Records Management What Is A Record

You’re staring at a shared drive with 47,000 files. Half the filenames are "Final_v2_REAL_final.Now, docx. " Nobody knows who owns the folder called "Project Alpha — DO NOT DELETE" from 2016. Legal just asked for every communication related to a vendor dispute from three years ago. You have until Friday.

Sound familiar?

This is the moment most organizations realize they don’t actually know what a record is. Now, they think they do. That said, they confuse records with documents, documents with data, and data with "stuff we should probably keep just in case. " That confusion costs money, creates legal exposure, and turns simple audits into month-long fire drills.

Let’s fix the foundation. Because until you can look at a file and confidently say this is a record* or this isn’t*, every retention policy, every classification schema, and every expensive ECM platform is just theater.

What Is a Record in Records Management

The shortest answer: a record is evidence of a business transaction. That’s it. But the devil lives in the distinction between evidence* and information*.

ISO 15489 — the international standard for records management — defines a record as "information created, received, and maintained as evidence and as an asset by an organization or person, in pursuance of legal obligations or in the transaction of business.Day to day, " Notice the words evidence* and obligations*. Those are the anchors.

A record isn’t defined by its format. An email can be a record. A Slack message can be a record. Because of that, a voice memo, a scanned contract, a database entry, a handwritten note on a napkin — if it proves something happened, it’s a record. Conversely, a beautifully formatted PDF sitting in your Drafts folder? Consider this: not a record. Because of that, not yet. It’s just a document.

The Three Tests

When I’m training teams, I give them three questions. If the answer to all three is yes, it’s a record. If any answer is no, it probably isn’t.

1. Does it provide evidence of an action, decision, or transaction? An approved invoice proves a payment was authorized. A draft of that invoice proves nothing except that someone was typing.

2. Is there a legal, regulatory, or operational obligation to keep it? "Operational" matters here. You might not be legally required to keep every project status report, but if your team can’t function without the history of why Scope Change #4 was approved, that’s an operational obligation.

3. Is it the official, final version — or a designated draft with significance? Final signed contracts. Final board minutes. The email thread where the VP wrote "Approved, proceed." Those are records. The seventeen earlier drafts? Usually not. Unless the process* of drafting is what’s under scrutiny — say, in a patent dispute or contract negotiation — then those drafts become records too.

Records vs. Documents vs. Non-Records

This is where most people trip.

A document is any structured unit of information. It has a lifecycle: draft, review, final, archive. Only the final (or significant intermediate) versions cross the line into record territory.

A non-record is everything else. Think about it: " The cat GIF your coworker sent in Teams. Reference copies. These have zero retention requirement. Spam. Zero legal weight. Convenience copies. Personal notes. Marketing brochures you downloaded "for inspiration.They should be deleted the moment they’re no longer useful — which is usually immediately.

The problem? Because of that, most organizations treat non-records like records. They back them up. Day to day, they migrate them. They pay for storage, eDiscovery review, and backup windows for files that have no business existing past Tuesday.

Why It Matters

You might think this is academic. It’s not. The definition of a record drives everything downstream.

Legal Defensibility

When litigation hits, the first question from opposing counsel (or the judge) is: "What is your records retention schedule, and did you follow it?" If you can’t define what a record is, you can’t have a defensible schedule. You end up over-retaining — which means producing thousands of irrelevant, embarrassing, or damaging files during discovery — or under-retaining, which looks like spoliation. In practice, both are expensive. One can be fatal to a case.

Storage and Cost

Unstructured data grows at 30–50% annually in most enterprises. A huge chunk of that is non-records. Duplicate files. Old installers. Because of that, thumbnails. If you classify correctly, you delete the noise. Here's the thing — you stop paying for tier-one storage for files nobody has opened since 2014. Also, you shrink backup windows. In real terms, you reduce cloud egress fees. The savings are real, even if they don’t show up as a line item on the budget.

Findability and Trust

When someone searches for "the signed NDA with Vendor X," they should get one result. That's why they start keeping personal copies on desktops, in personal OneDrives, in printed binders under the desk. Also, when the system returns noise, people stop trusting it. The record. In practice, not seventeen drafts, three scanned copies with different filenames, and an email thread discussing the NDA. Shadow IT is born from bad classification.

If you found this helpful, you might also enjoy how much did the hoover dam cost or george bush hearing about 9 11.

AI and Analytics Readiness

This is the new frontier. So you want to train a model on your contract history? Feed it your records.

AI and Analytics Readiness

This is the new frontier. Clean, well‑classified data is the single most valuable asset for any AI initiative. Even so, feed it your records. You want to train a model on your contract history? But feed it your non‑records—drafts, duplicates, the “Final_v3” files—and you’ll get a model that can’t tell the difference between a signed agreement and a personal note. The noise skews sentiment, inflates similarity scores, and, worst of all, produces legal risk if the model is used for compliance reporting. By treating non‑records as disposable, you give your analytics team a much cleaner training set, faster inference, and a lower risk of mis‑classification.

Putting the Theory into Practice

1. Establish a Clear Definition

   • Record = any document that, by legal, regulatory, or operational necessity, must be preserved in a specific form for a defined period.
   • Non‑record = any document that does not meet those criteria.
     Publish this definition in the organization’s data governance charter and make it the basis for all classification rules.

2. Automate Classification Where Possible

   • Use content‑based classification engines (NLP, machine‑learning classifiers) to tag documents on ingestion.
   • Implement file‑type heuristics (e.g., “.docx” → candidate record, “.tmp” → non‑record).
   • Combine with metadata (author, creation date, folder path) to reduce false positives.

3. Enforce Retention Schedules

   • Map each classification to an appropriate retention period.
   • take advantage of tiered storage: move records to cold, cost‑effective storage after the active period; delete non‑records immediately or after a short grace period.
   • Automate the lifecycle via DMS or archival solutions that enforce policies at the file system level.

4. Audit and Monitor

   • Run quarterly compliance audits to verify that records are retained correctly and non‑records are purged.
   • Use dashboards to show storage consumption by classification, identify outliers, and trigger alerts for policy violations.

5. Educate and Empower Users

   • Provide simple, contextual prompts in the document creation workflow (e.g., “Is this a final, signed agreement?”).
   • Offer quick‑start guides for naming conventions and folder structures that align with the classification schema.
   • Reward compliance: make it part of performance reviews or provide recognition for teams that keep their repositories lean.

The Bottom Line

Treating non‑records as expendable is not a cost‑cutting gimmick—it’s an operational discipline that protects your organization on multiple fronts:

Once you stop treating every file as a record, you give your people, your IT, and your legal teams a clearer, more efficient working environment. The cost savings are tangible, the risk exposure is dramatically lowered, and the data you rely on for decision‑making becomes far more reliable.

Take the first step today: audit one folder, classify its contents, and enforce a deletion policy on the non‑records. If that can be done in a week, scaling across the enterprise is a matter of a few more weeks. The discipline of distinguishing records from non‑records is simple, but the payoff is profound.