Most organisations budget carefully for the first version of a training library. Very few budget for the second. Two years on, an SOP has been rewritten, a regulator has published updated guidance, and the ERP system in your onboarding module now has a completely different interface. The library still plays, but half of it teaches behaviour your compliance team would fail on audit.
This is the quiet problem with training video. Production is a project; accuracy is a programme. The organisations that get years of value out of a training library are the ones that planned for change before they shot a single frame. The ones that treat every update as a fresh production tend to stop updating altogether, and the library dies of neglect somewhere in year three.
This article covers how to build and maintain a modular training library so that updating training videos when procedures change is a scheduled, contained piece of work rather than a crisis. It looks at shooting for modular replacement, deciding between a reshoot and a re-edit, retaining raw assets, and mapping version control to your LMS records so audit trails hold up.
Training video for Solas produced by One Productions
Ask an L&D manager how long they expect a compliance module to stay accurate and the usual answer is three to five years. Ask them to audit what they actually have and the picture is different. In regulated sectors, something in the content stack moves most quarters.
Track those four vectors against a library of thirty modules and you will find that something needs attention roughly every eight to twelve weeks. That is your real training content refresh cycle, and enterprise L&D teams tend to discover it retrospectively.
When there is no version-control approach, three things happen in sequence. First, out-of-date modules stay live because pulling them leaves a gap in mandatory training completion. Second, trainers start prefacing sessions with verbal corrections, which never make it into the record. Third, the whole library loses credibility with learners, who begin to treat all of it as approximate.
None of that shows up in a production invoice. It shows up in audit findings and in retraining costs.
The single highest-leverage decision in a training programme happens at pre-production, long before anyone thinks about maintenance. Structure the content so that individual pieces can be swapped without disturbing everything around them.
Training video for CPAS produced by One Productions
A twenty-minute module built as one continuous flow is a single object. Change one procedural step and you re-open the entire edit, potentially the entire shoot. The same twenty minutes built as eight discrete segments, each with a clean in and out point, is eight objects. Change one step and you replace one segment.
In practice this means:
Some content ages predictably. Screen recordings, dates, named individuals, statistics, regulation numbers, and site-specific detail all have short half-lives. Fundamental safety principles, behavioural coaching, and demonstration of physical technique tend to last.
Separate the two in the edit structure so that volatile content lives in replaceable layers. Screen captures held as overlay sequences with generic voiceover can be re-recorded and dropped in. Regulation references delivered as lower-third graphics rather than spoken lines can be updated with a graphics render and no talent involvement at all.
This is the core discipline of modular training video production, and it is a briefing decision as much as a technical one. Tell your production partner at the outset that the library must survive change, and the whole structure of the shoot changes.
Small scripting habits pay back for years. Avoid spoken dates where an on-screen graphic will do. Avoid naming specific software versions in voiceover. Refer to roles rather than individuals. Where a figure is likely to move, put it on screen rather than in the audio track, because updating a graphic is trivial and re-recording a voiceover line in a matching acoustic environment two years later is not.
When a change lands, the first question is scope. Getting this wrong in either direction is expensive. Over-cautious teams reshoot when a graphics update would have done. Over-optimistic teams patch with text overlays until a module looks like a palimpsest.
Most real-world updates land here. One segment out of eight needs new footage; the rest is fine. A half-day shoot with a small crew, matched to the original setup, replaces the affected segment and the module is back in service within a fortnight.
This is only possible if two conditions were met at the original production: the segments were built as discrete assets, and the technical setup was documented. A camera and lighting continuity sheet from the original shoot is worth more to a maintenance programme than any other single deliverable. If you are commissioning training video now, ask for one explicitly.
Training video for The UN produced by One Productions
Two years after delivery, the difference between a two-hour re-edit and a full reshoot is usually whether anyone kept the rushes. You’d be surprised how often the answer is no, or “probably, somewhere”.
Archive is a cost, and production companies do not hold multi-terabyte projects indefinitely by default. Settle three points before the project starts: how long raw assets are retained, who holds the archive, and whether project files transfer to you on delivery or remain with the producer.
For regulated training libraries, a five-year minimum retention is a sensible baseline, aligned to how long the training records themselves must be kept. Some clients take a full copy of everything on delivery and hold it in their own storage. That works, provided someone in the organisation knows the archive exists when the SOP changes.
Production-side versioning and LMS-side versioning are frequently disconnected, which is where audits get uncomfortable. The question an auditor asks is not “is the training current?” It is “can you demonstrate which version of this module a named employee completed on a given date, and what that version said?”
Keep it simple enough that non-production staff can use it. A three-part scheme covers most enterprise needs:
Carry that identifier everywhere: the file name, the LMS course record, the on-screen end card, and the production archive. When the auditor asks, you produce a single mapping table rather than an archaeology project.
Resist deleting old versions. Archive them as inactive but retrievable. The completion record of an employee who trained on v1.3 in 2024 is only meaningful if v1.3 still exists somewhere and can be produced on request. Training video version control in an LMS is fundamentally a records-management exercise, and deletion destroys evidence.
Tie the retraining trigger to version type rather than to a calendar. A major version release means everyone in scope repeats the module. A minor version means the content is silently updated and no completion records reset. Documenting that rule once, and applying it consistently, removes an argument that otherwise recurs with every update.
The most stable programmes review the whole library on a fixed cycle, typically annually, and treat urgent regulatory change as an exception that interrupts it. A scheduled review with a standing budget line is a fraction of the cost of periodic emergency remediation. It also means production capacity can be planned, so multiple segment updates are batched into a single shoot day instead of three separate mobilisations.
If you are at the earlier stage of building a library rather than maintaining one, our wider work across corporate training and internal communications video covers the production side of getting a programme off the ground.
An annual full-library review works for most organisations, with an exception process for regulatory changes that cannot wait. If your library contains significant software screen capture, move to a six-monthly review for those modules specifically, because UI changes arrive on the vendor’s schedule, not yours.
Segment updates are substantially less expensive, provided the original production was built modularly and the raw assets were retained. Where those conditions were not met, a full reshoot is often the more economical route, because reverse-engineering a monolithic module usually costs more in edit time than it saves.
Camera original footage including unused takes, editing and graphics project files, separated audio stems, a text-free master, and technical setup documentation. Agree retention periods and ownership in the production contract rather than assuming they are covered.
Version the source language master and treat localised versions as dependent children of it. When the master moves to v3.0, every localisation inherits a pending update flag. Text-free masters and separated audio stems make that process practical; without them, each language becomes an independent maintenance burden.
That depends on whether the change alters what the learner must do. Corrections to branding, terminology, or a regulation reference generally do not warrant retraining. Any change to a procedural step, a safety instruction, or a stated obligation normally does. Define the rule in your training governance documentation and apply it consistently, so the decision is not relitigated each time.
A training library becomes an asset rather than an operational burden when it is produced once and maintained well for years. That outcome is decided at pre-production, in decisions about segment structure, asset retention, and how volatile content is isolated from stable content. Retrofit it later and you are paying twice.
For L&D and compliance owners, the practical takeaway is to brief maintenance into the original scope. Ask your production partner how the library will be updated in year three, and judge the answer as carefully as you judge the showreel.
If you are planning a new training programme, or looking at an existing library that has drifted out of step with your procedures, we are happy to talk through the options. Get in touch with One Productions for a straightforward conversation about scope and approach.
Written by Freddie, One Productions.
A recruitment video consent form for employees, GDPR-compliant releases and practical ways to film on a live site while respecting staff who decline camera.
Accessible training video production explained: captions, transcripts, audio description and what the European Accessibility Act means for your team.
Learn B2B video attribution in HubSpot and Salesforce: connect video engagement data to influenced pipeline, revenue and reporting your board trusts.
