When Records Management Stopped Patenting the Future

Why the most important innovations in records management became standards instead of inventions

Imagine an economic historian, working a century from now, who sets out to write the innovation history of records management. She does what economic historians do. She goes to the patent record.

She finds a real body of evidence. Filings begin in the mid-1990s, thicken through the 2000s, peak around 2006, hold through the early 2010s, and then thin out. She checks 2020 and finds nothing at all. She checks the last few years before her cutoff and finds almost nothing.

She writes her conclusion carefully, because she is a careful scholar. Technical invention in records management appears to have plateaued in the mid-2010s and declined thereafter. Perhaps, she speculates, the problems were solved. Perhaps the market consolidated. Perhaps the profession simply stopped inventing.

Every sentence of that conclusion is defensible on the evidence she has. Every sentence is wrong.

I have been building the patent corpus she would use. It currently runs to 104 U.S. grants and published applications across 55 patent families, screened at the claims level to exclude filings where “record” means only a database row or a backup file. Forty-two of those 55 families have an earliest priority date before 2016. Thirteen come after. The single largest year is 2006, which alone accounts for 21 documents — fourteen of them a single cluster, the Electronic Records Archives filings made by Lockheed Martin and its partners on NARA’s behalf.

Set that curve beside what the same profession actually produced after 2015.

ISO 15489-1 was reissued in 2016 after three years of review. ISO/TR 22957 was substantially revised in 2018 to address enterprise content management. ISO 30301 was revised in 2019. ISO 23081-2 was revised in 2021. ISO/TS 7538, functional requirements for the disposition of records, was published in 2024 — as was ISO 18128, which did not merely revise the 2014 technical report on records risk but replaced it with a full International Standard under a new title. That last move is worth pausing on: a technical report is what a committee issues when the profession is still working something out. An International Standard is what it issues when the profession has reached a decision. Alongside them: capability models, defensible disposition as a repeatable practice, records management within Microsoft 365, cloud governance, machine-assisted classification, preservation metadata that actually survives migration, and a body of AI guidance currently being drafted.

That is not a profession that stopped innovating. That is a profession that stopped filing.

So the honest question is not whether innovation continued. It obviously did. The question is where it went, and why our standard instrument for measuring it went blind.

MetaArchivist is supported by readers who believe archives, records management, and standards deserve careful, original scholarship. If you'd like to encourage more work that combines archival research, technical standards, policy analysis, and historical investigation, I hope you'll consider becoming a paid subscriber.

---

Two ways a profession learns

Every field develops knowledge along two tracks.

The first is technological. New devices, new software, new mechanisms. This track has an intellectual property system built around it, an examination process, a citation structure, a public register, and a century of economic literature treating its output as a proxy for innovation. It is legible. It counts.

The second is institutional. New rules, new concepts, new vocabulary, new governance arrangements, new agreements about what “good” means. This track produces standards, frameworks, model regulations, professional doctrine. It has no register. It generates no counts. It is invisible to the instruments economists use.

Records management, over the past twenty-five years, migrated decisively into the second track. Not by accident, and not by failure. By fit.

---

What the early patents were actually trying to do

The earliest filing in my corpus is US 5,813,009, granted in 1998 on a 1995 priority, assigned to Univirtual Corporation: Computer based records management system method. Its claims cover retention schedules, suspension of disposal, certification of destruction, version history, custody, and location tracking. A 1996 filing combines a uniform subject classification scheme with retention functions. A 1998 filing by Andrea Jacobson claims the identification of official business email, the attachment of recordkeeping metadata, controlled read-only storage, and documented purge.

Read those claim sets, and something becomes clear. These are not really software inventions. They are professional doctrine, translated into apparatus so that it could be owned.

Retention scheduling was not invented in 1995. Certified destruction was not invented in 1995. What was new was the attempt to render an existing body of professional judgment as a machine — and the only vocabulary available for protecting that translation was the vocabulary of invention.

The 2006 ERA cluster is the purest case. Those filings — on authenticity, on preservation planning, on compound-record coherence, on access control across federated instances — explicitly invoke OAIS, InterPARES, and NARA lifecycle concepts in their specifications. The intellectual content is archival theory. The legal container is a patent. The inventors listed are contractors. The underlying archival concepts were part of a shared professional tradition rather than proprietary intellectual property — what could be claimed was the implementation, never the idea it implemented.

That mismatch is the whole story in miniature. The profession’s best ideas were already collective before anyone tried to file them.

---

The shift

Somewhere around the turn of the millennium, the governing question changed.

The patents ask: how should software store, classify, and dispose of records?

The standards ask something else entirely: what would make a record trustworthy, and how would anyone know?

That is not a refinement of the first question. It is a different order of problem. The first is soluble by engineering. The second is soluble only by agreement.

Australia asked the second question first, in AS 4390 (1996), the world’s first national records management standard. The United States asked a narrower version of it in DoD 5015.2-STD (November 1997), which NARA endorsed for federal use in November 1998. Europe asked it through MoReq, conceived in the DLM Forum after 5015.2 and funded into existence by the European Commission in 2000. Then AS 4390 was carried into ISO and became ISO 15489 in 2001.

Within five years the profession had built a new instrument. Not a machine. A settlement.

---

Standards as collective invention

Here is the philosophical core, and it has a real precedent in the economics literature.

In 1983 Robert C. Allen published a paper in the Journal of Economic Behavior and Organization about blast furnaces in the Cleveland district of northeast England between the 1850s and 1870s. He found that competing ironmaking firms routinely let rivals, visitors, and consultants inspect their plants and publish what they saw. Furnace height and blast temperature — the two variables that drove the era’s enormous productivity gains — rose steadily over twenty years through the accumulated adjustments of many firms, none of whom patented the improvements and all of whom benefited.

Allen called it collective invention.

The category has never, to my knowledge, been applied to records management. It should be.

A patent makes a claim about ownership: I made this, and here is the boundary of what is mine. A standard makes a claim about consensus: we determined this together, and here is what we will hold in common.

The second is not a lesser act of invention. It is a harder one. A working draft that survives to publication has passed through national committee review in dozens of countries, liaison negotiation with other technical committees, multiple ballot rounds, comment resolution line by line, and translation into languages whose recordkeeping traditions differ from the drafters’. ISO 15489 was adopted by more than fifty nations and translated into more than fifteen languages. The number of expert hours behind that is not smaller than the hours behind a patent portfolio. It is larger by orders of magnitude.

We simply have no counter for it.

---

The archive of innovation

Which brings me to what I actually went looking for, and what I found instead.

I ran 25 queries against the National Archives Catalog API, filtered to Record Group 64—the records of the National Archives and Records Administration itself. I was looking for the documentary trace of American participation in the development of ISO 15489.

Searching the exact phrase “ISO 15489” against RG 64 returns one result.

One.

NAID 633479323. Title: ISO - 15489 Records Management. Level of description: file unit. Parent series: Program Subject Files, NAID 313178464. Held at the National Archives at College Park, Textual Reference.

No start date. No end date. No creator recorded. No scope and content note. Not digitized. Zero digital objects.

A second folder sits nearby: ND - Survey Questionnaire of 15489, NAID 633478813, in the same series, described just as thinly. A third, in a different series also called Program Subject Files (NAID 23870286), is cataloged as ISO [International Standards Organization] - TC 46 / SCII Metadata — the subcommittee’s number transcribed as “II” rather than “11.” That folder is invisible to anyone searching “SC 11.”

The failures are as instructive as the finds. Querying “ISO/TC 46” against RG 64 returns zero results. “ISO/TC 46/SC 11” returns zero. “Standards Australia” returns zero. “Standards liaison” returns zero. Meanwhile, the bare string “TC 46” returns 252 hits, most of them noise.

The strongest contextual lead is a series that never mentions ISO at all: NAID 23870294, Official Managing Partner Records Relating to the E-Government Electronic Records Management Initiative, 2001–2007, created by the Office of the Archivist. Inside it: business cases submitted to OMB, meetings with Mark Forman, guidance on evaluating commercial off-the-shelf ERM applications, proof-of-concept pilot guidance, interagency review files, and one folder titled Issue Area 3 – Electronic Information Standards [May 2002 – May 2003] (NAID 33750084).

And here is the detail I keep returning to. The run produced 4,705 ranked results. Of the 4,631 in the lowest relevance band — the noise, the accidental keyword collisions — 4,303 have been digitized and are available online. Of the 74 results ranked high or very high, none are digitized. Not one. Every folder that actually bears on how this standard came to be sits in a box at College Park, reachable only by a reference request.

I want to be precise about what this does and does not show. A catalog search is not a survey of holdings. Description at folder level cannot expose every name and topic inside a box, and absence from an API result is not absence from the stacks. NARA says as much in its own research guidance, and I say it here for the same reason.

But grant the limits and the finding still stands, because the finding is not how few folders exist. It is what those folders are.

They are not administrative residue. They are the laboratory notebooks of an institutional invention. Where an engineering archive preserves prototypes and failed builds, this one preserves negotiation, consensus, professional judgment, translation, and influence. The survey questionnaire folder is a record of the profession polling itself about what it believed. The Issue Area 3 folder is a record of standards becoming procurement criteria becoming federal policy.

That pathway — draft to ballot to publication to regulation to purchase order — is the actual mechanism by which records management changed the world after 2001. And it is documented almost entirely in file units with no scope note, no date range, and no digital surrogate.

Records management may have stopped patenting its future because it began documenting how to build one collectively. The trouble is that we have not yet learned to describe that documentation as what it is.

---

Why standards outcompeted patents

The migration was not sentimental. It was structural.

Patents reward novelty, and reward it exclusively. The entire instrument is built to identify the boundary of what one party did that no one else had done, and to fence it.

Standards reward interoperability, and reward it inclusively. The entire instrument is designed to make many parties behave in a compatible manner.

Now consider what records management actually needs. Archives require interoperability, because records arrive from everywhere and must survive their creators. Government requires interoperability, because agencies must be comparable, auditable, and successor-proof. Long-term preservation requires interoperability by definition, since the whole enterprise is a bet that a future system will be able to read what a present system wrote.

A successful patent makes a technique scarcer. A standard that succeeds makes a technique universal. For a profession whose core obligation is continuity across time and organizational boundaries, only one of those is useful.

There is an obvious objection here, and it deserves a direct answer.

The post-2015 decline in my corpus has a competing explanation unrelated to professional migration. On 19 June 2014, the Supreme Court decided Alice Corp. v. CLS Bank International, holding that reciting a generic computer implementation of an abstract idea does not make it patentable. Records management claims are almost a textbook target: apply a retention rule on a computer. Filings in this space would have fallen after 2014 whether or not the profession’s center of gravity had moved. There is a second artifact, too: U.S. applications are published 18 months after their earliest priority date, so a corpus assembled in July 2026 is structurally blind to filings from 2025 and 2026, and partially blind to 2024. Some of the cliff is measurement, not behavior.

I think this strengthens the argument rather than weakening it.

Because the point was never that patent counts fell for a single, clear reason. The point is that patent counts are a poor instrument for this field, and Alice demonstrates exactly why. A single doctrinal ruling changed the measured innovation rate of an entire profession without changing a single thing the profession was doing. If the yardstick can move that far while the object stays still, the yardstick was never measuring the object.

Meanwhile, ISO/TS 7538 was published in 2024. No court ruling was required.

---

ISO 15489 as an invention

We are in the habit of calling ISO 15489 guidance. I want to propose that we have been underselling it.

It should be read as one of the profession’s great inventions — not because it introduced a technology, but because it reorganized professional thought.

National frameworks certainly existed before it — most notably Australia’s AS 4390 and the U.S. DoD 5015.2 standard, and neither was a small achievement. But they were national answers, bounded by the jurisdictions and administrative cultures that produced them. What ISO 15489 did was transform those national approaches into a genuinely international conceptual vocabulary: a shared account of what a record is, what makes one authentic, reliable, and usable, and what a records system must do to be trustworthy. Before 2001 there was no such vocabulary. After 2001 there was. That vocabulary propagated into national regulations, software requirements, procurement specifications, training curricula, consulting methodologies, university programs, and the compliance language of industries that have never employed an archivist.

It did all of that without a single patent claim. It could not have done any of it with one.

Judge it the way we judge inventions — by diffusion, by durability, by how much subsequent work builds on it — and ISO 15489 outperforms every patent in my corpus combined. Most of those patents have expired, been abandoned, or are held by companies that no longer make the product. The standard was revised in 2016 and remains the field’s conceptual foundation twenty-five years on.

---

Why this is an archival problem, not just an interesting one

Here is the twist, aimed squarely at us.

Archivists preserve the evidence of innovation. That is a professional commitment we have honored for a long time, and we have a well-developed apparatus for it: the inventors' papers, the laboratory records, the engineering notebooks, the drawings, the prototypes.

But innovation has changed form. Increasingly it consists of committee drafts, ballot comments, liaison correspondence, working papers, survey instruments, meeting minutes, mailing list threads, Git repositories, and issue trackers.

The records of standards development have quietly become one of the most important documentary genres of the digital age. Not only in records management — in every domain where the decisive intellectual work happens in a committee rather than a laboratory. Web protocols. Accounting rules. Clinical terminologies. Building codes. Data formats. In each case the artifact everyone knows is the published standard, and the record of how it was reached is scattered across institutional files that nobody has appraised as innovation records.

We are not preserving them badly. In many cases we are preserving them accidentally — as subject files, as office correspondence, as the administrative sediment of a program office — and describing them accordingly. A folder called ISO - 15489 Records Management, with no date range and no scope note, that sits in a general subject series is not a description failure. It is a category failure. Nobody cataloged it as evidence of invention, because we do not have that category.

---

The proposition

Which brings me to the idea I actually want to put into circulation.

Standards are a profession’s intellectual capital, preserved in documentary form.

A patent documents ownership of an invention. A records management standard documents the codification of collective professional knowledge — the moment a discipline decides that something it has been arguing about is now settled enough to be relied upon.

If that is right, then the records behind a standard are not administrative byproducts. The drafts, the ballots, the comments, the resolutions, the correspondence: those are the archival record of how a profession learns. They are the only surviving evidence of the reasoning, the dissent, the compromises, and the roads not taken. The published standard shows the conclusion. Only the file shows the argument.

And this gives archives a capability that patent data cannot match. Patent statistics answer who owned the invention. They cannot answer how a community came to agree that this idea was correct — who objected, on what grounds, and what changed their minds.

Only archives preserve that. Which means the appraisal decision facing us is larger than it looks. Standards development files are not low-value program records. They are the primary sources for the intellectual history of every knowledge profession that works by consensus. Which, increasingly, is all of them.

And this opens a research program, not just a cataloging problem. If standards constitute collective invention, then archives hold a largely unexplored source base for the history of innovation itself. Historians have spent decades mining patent databases for a simple reason: patents are searchable, countable, and conveniently centralized. The documentary record of how professions actually build shared knowledge is none of those things. It lies in committee drafts, ballot comments, liaison files, survey instruments, and working papers — scattered across institutional holdings, described at folder level, rarely digitized, and almost never appraised as evidence of invention. The archives of standards development may ultimately prove as significant to understanding the modern knowledge economy as patent records were to understanding the industrial one.

We will not know until we go and look. And we cannot go and look until we decide these files are worth keeping for that reason.

---

Coda

Every generation imagines innovation as a moment — someone in a room, something new, a date you can put on a plaque.

Archives suggest something less dramatic and more durable.

Most innovations survive not because they were patented. They survive because a community agreed to repeat them.

That agreement has a documentary history. For records managers, that history lives not in patent files but in committee drafts, ballot resolutions, meeting minutes, liaison correspondence, implementation guides, and the slow translation of a standard into a regulation into a procurement requirement.

Perhaps that is the profession’s real contribution to the study of innovation. We learned — earlier than most, and mostly without noticing — that the most durable inventions are not always machines.

Sometimes they are agreements.

---

Notes on sources

The patent figures come from a research inventory of 104 U.S. grants and published applications across 55 families, screened at claims level and current to July 2026. It is a research landscape, not an exhaustive count, and legal status fields are research snapshots rather than legal conclusions.

The RG 64 figures come from 25 queries run against the National Archives Catalog API v2 with a Record Group 64 filter in July 2026. Several queries were truncated at a configured page limit; the catalog searches description, not contents; and absence from results does not establish absence from holdings. This work uses the National Archives Catalog API but is not endorsed or certified by NARA.

Robert C. Allen, “Collective invention,” Journal of Economic Behavior and Organization 4, no. 1 (1983): 1–24.

Alice Corp. Pty. Ltd. v. CLS Bank International, 573 U.S. 208 (2014).

Standards chronology: AS 4390 (1996); DoD 5015.2-STD (1997, NARA-endorsed 1998); MoReq (2001), MoReq2 (2008); ISO 15489-1 (2001, rev. 2016); ISO 23081-1 (2006), -2 (2009, rev. 2021); ISO 30300/30301 (2011; 30301 rev. 2019); ISO/TR 18128 (2014), replaced by ISO 18128 (2024), which elevated the technical report to a full International Standard under the title Records risks — Risk assessment for records management; ISO/TR 22957 (2009, rev. 2018); ISO/TS 7538 (2024).

This article was prepared with the assistance of generative AI for research support, translation, drafting, and editorial review. The author directed the research, verified factual claims against the cited sources, and is solely responsible for the analysis, interpretations, and conclusions presented here.