When Should Practice Happen? Timing and Spacing for Adult Skill Learning

85% versus 33%. One classic spacing study varied the gap between practice sessions from seven minutes up to 24 hours, then measured retention several weeks later. Learners who practiced 12 hours apart held onto 85% of what they'd learned. Learners who practiced 24 hours apart kept about a third. Same material, same amount of practice – the only variable was the clock.

So when should practice happen? The research answer is specific: apply a new skill inside the training session itself, review it within about 12 hours, then space follow-up practice at roughly 10-20% of however long you need the skill to last. That approach – spaced practice, also called distributed practice – reliably beats massed practice, where all the repetition happens in one sitting. The underlying phenomenon is the spacing effect, and it's one of the oldest and most replicated findings in learning science.

I'm a product manager, not a memory researcher. But before Adam White and I designed our first Conflict Campaign scenario – Adam's the licensed counselor, I'm the product guy – I synthesized more than 200 studies on instructional design, and the timing research changed our plans more than almost anything else in the corpus. Most workplace training already defaults to 10-15% transfer. Timing is one of the levers that moves that number, and unlike most of the levers, it's close to free.

Why does spaced practice beat massed practice?

Spaced practice wins because consolidation happens between sessions, not during them. In the rest periods between practice, the brain reorganizes and stabilizes new information, and sleep reactivates and strengthens the neural pathways involved. When the next session arrives, the learner has to retrieve the skill rather than just repeat it – and retrieval after partial forgetting is exactly the kind of effort that makes memory durable.

Massed practice feels better while it's happening. Performance climbs steadily across a cram session, everyone leaves competent, and the trainer gets good reviews. Then the curve does what curves do. The spacing advantage holds across wildly different training domains – for interpersonal skills specifically, the research cites benefits in interviewing, surgical skills, and interpersonal simulations.

The shape of the schedule matters less than you'd expect. Equal intervals (12 hours, 12 hours, 12 hours) and expanding intervals (1 hour, 5 hours, 15 hours) both outperform massed practice, with equal spacing often showing slightly better recognition gains. Don't agonize over the pattern. Agonize over whether the second session exists at all.

How long should the gap between practice sessions be?

Roughly 10-20% of your retention target. That's the guideline the meta-analytic work supports: the optimal gap before review grows in proportion to how long you need the skill to last. Practicing for a presentation next Friday calls for a different schedule than building a skill your team should still have next year.

The adult-learning research translates the percentage into concrete schedules. For a one-week retention target, review at day 1 and day 4. For a one-month target, review at days 2, 7, and 20. Expanding schedules – practice at 1 hour, then 1 day, then 1 week – carry meta-analytic support as well.

Two timing details deserve special attention. Get an application rep in before anyone leaves the room, since adult learners cement skills most effectively when they try the technique within minutes of instruction, while attention is at its peak. Then schedule the first review within 12 hours – a brief quiz, a reflection prompt, or a micro-exercise is enough to catch the early consolidation window.

How fast do interpersonal skills decay without practice?

Faster than most other skill types, and on a fairly predictable schedule. Arthur, Bennett, Stanush, and McNelly's meta-analysis of skill decay – 53 articles, 189 data points – found losses ranging from an effect size of d = -0.01 immediately after training to d = -1.4 after more than 365 days of non-use, with the steepest decline in the first weeks and months. The same analysis showed which skills fall fastest: physical and speed-based tasks resist decay reasonably well, while cognitively complex, accuracy-based tasks lose ground quickest. Reading a tense conversation and choosing the right response is both cognitively complex and accuracy-dependent.

For interpersonal skills, substantial degradation begins within 30 to 90 days of training. That's the danger zone: far enough out that the training-day energy has dissipated, close enough that nobody has scheduled a refresher yet. Without reinforcement, the extrapolated half-life for conflict skills runs about 3-6 months.

Experience buffers some of this. Research on safety-critical professions found that novices who didn't practice after initial training decayed faster than those who did, and that people with more prior experience showed greater resilience. If your team was new to structured conflict skills when they trained, the follow-up calendar matters even more than usual.

What should a booster session contain?

Retrieval and application – not a replay of the slides. The point of a booster is to force the skill back out of memory and put it to work, which is a different activity than hearing the content again. A U.S. Army Research Institute review of interpersonal skills training put the stakes plainly: there's little reason to believe social skills training can be accomplished as a one-shot procedure with no follow-up.

A booster earns its calendar slot with three moves. Open with retrieval: people reconstruct the framework or walk through the technique cold, before any refresher material appears. Move to live practice, ideally on a scenario slightly harder than the original – Lacerenza and colleagues' meta-analysis of 335 leadership training studies found properly designed programs produced a 28% increase in on-the-job leadership behaviors, and graduated complexity is part of proper design. Close with a short debrief that ends in specific commitments about where the skill gets used next.

Reconvening the original training cohort makes the booster do double duty. Hughes, Zajac, Woods, and Salas's meta-analysis found peer support was the strongest environmental predictor of transfer – it explained more variance than supervisor or organizational support in a model accounting for 32% of transfer overall. Salas, Tannenbaum, Kraiger, and Smith-Jentsch's broader point applies here too: training works as a system, and what happens after the session counts as much as the session itself. A booster isn't remediation. It's the second half of the original design.

Does practicing past mastery help?

Yes. Overlearning – continuing to practice after learners first demonstrate competence – measurably slows decay. Arthur and colleagues found the degree of initial overlearning is negatively related to skill loss, with a moderate overall effect on retention. The practical implication runs against instinct: the moment someone finally handles the hard conversation well is the moment to run it once more, not the moment to celebrate and break for lunch.

That instinct to stop early is strong. It's why we built Operation Aetherfall's key encounters to be run more than once – the first success proves the skill exists, and the repetition after it is what the retention research actually pays off.

What should your team do at 1 week, 1 month, and 1 quarter?

Suppose your team ran a training day yesterday – a workshop, a simulation, a tabletop scenario. Here's a spacing plan the research supports, and it adds maybe three or four hours of total time across the quarter.

When	What to do	Why this timing
Same day, within 12 hours	A reflection prompt or 10-minute micro-exercise; everyone writes one if-then plan for using the skill	First review lands inside the early consolidation window
1 week later	A 30-60 minute retrieval session: reconstruct the framework cold, then each person applies it to one live situation; check in on the if-then plans	Catches the steepest stretch of the forgetting curve
1 month later	A booster: practice scenario slightly harder than the original, short debrief, new commitments	Lands inside the 30-90 day window where interpersonal skill decay begins
1 quarter later	A full practice run on a new, more complex scenario; compare results against your baseline	Unreinforced conflict skills have a rough half-life of 3-6 months; harder scenarios build schemas that resist decay

The if-then plans in the first row are implementation intentions – the most evidence-backed bridge between a training room and a Monday standup. The one-week session is also the natural moment to collect data; a one-week follow-up survey takes minutes and tells you whether anything is actually being used.

Notice what the plan doesn't require: new content, a hired trainer, or another full day. Every entry is retrieval, application, or peer contact – the ingredients the transfer literature keeps rewarding – arranged on an expanding schedule that tracks how forgetting actually behaves.

Here's the stance I'll defend to any skeptical budget owner: a shorter training day plus this follow-up plan beats a longer, richer training day with nothing after it. The single-event model asks one day to do a job the memory system won't let one day do.

Forgetting runs on a schedule. Practice should too.