Patent and Trademark Docketing Software: One Docket, Two Clocks

Patent and trademark docketing software is sold as one category, and the deadline rules underneath it are two. Patent dates run from a filing date or an issue date, both fixed and both published. Trademark dates run from a registration date, from a period of use the owner has to prove, and from dates held at WIPO that the USPTO never mails anyone. Whether you should run one system or two is downstream of a narrower question: can the engine you are looking at compute both rule sets without letting one inherit the other’s defaults? Here is what it has to get right, verified against the USPTO and WIPO, and how to test it.

What Patent and Trademark Docketing Software Has to Reconcile

A combined docket looks like one problem and is actually two. Patent dates hang off events the applicant controls or receives — a filing date, an issue date, a mailing date. Trademark dates hang off a registration date, a period of use the client has to prove, and dates set at WIPO that never appear in a USPTO record at all.

That is why the choice between one platform and two is not a preference. It is a question about whether a single rules engine can hold two unrelated bodies of law without one of them quietly inheriting the other’s defaults. The failure is always the same shape: a date that computed, looked right on the dashboard, and was derived from the wrong rule.

The published guidance on this question is thin. The two most visible articles on the topic — Alt Legal’s piece on the difference between IP, patent and trademark docketing software and its follow-up on best-in-class versus all-in-one — frame the trade-off in terms of cost, training and workflow, and they are useful on those points. Neither states a single statutory deadline. Anaqua’s PATTSY WAVE page markets combined patent and trademark docketing to law firms and reports that more than 200 use it, but publishes no deadline rules either.

So the buyer is asked to choose an architecture without being shown the rules that architecture has to compute. Below are those rules, verified this month against the USPTO and WIPO, and then a way to test any platform against them.

The Two Clocks, Side by Side

Every row below is a place where the patent answer and the trademark answer differ in kind, not just in length. A rules engine that treats them as variants of one concept will get at least one of them wrong.

Event	Patent side	Trademark side
Post-grant maintenance	Maintenance fees payable at 3–3.5, 7–7.5 and 11–11.5 years from the issue date, with grace periods at 3.5–4, 7.5–8 and 11.5–12 years on payment of a surcharge (USPTO).	Section 8 declaration of continuing use between the fifth and sixth years after the registration date; Section 9 renewal between the ninth and tenth years and every ten years after that; six-month grace with an additional fee (USPTO).
Trigger date	Issue date and filing date — both are on the face of the patent and never move.	Registration date, plus a period of use in commerce the owner must be able to prove. Use is a fact about the client’s business, not a field the office publishes.
Office action response	The shortened statutory period is typically three months, extendable in one-month increments to the six-month statutory maximum on payment of the extension fee under 37 CFR 1.136(a).	Three months for office actions issued during examination on or after 3 December 2022, with one three-month extension available for a fee, to a six-month maximum. Post-registration office actions moved to the same three-plus-three structure on 7 October 2023 (USPTO).
The exception that breaks engines	None comparable.	Section 66(a) applications — US designations under the Madrid Protocol — keep a six-month response period and cannot be extended (USPTO).
Optional strengthening filing	None. A patent is as strong on day one as it will ever be.	Section 15 incontestability, available once the mark has been in continuous use in commerce for five years after registration and is still in use, and filable within one year after the expiration of any such five-year period (USPTO).
International dependency	A PCT application enters the national phase at 30 or 31 months depending on the office. Once granted, national patents stand alone.	An international registration stays dependent on the basic mark for five years from the international registration date. If the basic mark ceases in that window, the international registration falls with it (Madrid Protocol).
Rescue window	Petition practice for unintentional delay, on the office’s own terms.	Transformation into national or regional applications, which must be filed within three months of the cancellation and keeps the international registration date (Article 9quinquies, WIPO).
Fee basis	Per case, with claim counts driving excess-claims fees.	Per class, per mark — a single registration in four classes is four fee events and can be four partial outcomes.

Read the last two rows together and the point of the whole exercise appears. The Madrid dependency clock runs from a date held at WIPO, ends five years later, and has a three-month rescue window that opens on a cancellation the USPTO never mails you. There is no patent equivalent, so there is nothing in a patent-derived data model for it to inherit. Our breakdown of Madrid Protocol deadlines walks the same dates from the international registration’s point of view.

Where One Combined System Actually Breaks

All-in-one platforms do not fail loudly. They fail on the rows above, in ways that look like ordinary records until the window closes. Four patterns account for most of it.

• Rule inheritance. The engine was built for one right and the other was added as a matter type. Symptom: a Section 66(a) registration gets a Section 8 due date instead of a Section 71, or a Madrid-based application is docketed with a three-month office action response and a three-month extension that does not exist for it.
• One reminder ladder for two risk profiles. A patent annuity that slips a week is recoverable with a surcharge. A Section 8 that slips past the grace period is a cancelled registration. The same 90/60/30-day ladder under-serves the second case.
• Per-case counting on a per-class right. If the system treats a four-class registration as one record, partial outcomes have nowhere to live — and a specimen refusal in one class is exactly a partial outcome.
• Use dates with no owner. First-use and continued-use dates come from the client, not the register. A platform designed around office data usually has a field for them and no workflow that makes anyone fill it in.

None of this argues against combined platforms. It argues that the combination has to be verified at the rule level rather than at the feature level, because every vendor in the category can truthfully say it supports both rights. The nine deadline rules in our guide to trademark docketing software are the specific ones a patent-derived engine tends to approximate.

Where Two Specialist Systems Break Instead

The split-docket answer is respectable and, for lopsided portfolios, often right. Alt Legal makes the fair point that two systems can cost less than one when pricing runs per matter: a firm with 2,000 patents and 8,000 trademarks is paying for 10,000 matters either way. What the argument leaves out is what a split costs in places that do not appear on an invoice.

• No single owner record. An assignment, a merger or a name change has to be recorded twice, in two schemas, by two people. The second one is the one that gets missed.
• Chain of title drifts. Patents and marks acquired in the same transaction diverge in the records within a year or two, and nobody notices until a due-diligence request asks for both.
• Conflict blindness. The team docketing an office action on a mark cannot see that the same brand is named in a patent-side licence, so no one asks the question.
• Two truths in reporting. Board and client reporting is assembled by hand from two exports, and hand-assembled reports are reconciled at exactly the moment someone asks for them.
• Orphaned matters. Copyrights, domains and design rights belong to neither system and end up in a spreadsheet, which is where deadlines go to be forgotten.

A split docket is a real answer when both engines are genuinely better at their own right and someone owns the reconciliation between them as a named job. It is a bad answer when it happens by accident — which is how most splits actually arise, usually through a lateral hire or an acquisition that brought its own platform along.

Choosing: One Docket or Two

Portfolio mix decides less than people assume. Three other variables decide more.

Question	Points to one system	Points to two
Who works the queue daily?	One team, or one person, covering both rights.	Separate patent and trademark teams that already work independently.
Is one right under 15% of the portfolio?	Yes — the minority right cannot justify its own platform, licence and governance.	No — both sides have enough volume to earn specialist tooling.
Do ownership records change often?	Yes. Acquisitions, name changes and assignments punish split records hardest.	Rarely, and a named person already owns reconciliation.
Does the trademark side include Madrid designations?	Only if the engine computes Section 71, the six-month unextendable response and the five-year dependency correctly — test it.	Yes, and the trademark specialist demonstrably handles them.
Who signs off a missed date?	One accountable owner across both rights.	Two owners, with an explicit rule for anything that touches both.

Cost belongs in the decision, but later than most firms put it, and it is not a headline number. Per-user and per-matter pricing invert as the team grows or the portfolio does, and implementation and data work are usually the larger line. Our breakdown of what trademark docketing software actually costs covers the fee structures that make two quotes non-comparable; on the patent side, the vendor-by-vendor comparison sets out the same evaluation categories.

It is also worth naming the third option honestly. A firm can run one platform and still put the trademark rule set in specialist hands — either an in-house docketer who owns it or a managed docketing service. The rules engine and the person who works it are separate decisions, and conflating them is why platform migrations so often fail to reduce missed dates.

Eight Scenarios to Run in the Demo

Feature grids do not separate these platforms; every one of them ticks “patents” and “trademarks”. Rule tests do. Take these eight into the demo, ask the vendor to enter each record live, and compare the dates the system produces against the answers below. Any vendor that has built the rules properly will enjoy the exercise.

1. A Section 66(a) registration. Enter a US registration that issued from a Madrid designation. Correct output: a Section 71 declaration due between the fifth and sixth years after the US registration date — not a Section 8 — plus a separate renewal of the international registration at WIPO on the international registration’s own ten-year cycle.
2. An office action on that same Madrid-based application. Correct output: six months, with the extension option unavailable. If the system offers a three-month extension, the engine has applied the domestic rule to a filing basis it does not cover.
3. An examination office action issued last month on a Section 1(a) application. Correct output: three months, with a distinct, separately dated extension request deadline that falls on or before the three-month date, and an outer bound of six months.
4. A post-registration office action. Correct output: the same three-plus-three structure, on the strength of the 7 October 2023 change. Older systems still carry six months here.
5. A registration approaching its fifth anniversary, in continuous use. Correct output: two dates, not one — the Section 8 window opening at year five, and the optional Section 15 declaration, which stays available for one year after the five-year use period ends.
6. An international registration two years old. Correct output: a docketed end of the five-year dependency period, and a rule that turns a ceasing of effect in the basic mark into a three-month transformation deadline in every designated territory.
7. A patent issued three years and one month ago. Correct output: the first maintenance fee window shown as open now, closing at 3.5 years, with grace to four years at a surcharge — window and grace as separate dates, not one deadline.
8. A four-class registration with a refusal in one class. Correct output: the record survives with three classes proceeding and one in refusal, each with its own dates. If the platform can only carry one status per registration, it will lose that distinction.

Two follow-up questions matter as much as the eight tests. First: when a rule changes, who updates the engine, how fast, and does the vendor recalculate open matters or only new ones? The December 2022 and October 2023 response-period changes are the live example — ask what happened to matters already docketed at the time. Second: what is the source of each date? A system that imports computed dates from a previous platform is inheriting that platform’s errors under a new interface.

Migrating a Combined Docket Without Losing Dates

Whichever architecture wins, someone has to move the records, and migration is where combined dockets are most often damaged. The rule is simple and widely ignored: migrate the facts, recalculate the dates.

1. Import source facts only. Registration dates, issue dates, filing bases, class lists, international registration numbers, first-use dates. Not the computed due dates.
2. Recalculate every deadline in the new engine from its own rules, then diff the result against the old system’s dates. The diff is the audit — every disagreement is either a migration defect or an error you have been carrying.
3. Reconcile against primary sources. Status and dates from the USPTO register and Patent Center, international registrations against WIPO records. Never from the export you are migrating.
4. Run both systems in parallel through one full cycle that includes at least one maintenance window and one office action response, and compare the queues weekly.
5. Retire the old system only after the diff is empty and someone has signed that it is. A read-only archive costs little and settles later arguments about what was known.

The diff step is the one firms skip, and it is the one that pays. Every disagreement between two engines is a date that at least one of them computed wrongly, surfaced at the only moment it is cheap to fix. The same method drives a standing trademark docket audit, and the errors that most often show up in a trademark docket are almost all visible in that diff before they are visible in a lost registration.

How PerspireIP Can Help

Choosing between one platform and two is a question about tooling. Whether the dates are right is a question about who works them. PerspireIP provides managed trademark docketing alongside your patent docket, whatever platform it sits in — the trademark rule set and a trained operator, sized to your portfolio, with no implementation project to run.

Every engagement opens the same way: we recalculate your trademark dates from the governing rule and reconcile them against the register, then hand you the diff. Talk to our trademark docketing team before your next demo, and take the eight scenarios above in with you.

Frequently Asked Questions