Microlearning

Microlearning: Why Two-Minute Videos Miss the Point

Short isn't the active ingredient. Only 8% of viewers want instructional videos under three minutes, and the research credits spacing, recall and feedback. Here's what works, and how to build it.

A bearded man in a grey jumper, seen from behind, taps through a lesson on a tablet at a white timber table, a cup of black coffee beside him
A few minutes, one idea, a cup of coffee: the shape most people picture when they hear microlearning.Photo: Karolina Grabowska, Kaboompics
The short answer

Microlearning is teaching in short, focused pieces, usually a few minutes each, built around one task or idea. It works best as a way to deliver spaced practice and frequent recall with feedback, which have strong research support. Shortness alone isn't the active ingredient, so design each piece around a question, space the pieces out, and measure changes in behaviour.

Key takeaways

  • Two in three organisations in ATD's 2025 benchmark already use microlearning, while formal learning hours per employee fell from 35 in 2020 to 13.7 in 2024.
  • The strongest support for microlearning comes from spacing: a 2024 meta-analysis found spaced digital education beat massed delivery for knowledge, skills and clinical behaviour.
  • Shortness isn't the active ingredient. Short pieces help because they make spacing, recall and feedback easy to schedule.
  • Short doesn't mean tiny: in TechSmith's 2024 survey, only 8% of viewers preferred instructional videos under three minutes.
  • Most microlearning studies stop at knowledge tests; judge yours by whether people do the task differently.

The two-minute video is the face of microlearning, and it’s the part the evidence supports least. Microlearning is teaching in short, focused pieces, usually a few minutes each, built around one task or idea. It works when those pieces carry the things the research supports: spaced practice (coming back to an idea over days and weeks), recall (answering from memory rather than re-reading) and feedback. Cutting a long course into short videos, on its own, does little. Design each piece around a question, space the pieces out, and judge the result by what people do differently.

That’s the short version. Most of the guides ranking for “microlearning” are written by companies that sell microlearning software, and they lean on engagement figures that rarely link to a study. This guide starts from the research and the latest industry benchmarks, then turns them into seven design steps with examples. For the definition and the history of the term, see our library entry, Microlearning defined.

66%

of organisations used microlearning and short-form learning in 2024

ATD, 2025 State of the Industry, p. 41

13.7

formal learning hours per employee in 2024, down from 35 in 2020

ATD, 2025 State of the Industry, p. 10

70%

of US employees multitask during training

TalentLMS, 2026 report, p. 15

8%

of viewers prefer instructional videos under three minutes

TechSmith, 2024 study, p. 15

Why microlearning is everywhere now

The simplest reason is time. The Association for Talent Development (ATD) surveys hundreds of organisations each year about their training. In its benchmarks, the formal learning hours each employee used fell year after year, then turned.

Formal learning hours per employee fell by more than half in four years, then edged up

Average formal learning hours used per employee a year, organisations in ATD's State of the Industry surveys

Interactive chart. The figures are in the table below.

Formal learning hours fell by more than half. Organisations in ATD’s surveys reported 35 hours of formal learning per employee in 2020 and 13.7 hours in 2024. Formal learning means scheduled courses and programmes; on-the-job and informal learning aren’t counted.
Then they edged up. In ATD’s 2026 report, covering 2025, hours rose to 16.7 per employee, “for the first time in five years”. Over the same year, direct learning spend per employee fell to USD 846, so the extra hours came with less money behind them.

Each year is a different group of organisations, so treat the line as a direction rather than a precise trend. The direction matches what employees say. In TalentLMS’s 2026 benchmark of 1,000 US employees, 70% said they multitask during training, a three-year high, and 53% said high workloads leave little room for it (TalentLMS, 2026). We looked at that squeeze in more detail in the workplace learning paradox.

Microlearning is the industry’s answer, and it’s already mainstream:

Two in three organisations already use microlearning; one in five more are considering it

Share of organisations using or considering each technology-based learning method, 2024

Interactive chart. The figures are in the table below.

Show the data
UsedConsidering
Simulations and scenarios69%20%
Microlearning and short-form66%22%
Podcasts and videos63%19%
Immersive learning55%29%
AI embedded in learning40%36%
Games and gamification31%33%
539 organisations reported 2024 data. ATD's category label is "Microlearning and short-form learning". Source: ATD, 2025 State of the Industry: Talent Development Benchmarks and Trends, page 41.

Two in three organisations used microlearning in 2024, and another 22% were considering it. The fastest movement is in AI: 40% already embed it in learning and 36% more are considering it. That matters for microlearning because AI tools can now generate short quizzes, flashcards and scenario questions in minutes, which makes it cheap to produce many small pieces. Cheap to produce isn’t the same as good to learn from, which is where the evidence comes in.

What the research supports, and what it doesn’t

There’s less direct research on “microlearning” than the market suggests, and the best of it points to a clear conclusion: the shape helps because of what it makes easy, not because short is better.

Spacing has the strongest support. Spaced learning means returning to the same material in several sessions with gaps between them, instead of covering it all at once (massed learning). A 2024 meta-analysis pooled 23 randomised and quasi-randomised trials of spaced digital education for doctors, nurses and other health professionals (Martinengo et al., 2024). Spacing came out ahead on knowledge, skills and what clinicians did in practice:

Spacing digital lessons out beat cramming them together, most of all for skills and behaviour

Pooled effect of spaced versus massed digital education for health professionals, with 95% confidence intervals, from a meta-analysis of 23 trials (2024). Clinical behaviour compares spaced online education with massed delivery or no training · standardised mean difference (SMD)

0Knowledge, online (9 studies)0.32Clinical behaviour (3 studies)0.67Surgical skills, simulation (2 studies)1.15
Source: Martinengo, L., Ng, M. S. P., Ng, T. D. R., Ang, Y.-I., Jabir, A. I., Kyaw, B. M., & Tudor Car, L. (2024), Figures 3, 6 and 8 (pooled estimates, also reported in the abstract). doi.org/10.2196/57760

The effect on knowledge was small, the effects on skills and on what clinicians did in practice were larger, and the confidence intervals are wide where few studies exist. The authors rated most studies (19 of 23) at unclear or high risk of bias, so this is moderate evidence, not proof. What’s useful for designers is how the spaced programmes were built: in 15 of the 17 online studies, each piece was a case or small learning unit followed by one or more multiple-choice questions, most often delivered by email. That’s microlearning, but microlearning with a question in every piece.

The gap between sessions matters, and the right gap grows with how long people need to remember. A synthesis of 317 experiments on spaced practice found exactly that (Cepeda et al., 2006):

In plain terms (ISI is the gap between study sessions): if people need something next week, review it over a few days; if they need it next year, spread the reviews over weeks and months.

Recall beats re-reading. A meta-analysis of the testing effect found that practising recall produced better retention than restudying the same material, with larger benefits from recall questions than from recognition ones (Rowland, 2014). Short pieces make it easy to put a question in front of learners often.

Shortness alone has weak support. A scoping review of 17 microlearning studies in health professions education found positive effects on knowledge and confidence, but none of the studies defined microlearning, and few looked beyond the test (De Gagne et al., 2019):

None of the 17 measured results for patients or organisations. The popular six-minute video rule has a similar problem: it came from a study of how long people watched, and its authors warned against drawing conclusions about learning from it. Our guide to how long course videos should be covers that study in detail.

How short is short?

Microlearning is often sold as two-minute videos. Viewers don’t seem to want that for instruction. TechSmith, which makes video software, asked 929 people in six countries, Australia among them, how long they prefer instructional videos to be.

Only 8% of viewers prefer instructional videos under three minutes

Preferred length of instructional videos, share of 929 viewers in six countries, May 2024

Interactive chart. The figures are in the table below.

Preferences spread across a wide range. Most people prefer instructional videos somewhere between three and 19 minutes.
Very short is a minority taste. Only 8% preferred instructional videos under three minutes, and almost nobody chose under 30 seconds.
The single most popular choice was 10 to 19 minutes, picked by 26%. At the other end, 60% of viewers said they would not watch a video of 20 minutes or longer.

This is a survey of preferences, not a test of learning, and a vendor’s survey at that. But it’s a useful corrective. “Micro” should describe the scope of a piece, one task or idea, not a target running time. A recall question takes 30 seconds; a worked example of a tricky procedure might take eight minutes. Both can be microlearning.

How to design microlearning that works: seven steps

1. Start from a task, not a topic

Write down what people need to do differently, in observable terms. “Understand our returns policy” is a topic. “Process a return for a damaged item in under three minutes without calling a supervisor” is a task. Microlearning suits tasks that recur, have clear right answers or common mistakes, and are easy to forget between uses.

Example: a retail team keeps miscoding damaged returns. The microlearning goal is the coding decision, not “returns training”.

2. Give each piece one objective

Each piece should do one job: teach one step, practise one decision, correct one misconception. If you can’t name the piece’s single objective in a sentence, split it. Our guide to writing learning objectives covers how to word them.

Example: “Choose the right return code for a damaged item” is one piece. “Explain the returns process” is five.

3. Make most pieces ask a question

The spaced programmes in the 2024 meta-analysis almost all paired a small case with questions. Aim for most pieces to ask learners to recall, decide or apply, with explanation only where it’s needed to answer. Scenario questions (“A customer hands you this. What do you do?”) beat fact questions for workplace skills, because they practise the decision itself.

Example: a 60-second piece shows a photo of a damaged item and the receipt, then asks which code to use.

4. Space the pieces out

Schedule pieces across days and weeks, and bring back earlier ideas instead of moving steadily forward. Following Cepeda and colleagues’ finding, set the gaps by how long people need to remember: days for something needed next week, widening to weeks for something needed all year. Adaptive tools that bring back the questions a learner got wrong sooner, and the ones they got right later, apply the same idea automatically.

Example: the return-code question comes back on day 2, day 7 and day 21, with a fresh photo each time so learners can’t memorise the picture.

5. Give feedback that explains

After each answer, say whether it was right and why, briefly. Feedback that explains the reasoning helps learners more than a tick or a cross. Our library entry on effective feedback summarises the evidence.

Example: “Code D2: the damage happened in transit, so it’s the carrier’s claim, not a store write-off.”

6. Keep the pieces connected to the whole

Short pieces are easy to lose in. Show learners where each one fits (a simple map or progress view), link related pieces, and keep longer formats for what needs sustained attention: a new mental model, a complex procedure, a discussion. Microlearning is often best as the practice and review layer around a longer course, not as a replacement for it.

Short videos are often sold as the answer for learners raised on social media, but students tend to skip shorter videos or treat them as less important. Surprisingly, we find the most engagement and the highest retention (assessed through quizzes and reflection activities afterwards) with videos that sit in the 8-10 minute mark. That might be because a video 8 or 10 minutes long is seen as something a learner shouldn’t miss.

Example: a 40-minute workshop introduces the returns system; microlearning handles the practice for the next month.

7. Measure behaviour, not views

Completion rates and opens tell you people saw the pieces. Decide before launch what change you expect to see in the work, and check it: fewer miscoded returns, faster handling, fewer escalations. Track question accuracy over time too. If people keep getting the same question wrong on its third return, the piece needs rewriting, not repeating.

Example: the team tracks the share of returns coded correctly in the month before and the two months after.

Microlearning examples, by format

FormatExampleWhat makes it work
Spaced scenario questionsA short case by email or app every two days, with feedback after each answerSpacing, recall and feedback in one piece
FlashcardsA spaced set of key terms or drug doses, with the ones you miss coming back soonerRecall on a schedule that adapts
Demo plus practiceA 90-second screen recording of one task, then a task to try straight awayOne objective, immediate application
Worked exampleA short walk-through of one solved problem, then a similar problem to solveShows the thinking before asking for it
Job aid with self-checkA one-page checklist used on the job, with three questions at the endSupport at the moment of need
Reflection promptA question after a real task: what went well, what would you change?Connects learning to the work

When microlearning is the wrong tool

Microlearning is a poor fit when learners need to build a new, connected understanding from scratch, when a skill has many interdependent steps that only make sense together, or when the goal is discussion and practice with other people. It also can’t fix a lack of time to apply what’s learned: if workloads leave no room for practice, shorter lessons won’t create it. In those cases, use a longer lesson or a live session for the core, and microlearning for the spaced practice afterwards.

What to do next

Pick one task your learners keep getting wrong. Write three scenario questions about it, each with a short explanation of the right answer, and schedule them to reach learners two, seven and 21 days after the main lesson. Compare their answers on day 21 with day 2, and check the task itself a month later.

Check your understanding

Three quick questions on this article. Recalling it now helps it stick.

Answer to see the results

1. According to the research in this guide, what mainly makes microlearning work?
2. In TechSmith's 2024 survey of 929 viewers, which length of instructional video was the most popular preference?
3. What's the best evidence that a microlearning programme worked?
0 of 3 answered

Questions people ask

What is microlearning?

Microlearning is teaching in short, self-contained pieces, each focused on one idea, task or skill. A piece might be a two-minute video, a three-question quiz, a worked example, a flashcard set or a short scenario. There's no agreed length: researchers describe anything from a few seconds to about 15 minutes. What matters more is that each piece asks the learner to do something and connects to a larger goal.

What are some examples of microlearning?

Common examples are a daily scenario question sent by email or app, a spaced flashcard set, a 90-second demonstration followed by a practice task, a one-page job aid with a quick self-check, and a short case with a decision to make. The best ones ask learners to recall or apply something, give feedback on their answer, and return to the same idea a few days later.

Does microlearning actually work?

Early evidence is positive, mostly for knowledge and confidence, and stronger when short pieces are spaced out. A 2024 meta-analysis of 23 trials in health professions found spaced digital education beat massed delivery for knowledge, skills and clinical behaviour, though most studies carried an unclear or high risk of bias. Evidence that shortness alone improves learning is weak.

How long should a microlearning lesson be?

Long enough to finish one task or idea, and no longer. For a recall question or flashcard, that's under a minute. For a demonstration or worked example, a few minutes. Viewer surveys suggest people don't want instructional videos to be tiny: in TechSmith's 2024 study, only 8% preferred instructional videos under three minutes. Split by idea, not by the clock.

What's the difference between microlearning and eLearning?

eLearning is any learning delivered digitally, from a two-hour module to a full online degree. Microlearning is one way of shaping it: small, standalone pieces, usually taken in a few minutes and often spread over days or weeks. A course can mix both, with longer lessons for complex skills and microlearning for practice, review and reminders.

What are microlearning best practices?

Start from a task people need to do, give each piece one objective, make most pieces ask a question rather than present information, space the pieces out over days and weeks, give feedback that explains the answer, keep the pieces connected to a larger goal, and measure whether behaviour changes. Avoid cutting a long lecture into short videos and calling it microlearning.

Sources

  1. Martinengo, L., Ng, M. S. P., Ng, T. D. R., Ang, Y.-I., Jabir, A. I., Kyaw, B. M., & Tudor Car, L. (2024). Spaced digital education for health professionals: Systematic review and meta-analysis. Journal of Medical Internet Research, 26, e57760. https://doi.org/10.2196/57760
  2. Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380. https://doi.org/10.1037/0033-2909.132.3.354
  3. De Gagne, J. C., Park, H. K., Hall, K., Woodward, A., Yamane, S., & Kim, S. S. (2019). Microlearning in health professions education: Scoping review. JMIR Medical Education, 5(2), e13997. https://doi.org/10.2196/13997
  4. Association for Talent Development (2025). 2025 State of the Industry: Talent development benchmarks and trends. ATD Research. https://assets.td.org/m/483691e33fbdf4db/original/State_of_the_Industry_PDF_Download_2025.pdf
  5. Association for Talent Development (2026). Highlights from ATD's 2026 State of the Industry report. ATD blog, 11 May 2026. https://www.td.org/content/atd-blog/highlights-from-atd-s-2026-state-of-the-industry-report
  6. TalentLMS (2026). 2026 Annual L&D Benchmark Report. TalentLMS. https://www.talentlms.com/research/learning-development-report-2026
  7. TechSmith (2024). 2024 Video Viewer Study. TechSmith. https://assets.techsmith.com/docs/TechSmith-Video-Viewer-Study-2024.pdf
  8. Rowland, C. A. (2014). The effect of testing versus restudy on retention: A meta-analytic review of the testing effect. Psychological Bulletin, 140(6), 1432–1463. https://doi.org/10.1037/a0037559
Jim Lounsbury

Learning designer, documentary filmmaker and founder of The Learning Curve. Jim designs microcredentials and online courses for universities and businesses. More about Jim

Read next