Course File Management: How to Organise a Training Library That Keeps Growing

Why Does a Course Library Become Unmanageable?

The usual explanation is volume, and it is wrong. Plenty of teams run four hundred courses comfortably and plenty of teams lose control at thirty. What breaks a library is that several different kinds of thing accumulate at different speeds, and normally only the courses themselves have an agreed home.

Six specific failures do the damage, and each one has a fix that costs nothing.

1. Copies outnumber originals. Somebody cannot find a course, so they build it again. Now there are two, both plausible, and the next person picks one at random.
2. Nobody can tell which version is live. The course in the library and the package sitting in the learning management system drifted apart months ago and no record says when.
3. Media is re-uploaded instead of reused. The same logo arrives eleven times under eleven filenames, which wastes storage and guarantees that updating the brand means finding all eleven.
4. The source material is somewhere else. The course was built from a policy document that lives in a shared drive nobody linked to, so when the policy changes there is no trail back.
5. Ownership is implicit. Everyone assumes the person who built a course still maintains it. They left in March.
6. Nothing is ever deleted. Deletion feels risky, so retired courses sit alongside current ones and the library slowly stops being a list of what you offer.

Notice that five of the six are findability problems and only one is a storage problem. That distinction matters when you are choosing a tool, because vendors sell storage and what you need is retrieval.

Three Libraries, Not One

The single most useful shift is to stop thinking about "the course library" as one thing. There are three, they have different lifespans, and filing them together is what produces the mess.

The last row is the one that causes the most expensive kind of failure. Courses and media normally sit inside the authoring platform. The source material sits in a document management system, an intranet or somebody's inbox, and the connection between a course and the document it was generated from exists only in the memory of whoever built it. When the source changes, the course does not, because nothing pointed back.

The fix is one line of text. Put the source location in the course description field, every time, without exception. It costs five seconds at build time and it is the difference between a twenty minute update and a rebuild.

How Should You Structure Course Folders?

Most teams file by topic because topic is how they think about training. Topic is a poor filing key, because a course belongs to several topics and to exactly one owner.

File by who owns it and how long it lives. Five rules make that concrete.

1. Top level is the owning group. A department, a client, a region, a business unit. Whatever unit actually signs off content. If two people disagree about which folder a course goes in, you have found a genuine ownership question and it is better raised now than at audit.
2. Second level is the programme, not the subject. "Onboarding 2026" rather than "Health and Safety". A programme has a start, an end and a person responsible. A subject has none of those.
3. Third level exists only when a programme has cohorts or variants. Regional versions, language versions, role-specific tracks. If there is no variation, stop at two levels. Depth you do not need is friction you pay for daily.
4. Retired content gets a sibling, not a subfolder. An "Archive" folder alongside the live ones, at the same level, so a stale course can never be reached by browsing the current structure by accident.
5. One folder per client, always, if you are a consultancy. Client separation is the one case where the folder tree is doing compliance work rather than convenience work, and it should never be improvised.

The test of a structure is not whether it looks tidy. It is whether a new starter, given a course title and nothing else, can predict which folder it is in before they look. If they cannot, the structure is decoration.

A Naming Convention That Survives a Team

A convention only works if it is short enough that people follow it when they are busy. Seven rules are about the limit.

1. Front-load the distinguishing word. Lists sort alphabetically and truncate on the right, so "Fire Safety, Warehouse" beats "Warehouse Fire Safety Course" every time.
2. Never put "new", "final", "latest" or "current" in a title. Every one of those words becomes a lie within a quarter, and "Final v2" is the recognised symptom of a library with no convention at all.
3. Use ISO dates or no dates. 2026-04 sorts correctly and reads the same in every country. April 2026 does neither.
4. Name the audience when the audience is the variable. If there are three versions of the same course for three regions, the region belongs in the title and nothing else needs to change.
5. Name the language in the title, not just the metadata. Translated courses are the single most duplicated thing in a multilingual library, because the original and the translation look identical in a list.
6. Drop the word "course" and the word "training". Everything in a course library is a course. The word carries no information and it eats the visible part of the title.
7. Write the convention down in one paragraph and put it where the folders are. A convention that lives in one person's head is a preference, not a convention.

How Do You Handle Versions Without Version Control?

Be clear about what authoring platforms actually give you, because the marketing word is "version history" and the mechanic is usually something narrower.

What you can generally rely on is recovery and attribution. Deleted work can be brought back, edits are attributable to an author, and lists can be sorted by date modified so the most recently touched course is visible. What you generally cannot do is roll a course back to how it looked on a specific Tuesday, the way a developer rolls back code. If your compliance position depends on proving what a course said on a given date, do not assume the platform will answer that.

There is a straightforward answer that almost nobody uses deliberately. Your SCORM export is your version archive. A genuine SCORM package is a self-contained zip file holding the course exactly as it stood at the moment you exported it. Export at every deployment, store the zip with the date in its filename, and you have a dated, openable record of every version you ever published, independent of the vendor and independent of whether you still have an account.

That only works if the export is a real file. Some platforms ship what looks like a SCORM package but is actually a thin wrapper that calls back to their servers each time a learner opens it, which means it stops working when the subscription does and preserves nothing. Check which kind you get before you build a retention process on it.

Three habits carry the rest.

1. One live course, never two. If a course is being rewritten, the draft lives in the archive folder or it does not exist. Two plausible current versions is the condition you are trying to avoid.
2. Put the review date in the description, not the title. Titles are read by learners. Descriptions are read by the people who maintain the library.
3. Record the deployment, not just the build. The date that matters for an audit is the date the package reached the learning management system, which is rarely the date it was finished.

The Media Library Is the Half Nobody Plans For

Course folders get designed. Media almost never does, and media is where the volume lives. A hundred courses might hold four thousand images between them.

Four practices keep it usable.

1. Approved brand assets go in one shared folder and everything else references it. Logos, colour swatches, approved photography, the standard course cover. If an author has to hunt for the logo they will use whichever copy they find first, and one of those copies is out of date.
2. Upload once, reuse everywhere. A media library exists so a file can be used in forty lessons without existing forty times. Reuse is what makes a brand refresh a one-afternoon job instead of a project.
3. Crop at upload, not at use. A single correctly sized asset beats the same photograph stored at five different crops, and the difference compounds across a library.
4. Sweep orphans on the same cycle as courses. Media whose course was deleted is invisible and permanent. Nothing prompts you to remove it, so it has to be a scheduled job.

What Storage Limits Actually Constrain

Storage limits are quoted in gigabytes, which makes them sound like a course count. They are not. Text and interactions are tiny. Video is not, and video is the only thing most libraries will ever run out of room for.

QuikAuthor includes 500 MB of storage on the free plan, 25 GB on PRO, 100 GB on TEAM, and flexible storage on Enterprise. Read those figures as video budgets. Five hundred megabytes is comfortable for a library of interactive, image-led courses and tight the moment you start hosting your own footage. If your training is video-first, the storage line is the plan line that should decide your choice, not the course count. Current figures for every plan are on the pricing page.

The related limit worth reading carefully is what happens to your content if you stop paying. On QuikAuthor, courses and learner data are never deleted: downgrading to the free plan keeps access to your first 25 courses and everything else is preserved and available again the moment you upgrade. Ask every vendor the same question in the same words, because the answers vary far more than the pricing does.

Deleting, Archiving and the Courses Nobody Owns

Libraries grow because deleting feels risky and archiving feels like admin. Both are cheaper than the alternative, which is a library where half the contents cannot be trusted.

Separate two things that get conflated. The completion record is what an auditor asks for and it needs to be retained for as long as your policy says. The course itself is a teaching artefact and it has no independent evidential value once it has been superseded. Retiring a course does not throw away the record that people completed it.

A workable retirement process has four steps. Move the course to the archive folder so it cannot be assigned by accident. Note the date and the reason in the description. Keep the last SCORM export as the frozen copy. Then, after whatever interval your retention policy sets, delete it properly, including its orphaned media.

Do this annually and it takes an afternoon. Leave it for four years and it becomes a project nobody funds.

What to Check in an Authoring Tool

Work through these with the vendor's own documentation open, not their sales page.

1. Are folders available on every plan, or only on a team tier? Organisation is sometimes packaged as a collaboration feature, which means the solo author who most needs structure is the one who does not get it.
2. Is there a depth limit on nesting? Two levels is fine for most teams and painful for consultancies with many clients.
3. Does search cover descriptions and media, or only course titles? Title-only search fails exactly when the library gets big enough to need it.
4. Can media live in folders of its own? A flat media list of four thousand items is not a library, it is a pile.
5. Is a deleted course recoverable, and for how long? "Recoverable" with no stated window is a promise you cannot plan around.
6. Can you export a single lesson, or only whole courses? If updating one module means republishing everything, your maintenance cost is set by the tool rather than by the change.
7. Is the export a self-contained file or a link back to the vendor? This decides whether your archive is an archive.
8. What is the storage allowance, and what happens when you reach it? Some platforms block uploads, some charge, some degrade quality. All three are survivable if you knew in advance.

How File Management Works in QuikAuthor

Folders and sub-folders are available on every plan, including the free one, with no limit on how deep the structure goes. New folders take a name, a description and an optional cover image, and the project library shows breadcrumbs as you move down the tree. In list view, courses and folders can be dragged between folders directly, and the library sorts by date modified, date created or title.

Search runs across the course library, and the media library has its own folders and its own search across assets. Media uploaded once can be used in any lesson, Unsplash is integrated into the picker for stock photography, and landscape images can be cropped in place rather than re-uploaded at a different size. On TEAM, shared folders make approved assets available to every author, comments sit inline on the slide they refer to, and every course has an attributable author.

Deletions are recoverable rather than final: removed slides can be restored from the trash panel inside the lesson builder, and removed courses can be brought back. Exports are standalone SCORM 1.2 and SCORM 2004 packages with no calls back to our servers, and individual lessons export as their own packages as well as whole courses, which is what makes module-level updates and dated archiving practical. The file management feature page covers the library itself in more detail.

Two honest limits. QuikAuthor does not offer point-in-time rollback of a course to a previous state, so the dated SCORM export described above is the archiving method we would actually recommend rather than a workaround. And the free plan's 500 MB is a genuine ceiling, not a soft one, which is the right constraint to check first if your courses carry your own video.

Frequently Asked Questions

What is file management for course creators?

File management for course creators is the practice of organising three things that accumulate separately: finished courses, the media used inside them, and the source documents the courses were built from. Most authoring platforms organise the first two and leave the third outside the tool entirely, which is why a course and the policy it came from usually lose contact with each other.

How should a course library be organised?

Organise by ownership and lifespan rather than by subject. Make the top level the group that signs off content, the second level the programme rather than the topic, and add a third level only where a programme genuinely has regional, language or role variants. Keep retired courses in an archive folder that sits alongside the live structure rather than inside it.

Should courses be filed by department or by topic?

By department, or by whichever unit actually owns and approves the content. A course belongs to several topics at once and to exactly one owner, so topic produces arguments about where things go and ownership does not. The working test is whether a new starter can predict the folder from the course title alone.

What is the best naming convention for e-learning courses?

Put the distinguishing word first, because lists sort alphabetically and truncate on the right. Use ISO dates such as 2026-04 or no dates at all. Name the audience and the language when either is the variable. Never use "new", "final", "latest" or "current", since all four become inaccurate within a quarter, and drop the words "course" and "training", which carry no information in a course library.

Do authoring tools have version history?

Usually less than the phrase suggests. Most offer recovery of deleted work, attribution of edits to an author and sorting by date modified, which is not the same as restoring a course to how it looked on a particular date. If you need a dated record of what a course said, export a self-contained SCORM package at each deployment and keep the zip with the date in its filename.

How much storage does a course library need?

It depends almost entirely on video, since text, images and interactions are small by comparison. A library of interactive, image-led courses fits comfortably in a few hundred megabytes, while self-hosted video moves the requirement into tens of gigabytes. QuikAuthor includes 500 MB on the free plan, 25 GB on PRO, 100 GB on TEAM and flexible storage on Enterprise.

What should happen to a course that is out of date?

Move it to an archive folder so it cannot be assigned by accident, record the date and the reason in its description, and keep the last SCORM export as the frozen copy. Delete it properly, along with its orphaned media, once your retention policy allows. Retiring a course does not remove the completion records of the people who took it, and those records are what an audit asks for.

Can two authors work in the same course library without overwriting each other?

That depends on the plan rather than the product. QuikAuthor PRO is a single-author workspace, and TEAM adds three authors sharing one library with shared asset folders, inline comments on slides and an attributable author on every course. Beyond three authors, Enterprise is the route. Check the author count on any platform before assuming a shared library is included.

Where This Guide Stops

Two things are outside what any filing structure can fix.

A convention nobody enforces is worse than no convention, because it produces a library that looks organised and is not. If you introduce one, somebody has to own it, and the honest version of that sentence is that the owner has to be willing to rename other people's courses.

And none of this tells you whether a course is any good or whether it is still accurate. Findability and accuracy are different problems with different owners. A perfectly filed library of out-of-date training is still out-of-date training, and the review cycle that fixes that is programme work rather than file management.

Organise One Folder and See If It Holds

Pick the messiest programme you own, apply the two-level structure and the naming convention to it alone, and see whether a colleague can find things without asking. QuikAuthor's free plan includes folders, sub-folders, the media library and 25 courses with no credit card and no expiry.

Related reading: Course File Management for L&D Teams · Rapid Authoring Tools · SCORM Export