What is microlearning? Definition, formats, and limits

A term without an official definition

Microlearning is not a protected or standardized term. Neither the Federal Institute for Vocational Education and Training nor the German Institute for Adult Education maintains a glossary definition for it, and it does not appear in any German law or technical standard. Anyone who uses it is using a term from the specialist literature and from vendor marketing.

The most robust German-language definition comes from a literature review by Melanie Schall in the Journal of Vocational and Business Education. There, following a definition by Theo Hug, she describes microlearning as technology-supported learning activities of short duration, or as learning with small screens, small learning units, mobile devices, or micro-content. Hug himself describes micro-learning as a relational concept — it is only ever small in relation to something else.

In practical terms: when two people talk about microlearning, they do not necessarily mean the same thing. It is worth defining at the start of a project which duration, which scope, and which objective are meant.

Which characteristics can be substantiated

Duration. The literature review cited gives predominantly 3 to 15 minutes for the studies evaluated, in individual cases up to 20 minutes. That is a description of practice, not a requirement.
Technology-supported. Independence of place and time is regarded in the literature as a defining characteristic, not as an added benefit.
Small scope. Small screens, mobile devices, and micro-content are a constitutive part of most definitions.

Frequently cited characteristics such as "exactly one learning objective per unit" or a built-in repetition logic could not be substantiated as part of the definition in the specialist sources examined. They are didactic conventions of good practice — sensible, but not part of the term.

What learning psychology actually supports

Two effects are substantiated: distributed learning and active retrieval. The forgetting curve is not one of them — and this is precisely where the most exaggeration occurs.

The forgetting curve shows less than is attributed to it

Hermann Ebbinghaus published his studies on memory in 1885. What he examined was the retention of meaningless syllable sequences, in self-experiments, with himself as the only test subject. A replication by Murre and Dros from 2015 confirmed the curve — again with meaningless syllables and again with a single test subject.

The curve is thus a solid finding about the retention of meaningless stimuli. It is not proof that a five-minute video on the General Data Protection Regulation works better than a one-hour seminar. Anyone who cites it as proof of a product's effectiveness is overstretching it.

Spacing and retrieval practice are the robust part

What is well researched, on the other hand, is that distributed learning works. The meta-analysis by Cepeda and colleagues from 2006 evaluates 839 comparisons from 317 experiments; a key finding is that the optimal learning interval increases with the desired retention interval. The testing effect is also well researched: in a meta-analysis on practice testing from 2017, active retrieval performs better than repeated reading.

Both speak in favor of short, recurring units with quizzes. But they speak in favor of a pattern, not a format — a well-timed series of in-person sessions achieves the same effect.

Learning styles are not an argument

The German Institute for Adult Education classifies the widespread four learning types according to Vester as myths of learning psychology. A micro-video therefore does not make sense because it serves "visual learning types". It may make sense for other reasons — this is not one of them.

The common formats

Micro-video — a short clip on exactly one clearly delimited subject.
Flashcard — a question-and-answer pair that can be repeated at intervals.
Quiz — retrieval instead of rereading; this is the implementation of the testing effect.
Simulation — a highly limited, interactive replica of a work situation.
Nudge or push notification — a triggered reminder in the flow of work.
Quick guide at the workplace — retrievable at the moment of need, not learned in advance.

This list is an editorial systematization, not a standard. It helps with selection but does not replace a didactic decision.

How short units become verifiable

A short unit only becomes verifiable when the person, time, and content are logged in an individually attributable way. For mandatory topics, this is the decisive question. If training is to be proven to a supervisory authority or an accident insurance institution, each unit must remain individually attributable — with person, time, and content. The information required for this is described in Proving mandatory training; how strictly a supervisory authority checks this in an individual case is shown in the article on GwG training records.

Technically, the standard determines whether this succeeds. SCORM, in the versions maintained by the ADL Initiative 1.2 and 2004, is designed for browser-based learning content; for units that run outside the browser, such as in a native mobile application, this model reaches its limits. The Experience API, or xAPI for short, is designed as a data and interface standard to capture learning and performance data outside the browser as well; cmi5 is a profile that brings xAPI together with classic learning platforms. The differences are compared in detail in SCORM versus xAPI; the role the platform plays in this is classified in What is an LMS?.

Where microlearning reaches its limits

The limits are a lack of evidence of effectiveness, fragmented learning, and an unsolved transfer problem. The literature review cited names them clearly: representative studies on actual learning outcomes were not found; expectations of the format are spread without questioning its added value. Disjointed, fragmented learning is cited as a risk, as are high demands on the self-discipline and digital competence of learners.

Then there is the transfer problem. According to the study by Michael Gessler on learning transfer in professional development, transfer is only considered to have taken place when what has been learned is applied in the real work context and over a longer period. The learning field and the work field interact — the notion of "train first, then apply" does not hold. A five-minute unit does not solve this problem; it only makes the entry point smaller.

And a note on the data situation: reliable figures on in-company professional development in Germany have become scarce. The official survey of the Federal Statistical Office on professional development in companies is most recently available for the reporting year 2020 — at that time, 77 percent of companies offered professional development — and was subsequently discontinued. There are no official German statistics on microlearning itself. Anyone who reads retention rates or effectiveness percentages should ask about the survey year, sample, and client. Where the figures are missing, the rest of the selection is all the more important: does the format suit the learning objective, and can the result be proven?

How long is a microlearning unit?

In the 2020 literature review by Melanie Schall, predominantly 3 to 15 minutes, in individual cases up to 20 minutes. There is no binding requirement, because the term is neither standardized nor officially defined. The time specification therefore describes what was found in studies, not what a format must fulfill to be called microlearning.

Is microlearning more effective than conventional training?

That cannot be said on the basis of the available sources. A literature review in the Journal of Vocational and Business Education expressly states that representative studies on actual learning outcomes are lacking. What is substantiated are general principles such as distributed learning and active retrieval — but these apply to any format that implements them and are not an argument for short units in particular.

Does the Ebbinghaus curve prove the benefit of microlearning?

No. In series of experiments around 1880, Ebbinghaus studied the retention of meaningless syllables in self-experiments with himself as the only test subject and published the results in 1885; the replication by Murre and Dros from 2015 did the same, likewise with one test subject. The curve thus describes the forgetting of meaningless stimuli and makes no statement about work-related competence development or about the effect of a particular learning product.

Is microlearning suitable for mandatory training?

There is a lot to be said for it when it comes to repeating and refreshing knowledge content, because distributed learning is well researched. The format alone is not designed for instruction with a practical component or for complex practical competence. It is also decisive that every unit remains individually verifiable — otherwise learning activity takes place, but no auditable proof is created.

Which technical standard is suitable for short units?

SCORM 1.2 and SCORM 2004 are tied to the browser context and not designed for native mobile applications. The Experience API is intended for micro-units outside the browser, and cmi5 connects it as a profile with classic learning platforms. Which standard fits depends on what your own platform supports and how granular the proof needs to be.

All sources accessed on 6 Aug 2026 unless otherwise stated in the entry.

Microlearning: definition, formats, and what is really substantiated