The forgetting curve in corporate training: what Ebbinghaus did and didn't show

Part of The science, honestly.

The slide everyone has seen

If you work in L&D, you've probably seen a curve dropping off a cliff, labelled "the forgetting curve", with a claim that people forget most of a training session within a day. It's usually credited to Hermann Ebbinghaus. The real study is more interesting, and more useful, than the slide.

What Ebbinghaus actually did

Ebbinghaus published his memory experiments in 1885; the standard English translation appeared in 1913 [1]. He was his own and only subject. He learned series of nonsense syllables (made-up consonant-vowel-consonant combinations), so that meaning and prior knowledge wouldn't help him.

He didn't measure how much he could recall. He measured savings: how much less time it took to relearn a series after a delay, compared with learning it the first time. The more time he saved, the more had been retained.

His results showed a clear shape. In his own words, forgetting was very rapid at the beginning and very slow at the end [1]. After a day, relearning a series still took about two-thirds of the original effort; in savings terms, about a third of the original effort was saved after 24 hours, about a quarter after six days, and about a fifth after a month [1].

What it doesn't show

That is a long way from "people forget 70% of training in 24 hours". Three things get lost on the way to the slide:

1. One person. Ebbinghaus tested himself. It was careful work, but it was one learner.
2. Nonsense material. He chose syllables with no meaning on purpose. Workplace training is the opposite: it is meant to connect to what people already know and do.
3. Relearning time, not recall. His "amount forgotten" column is relearning time not saved. It isn't the share of material that someone could no longer remember.

So the curve tells us something real about the shape of forgetting. It doesn't give a percentage you can apply to a sales course or a clinical procedure.

The replication

In 2015, Murre and Dros repeated the experiment with one learner, using Ebbinghaus's method and similar intervals, from 20 minutes to 31 days [2]. They concluded that the forgetting curve had been replicated, with one wrinkle: the curve isn't completely smooth, and probably jumps upwards at the 24-hour point. Most of the loss came early, as before.

They were also open about the limits. It was again a single-subject study, and in their own recall data the curves were much shallower than the savings curves [2]. In other words, how you measure forgetting changes how steep it looks.

What spacing adds

If forgetting is steep at first, the obvious question is when to revisit material. A 2006 meta-analysis by Cepeda and colleagues pulled together 839 assessments of spaced practice from 317 experiments [3]. Its main finding, in the abstract, is that the best gap between study sessions depends on how long you need to remember: the longer the retention interval, the longer the best gap.

Two cautions. That review covered verbal recall tasks, such as word lists and facts, not complex workplace skills. And I've checked it at the level of the published abstract, not the full paper. It supports spacing as a design principle. It doesn't give a timetable for your programme.

What this means for your programmes

• Don't quote a percentage. "Most is lost early" is supported. "70% in a day" isn't.
• Plan the second contact, not just the first. If the early drop is steep, a single event with nothing after it is a weak design.
• Space practice to match the need. Material people must use next week needs a different rhythm from material they must still have in a year.
• Measure use, not memory alone. In corporate training, the question isn't only "can they recall it?" but "are they doing it?" That needs a check weeks later, not a quiz at the end of the session.

Julian's workshop Forgotten by Friday: designing training for brains that forget rebuilds one real programme around these ideas.

General information on training design, not professional advice for your organisation.