a reader who has not worked on it.

[ ] No ground rests on a reference whose public availability date is unproven.
[ ] No chart quotes a passage without identifying the words that meet the limitation.
[ ] No motivation to combine rests on shared field of endeavor alone.
[ ] No claim construction is proposed that is not needed for a ground.
[ ] No construction proposed here contradicts a position taken in district court.
[ ] No dependent claim is challenged without its own analysis.
[ ] No real party in interest is omitted.
[ ] No related matter is omitted.
[ ] The declaration reads as the expert's work.
[ ] The petition is within the word limit without having cut the charts.
[ ] Every exhibit number matches between petition, declaration, and exhibit list.
[ ] The estoppel inventory is in the file and the client has seen it.

Phase 15. What to tell the client, at each stage

[ ] At the filing decision. The petition buys a decision, not a result. It is a bet on the search, and the search has already told you whether the bet is worth making.
[ ] At institution. Institution is not a merits ruling and denial is not vindication. A discretionary denial is not a finding that the patent is valid, and describing it that way to a client creates a problem when the district court case continues.
[ ] During the trial phase. Discovery is one deposition per declarant. There will be no document production, no interrogatories, and no motion practice of the kind district court litigation involves.
[ ] At settlement. Termination before a final written decision generally avoids estoppel, which is why a large share of these proceedings never reach decision. The Board retains discretion to proceed anyway where the record is complete.
[ ] At the decision. All claims cancelled ends the matter as to that patent, subject to pending continuations. Some claims cancelled leaves an infringement case on the survivors. All claims surviving makes the patent materially stronger and moves settlement value against the defendant.
[ ] On appeal. The Federal Circuit reviews legal conclusions without deference and factual findings for substantial evidence. What a reference teaches is factual. Appeals succeed more often on construction and procedure than on the technical merits.

Phase 16. The patent owner's full response, after institution

Institution changes the posture. The Board has already found a reasonable likelihood on at least one claim, and the response is written against that finding rather than in ignorance of it.

[ ] Read the institution decision as a roadmap.
- Why. It says which arguments the panel found persuasive and which it passed over. A response that does not address the panel's stated reasoning is answering a different petition.
[ ] Attack the combination first.
- Why. Petitions fail on motivation far more often than on what a reference discloses, and the combination rationale is where a hindsight argument does its work.
[ ] Show teaching away with the reference's own language where it exists.
- Why. A reference that describes the petitioner's modification as undesirable, or that solves the problem a different way for stated reasons, is the strongest evidence available.
[ ] Show the modification would defeat the primary reference's purpose.
- Why. A combination requiring the principal operation of the primary reference to change is not the combination a skilled artisan would have made.
[ ] Develop objective indicia with a nexus.
- Why. Commercial success, long-felt need, failure of others, industry praise, licensing, and copying are part of the obviousness analysis, and the evidence is in the patent owner's hands and nowhere else.
- Trap. Sales figures without a showing that the success traces to the claimed features rather than to marketing, price, or an unclaimed feature.
[ ] Support the response with a declaration that does independent work.
- Why. The same discount that falls on a petitioner's mirror declaration falls here.
[ ] Confirm every construction argued is one the infringement case survives.
- Trap. Winning the review on a construction that makes the accused product non-infringing.
[ ] Take the petitioner's expert's deposition before finalizing the response.
- Why. The transcript is the only discovery available and it frequently supplies the response's best material.

Phase 17. Parallel proceedings housekeeping

[ ] Maintain a running map of grounds: at the Board, in district court, estopped on decision.
- Why. The invalidity case is one case with two forums, and the map is what prevents an estopped ground from appearing in an expert report.
[ ] Keep claim construction positions consistent across forums.
- Why. A position taken at the Board is available in district court whether or not it formally binds.
[ ] Use one expert across both proceedings, or accept the risk.
- Trap. Two experts whose positions differ, which hands the other side a cross-examination it did not have to build.
[ ] Confirm the duty of disclosure is satisfied for any co-pending application in the family.
- Authority. 37 C.F.R. § 1.56.
- Why. Art cited in the petition is material to any pending continuation, and failing to disclose it there is a separate problem entirely.
[ ] Preserve documents notwithstanding a stay.
- Trap. Treating a stay as a suspension of preservation obligations.
[ ] Track the International Trade Commission schedule if a section 337 investigation is pending.
- Why. The Commission does not wait, estoppel from a final written decision reaches it, and the two schedules do not align.
[ ] Reassess settlement at each milestone.
- Why. Filing, institution, and the close of briefing each change the risk profile materially, and each is a natural moment to revisit numbers.

Phase 18. Choosing the right proceeding

Before any of the boxes above, confirm the Board is the right forum at all.

[ ] Confirm the best invalidity argument is documentary.
- Why. Only patents and printed publications under sections 102 and 103 are available. A case resting on eligibility, indefiniteness, enablement, an on-sale bar, or a physical product as prior art has no home in an inter partes review.
- Authority. 35 U.S.C. § 311.
[ ] Check whether the nine-month post-grant window is still open.
- Why. Post-grant review permits every ground, including the ones inter partes review excludes, and it is the better tool where it is available. The estoppel is correspondingly broader.
- Authority. 35 U.S.C. § 321.
[ ] Consider ex parte reexamination where the bar has passed or the budget will not stretch.
- Why. No estoppel, no identification of the real party required, filable at any time. The requester has no further role, the proceeding has no deadline, and the patent owner amends and argues freely with the examiner.
- Authority. 35 U.S.C. § 302.
[ ] Consider recommending no Office proceeding at all.
- Why. Where the art is what the examiner already saw, where the district court trial date makes discretionary denial likely, or where the strongest arguments are the excluded ones, the money is better spent in district court. A recommendation not to file is a real recommendation and belongs in the file as one.
[ ] If the client is the patent owner, understand where reissue leads.
- Authority. 35 U.S.C. § 251.
- Why. Broadening reissue is available only within two years of grant, and a reissue opens the claims to fresh examination on every ground.
[ ] Confirm the client understands the cost of a patent that survives.
- Why. A patent that emerges from review intact carries estoppel against the petitioner on the documentary grounds, reads to a jury as twice-examined, and prices higher in every negotiation that follows. That outcome is the real downside of filing, and it should be described before the decision rather than after it.

Phase 19. The record you leave behind

Everything filed in these proceedings is public and permanent, and it is read later by people who were not in the matter.

[ ] Assume every argument will be quoted back.
- Why. A construction argued to the Board, a characterization of what a reference teaches, an expert's concession at deposition — all of it is available to any future adversary, in any forum, against any patent in the family.
[ ] Check the petition against positions taken in other matters.
- Trap. A company that argued a term broadly to capture an accused product in one case and narrowly to avoid prior art in another, with both filings public.
[ ] As patent owner, watch what the response says about scope.
- Why. Distinguishing a reference by describing the claim narrowly is the Board's version of prosecution disclaimer, and district courts read those statements.
[ ] Keep the estoppel map current in the client file, not just in counsel's head.
- Why. The lawyers who handle the next assertion will not be the ones who filed the petition, and the map is what tells them which grounds are gone.
[ ] Preserve the search results and the searcher's materiality assessment.
- Why. They document what a diligent search found, which is the measure estoppel uses, and they support a later argument that a reference was not reasonably available.
[ ] Debrief after the decision, whichever way it went.
- Why. The useful questions are narrow: did the search aim at the right limitation, did the expert's analysis hold up at deposition, was the ground that was cut the right one to cut, and did the stay motion timing work. These answers improve the next petition and they are forgotten within a month if nobody writes them down.

Outcome. The search returned a standards contribution and a foreign application, neither before the examiner, both disclosing the timing limitation. A third candidate ground was cut. The petition was filed in month seven, the stay was granted after institution, and the independent claims were cancelled. The on-sale defense, preserved throughout, was never needed.

Key Authorities at a Glance

| Authority | Proposition | Phase | |---|---|---| | 35 U.S.C. § 102 | Novelty; public accessibility of publications | 3, 6 | | 35 U.S.C. § 103 | Obviousness; motivation to combine | 6, 7 | | 35 U.S.C. § 101 | Not available; survives estoppel | 5 | | 35 U.S.C. § 112 | Not available; survives estoppel | 5 | | 35 U.S.C. § 282 | Presumption of validity; absent at the Board | 5 | | 35 U.S.C. § 302 | Ex parte reexamination | 1 | | 35 U.S.C. § 311 | Inter partes review; scope | 6 | | 35 U.S.C. § 314 | Institution threshold; all or nothing | 6, 9 | | 35 U.S.C. § 315(b) | One-year bar | 1, 4 | | 35 U.S.C. § 315(c) | Joinder | 6 | | 35 U.S.C. § 315(e) | Estoppel | 5, 10 | | 35 U.S.C. § 316 | Conduct; twelve-month decision; amendment | 10 | | 35 U.S.C. § 321 | Post-grant review; nine-month window | 1 | | 35 U.S.C. § 251 | Reissue | 10 | | 35 U.S.C. § 141 | Appeal to the Federal Circuit | 10 | | 37 C.F.R. § 42.100 | Rules for inter partes review | 7 | | 37 C.F.R. § 1.56 | Duty of disclosure | 2 |

The five things people get wrong

One: they file late. The one-year bar is not the deadline; month six is. A petition drafted in six weeks against a bar three weeks away loses on institution for reasons that have nothing to do with the art, and a late petition also hands the patent owner its best discretionary denial argument.

Two: they file on the examiner's art. References already on the face of the patent were considered, rejected as insufficient, and are now being offered again. That is a hard argument on the merits and an easy denial on discretion.

Three: they include the marginal ground. All-or-nothing institution makes the weakest ground the one that decides whether any of them are tried. Cutting until it hurts is the correct instinct.

Four: they let counsel write the declaration. A declaration that tracks the petition verbatim adds nothing the Board can weigh, and the discount usually falls on the obviousness ground that needed the expert most. 35 U.S.C. § 103.

Five: they skip the real party in interest analysis. It is the failure mode with no connection to the merits, it terminates petitions, and it can import a bar from a complaint served on a supplier two years earlier. See Filing or Defending an Inter Partes Review.