Learning & development · What the studies say

Search for the benefits of e-learning and within minutes you will find a dozen impressive percentages. Very few of them survive a check of their sources.

"60 percent faster employee development." "Up to 80 percent higher retention." "Knowledge retention rates of 25 to 60 percent versus 8 to 10 percent in the classroom." These sentences appear on hundreds of vendor pages. Trace them back and you almost always land on another vendor's blog, which in turn cites a third. A verifiable primary study is rarely at the end of the chain.

That is a shame, because e-learning does have demonstrable advantages. They are simply more sober than the marketing suggests. This article separates the two: what current surveys actually show, what learning research has supported for decades, and where the limits are.

What the current figures from the DACH region show

Two surveys provide reliable reference points for German-speaking Europe: the eLearning BENCHMARKING Study by eLearning Journal together with Haufe Akademie, and the mmb Trendmonitor "Learning Delphi" by the mmb Institute, whose twentieth wave was conducted between December 2025 and February 2026.

33.4 %

of training budgets go to digital learning. It is no longer a side channel but a third of the mix.

48.8 %

of companies treat digital learning as a central part of their training strategy; another 31.6 % use it as a supplement.

68.8 %

adoption of microlearning, up 15.8 percentage points. The fastest-growing format.

More interesting than the adoption rate is the direction. Video has overtaken the classic web-based training as the most used format, and short learning units are growing faster than anything else. Both point the same way: away from the multi-hour course, towards units that fit between two meetings.

The four advantages that can actually be evidenced

Spaced learning beats the block session

The spacing effect is one of the best-established findings in the psychology of learning: the same material, spread across several sessions, is retained more reliably than in a single pass. A full training day is hard to split; a course series of ten-minute units is built for exactly that. This is the real didactic advantage of digital formats, and marketing rarely mentions it because it sounds unspectacular.

Retrieval works better than re-reading

The second robust finding is testing as a learning method. Actively retrieving content rather than reading or watching it again leads to markedly longer retention. Digital formats can build this in systematically, through interim questions, decision scenarios and repetition after several weeks. In a classroom session, the questioning usually ends when the day does.

Differentiation by role instead of one size for all

A classroom session has to find a compromise for the room. Half the participants are bored, the other half cannot keep up. Digital courses can be tailored to prior knowledge and job. That is why personalisation sits at the top of the agenda in current surveys, though also as one of the widest gaps between ambition and practice.

Evidence without extra effort

This point gets overlooked in didactic debates but decides everything in regulated areas. Who worked through which content and when is produced as a by-product of digital formats. After a classroom session, at best there is a signature sheet, which says nothing about whether anyone was listening.

Where e-learning is not better

An honest article has to name the other side, otherwise it joins the promotional prose criticised at the start.

Completion rates

The biggest weakness. Open online courses are notorious for completion rates in the low double digits. Mandatory corporate courses sit far higher, but voluntary participation without structure reliably produces courses that are started and never finished.

Negotiating and disagreeing

Where attitudes, conflict or cultural change are at stake, no module replaces a room in which people contradict each other. For knowledge and procedure, digital wins. For negotiation, it does not.

The time myth

Digital learning saves travel time, not learning time. The surveys name lack of time and lack of resources as the biggest obstacles. Without a slot carved out of the working day, no training happens, whatever the format.

Two myths that refuse to die

The so-called learning pyramid, with its fixed retention percentages per learning mode, is unsupported and traces back to a source that was never published. Equally untenable is the division into visual, auditory and kinaesthetic learner types: repeated studies have failed to show any learning benefit from matching material to a supposed type. Both concepts still turn up in training concepts today.

The shift brought by AI

The twentieth wave of the mmb Learning Delphi paints an unambiguous picture: for every question about business opportunities, learning topics or didactically significant technologies, the most common answer was artificial intelligence. Chatbots and AI learning assistants lead the field, while video tutorials and augmented and virtual reality applications lose ground clearly.

What stands out is how sober the experts are. In the previous wave, 93 percent considered the use of large language models by education providers a given within three years. At the same time, 71 percent expected that more time would be needed to check the quality of AI output, and two thirds reckoned that generative AI would also cause skills to be lost. The industry benefiting most from AI is also the one warning most clearly about deskilling.

One finding from the current wave deserves separate mention: compliance is losing importance as a learning topic while AI literacy sits in first place. That is notable, because regulatory requirements have grown over the same period. Both cannot hold at once.

What follows in practice

Anyone introducing or reworking digital training can orient themselves around four points that emerge from the survey data and the learning research together.

Short units instead of full-day events. Not because attention spans have shrunk, which is itself a myth, but because spaced learning demonstrably sticks better and short units are the only kind that fit into a working day at all.

Build in retrieval, do not just play content. Every module needs a point where learners have to decide or recall something. Without it, a course is a video with a progress bar.

Assign time formally. The most common reason programmes fail is not poor content but the absence of an appointment. A reserved slot in the calendar does more than any gamification.

Plan for repetition. One-off training ages, and in regulated subjects it ages within months. Recency is not a comfort feature, it is part of the effect.

Conclusion

E-learning is not inherently superior to classroom training. It is superior for everything that has to be spaced, repeated, differentiated and documented. That covers a large part of corporate training, but not all of it.

The demonstrable advantages do not need invented percentages. They are good enough without them.

Sources

eLearning BENCHMARKING Study, eLearning Journal in cooperation with Haufe Akademie, 2024 and 2025 editions · mmb Trendmonitor "Learning Delphi", mmb Institute, waves 2024/2025 and 2025/2026, the latter surveyed between December 2025 and February 2026 among 77 experts from the education sector.

Last updated: 16 September 2026.

What E-Learning Really Delivers: The Evidence Beyond the Marketing Numbers