How Long Should a Microlearning Module Be?

How long should a microlearning module be?

Three to seven minutes, covering one objective. That is the working answer, and everything else on this page is about why the number matters less than the constraint behind it.

The honest version of the research

You will see "attention spans are eight seconds" quoted constantly. It is not a real finding about learning, and you should ignore it. What is well supported is narrower and more useful:

• Working memory is limited. People hold a handful of new items at once. Overload it and retention collapses regardless of how long the module is.
• Spacing beats massing. The same content split across several sessions is retained better than one long block.
• Completion drops with length. Not because of attention span, but because a 25-minute module needs a diary slot and a 5-minute one does not.

None of that gives you a magic number. It gives you a design rule: one objective, minimum viable content, then stop.

Practical lengths by purpose

If a module runs past about ten minutes, it almost always contains two objectives. Splitting it is nearly always better than trimming it.

Use word count, not minutes, while you are writing

Minutes are hard to judge in an editor. Reading pace for learning content is roughly 130 to 150 words per minute, slower than casual reading. So:

• 3 minutes is around 400-450 words
• 5 minutes is around 650-750 words
• 7 minutes is around 900-1,000 words

Add time for video, interaction and knowledge checks on top - a 60-second video and three questions add two to three minutes to those figures.

The real constraint is the objective, not the clock

Write the single objective first. "Explain when to escalate a safeguarding concern." Then include only what is needed to meet it, and assess exactly that. If content does not serve the objective, it belongs in another module or in a reference document.

This is why a five-minute module built properly outperforms a five-minute module produced by cutting a thirty-minute one in half. The second one is a fragment; the first is complete.

Where compliance overrides the guidance

Some statutory training carries an expected duration, and some auditors treat time-on-task as evidence. If your regulator or insurer expects a certain length, meet it - but structure it as a sequence of short objective-led sections rather than one continuous run.

Building at this length

The reason organisations default to long modules is that short ones are inefficient to produce by hand: the setup cost per module is the same whether it is four minutes or forty. AI generation removes that overhead, which makes splitting one topic into six focused modules a reasonable thing to do rather than six times the work.

If you want to try it, take an existing long course and rebuild it as a sequence during a free trial. Related reading: microlearning vs traditional eLearning and how to convert an existing course to microlearning.