People willingly spend hours watching YouTube, yet groan when asked to watch an eight-minute training video. Why is that?
At Cherryleaf, we develop videos for our e-learning courses on technical communication and publish our own marketing videos on YouTube. But we are not, by any stretch, social media influencers. To look beyond our own experience, we asked several chatbots what L&D teams could learn from successful creators. This article combines their suggestions with lessons from our own work and research.
One idea kept surfacing:
Internal training videos should stop being treated as presentations that happen to have been recorded. They should be designed as pieces of media for a particular audience and purpose.
That shift affects the opening, script, presenter, visuals, editing, and title. It also affects what happens after publication.
L&D teams can borrow a great deal from social media creators. But they should not copy them wholesale.
Attention is necessary, but it is not the outcome
Creators often optimise for watch time, shares, and reach. L&D teams need people to understand, remember, and apply something.
Those aims overlap, but they are not identical.
A cliffhanger might keep someone watching. It will not necessarily help them approve an invoice correctly or handle a difficult conversation.
The better goal is useful attention: the minimum attention someone needs to understand and perform a task.
That distinction should guide every production decision.
Open with the reason to watch
Creators know that the opening must quickly fulfil the promise made by the title or thumbnail. YouTube even treats the first 30 seconds as a distinct section in its audience-retention reports.
A workplace video does not need a sensational hook. It needs a relevance contract.
For example:
“This claim was rejected because one box was left unticked. Here is how to prevent that happening to yours.”
A useful opening usually does one of four things:
- Shows the problem.
- Demonstrates the desired result.
- Presents a decision the learner must make.
- Promises a specific, useful outcome.
Give each video one clear job
Successful creator videos tend to have a recognisable proposition. The viewer knows what they will get in return for their time.
Training videos are often less disciplined. A single video might attempt to explain an entire policy, process or management topic.
A better title might be:
- Approve an expense claim in Workday.
- Prepare for your first one-to-one.
- Report a near miss.
- Reset a customer’s multi-factor authentication.
Each title describes a task rather than a place in a course.
Shorter, focused videos are also easier to find, reuse and update. A software change might require replacing one four-minute demonstration rather than rebuilding a 45-minute programme.
Research based on 6.9 million MOOC viewing sessions found that shorter videos were associated with greater engagement. It also found that purpose-made online videos performed better than recorded classroom lectures. However, the researchers measured viewing behaviour, not workplace performance. Their findings do not prove that every subject should fit into six minutes. Read the study by Guo, Kim, and Rubin.
Length should follow the subject
A three-minute refresher can be ideal before completing an unfamiliar task. A difficult ethical judgement might need a longer scenario, discussion and practice.
Show the work
YouTube and TikTok are visual media. Creators demonstrate products, show examples, and put claims on screen.
Workplace video too often shows a presenter while they describe something that could be demonstrated. This has often been because organisations have found it difficult to create visuals. With AI, that’s now a lot easier to do.
If the video explains how to use a system, show the system. If it covers customer conversations, show a realistic conversation. If it explains a safety check, show both the correct action and the common mistake.
Useful visuals include:
- Screen recordings with the relevant control enlarged.
- Before-and-after comparisons.
- Real forms, tools and equipment.
- Short role-play scenes.
- Diagrams that reveal a relationship or sequence.
- Key values, warnings and decision rules displayed on screen.
Generic office footage rarely helps. A slow-motion shot of smiling colleagues walking through a glass building does not explain a process. Every visual should earn its place.
Use credible people
Creators build a relationship with viewers through direct speech, personality, and a recognisable voice.
L&D teams can do the same without asking subject-matter experts to behave like influencers.
A knowledgeable colleague speaking naturally to the camera will often feel more credible than an unseen narrator reading committee-written copy over slides. Minor imperfections can make the delivery feel human.
That does not excuse poor preparation or inaccurate information.
Give presenters a clear structure and rehearse the difficult sections. Let them use their own vocabulary where possible. You could experiment with scripting only the wording that must be exact.
For regulated material, a good principle is approved meaning, natural delivery. Legal or safety-critical statements can remain fixed while the surrounding explanation sounds like speech.
We use a teleprompter, and one of the benefits of doing so, is it solves accuracy problems.
Edit out friction, rather than thinking time
Creators remove pauses, repetition, false starts, and unnecessary setup. L&D teams should be equally ruthless.
Editing pace and learning pace are different. A presenter can move quickly through familiar context, then slow down for a difficult software action or safety decision.
Use close-ups, pauses, and repetition where the learner needs to inspect something. Remove dead time everywhere else. AI can help with the editing.
A faster delivery can add energy. Mechanically speeding up a dense explanation gives the learner less time to think.
Write for the ear
Corporate scripts often sound like documents being read aloud.
Compare:
“In circumstances where employees encounter discrepancies within the submitted documentation, the matter should be escalated.”
With:
“If the figures do not match, stop and ask your manager to check them.”
Creator scripts tend to use short sentences, contractions, direct address, and ordinary words. Training scripts should do the same.
Read the script aloud before recording. If the presenter runs out of breath or loses the thread, rewrite it.
Natural language does not mean vague language. Names, values, warnings, and technical terms still need to be precise.
Design for the real viewing environment
Creators understand the platform on which their work will appear. Corporate video producers cannot assume a single platform.
A training video can appear in an LMS, Teams channel, knowledge base, email, or application window. Someone might watch it on a phone, a laptop, or beside the system they are learning to use.
Accessibility also needs to be part of the original production. Provide corrected captions and a clean transcript.
Make useful videos easy to find
Creators pay close attention to titles, descriptions, thumbnails, playlists, and series. Internal video libraries often contain filenames such as Onboarding_Module_4_FINAL_v3.
Employees usually search for the problem in front of them. Use those words.
Organise videos into task-based collections such as:
- First 30 days.
- People manager essentials.
- Common CRM fixes.
- Two-minute safety checks.
Add synonyms and transcripts to the search index where the platform allows it. Link videos from the relevant procedure, system, or job aid.
This matters because skills are often developed most effectively through practice and support close to the work itself. The CIPD’s guidance on skills development also stresses accessible interventions at the point of need.
Every video should also have an owner, version, review date, and retirement date.
Publish, measure, and revise
Creators rarely regard publication as the end of production. They inspect the response and use it to improve the next video.
L&D teams can use viewing data in the same way.
Viewing data is diagnostic evidence. It is not proof of learning.
A sharp drop could indicate a slow introduction or an irrelevant section. Repeated viewing might signal strong interest, but it might also mean that an explanation was unclear.
A sensible cycle is:
Form a hypothesis → test a version → measure → interpret → revise
What L&D should leave on social media
Some creator techniques do not transfer well into workplace learning.
Clickbait damages trust, particularly when training is compulsory. Artificial suspense delays information that employees needs quickly.
Trends and memes can date long before the underlying procedure changes.
Fast cuts can remove the time needed to inspect a screen, or consider a decision.
Completion is also a poor final measure. Someone can watch every second of a video without learning anything.
The same warning applies to video itself. A performance problem might be caused by a broken process, unclear responsibility, poor system design, or conflicting incentives. Producing a video will not fix those problems.
Borrow the craft, not the incentive system
Social media creators have become very good at earning attention. They understand openings, pacing, personality, visual explanation, discoverability, and rapid iteration.
Those are valuable skills for L&D teams.
The aim, however, is to help someone make a better decision, or perform a task correctly.
The best training video can therefore be the one an employee watches for three minutes, uses immediately, and never needs to watch again.
Leave a Reply