Multitasking During Training: How to Win Attention Back

Seven in ten employees are doing something else while your training plays. That is not a guess: the TalentLMS 2026 L&D Report found 70% of employees multitask during training, up from 58% in 2024, the highest figure in three years of tracking. Completion rates will not show you this. Your dashboard says the course was finished; the learner's attention says otherwise.

That gap between attendance and attention is the real training problem in 2026, and it has a fix that has nothing to do with shorter videos.

Why 70% of employees multitask during training

Employees multitask during training because the work does not stop while they learn. The same TalentLMS survey of 1,000 U.S. full-time employees and 101 HR managers found 53% say their workload leaves little room for training, even when it is needed, and 65% report rising performance expectations.

Put those numbers together and the behavior is rational. A learner who is measured on output and handed a 40 minute module will run both at once.

Two more findings from the report sharpen the picture:

• Lack of time has been the top training obstacle for three consecutive years. This is not a fad, it is a structural condition.
• 31% say their training is too theoretical, with not enough hands-on practice. That is the second biggest blocker employees name.

The second point matters more than it looks. Theoretical content is exactly the kind a learner can safely half-watch. Nothing in a lecture-style module ever calls on them, so splitting attention carries no visible penalty.

What divided attention actually costs learning

Divided attention measurably lowers what learners retain, and the penalty spreads to people nearby. The most cited demonstration is Sana, Weston and Cepeda's 2013 study in Computers & Education, which found that students who multitasked on laptops during a lecture took poorer notes and scored lower on comprehension tests, and that students merely seated in view of a multitasking peer also scored lower.

The effect holds up across the wider literature. A 2025 meta-analysis in the International Journal of Cognitive Research in Science, Engineering and Education pooled 11 studies and found a negative correlation between academic media multitasking and achievement of r = -0.25.

That is a modest coefficient, and it is worth being honest about it: multitasking is not a catastrophe on its own. It is a steady tax. Apply a 25% drag across every compliance module, every onboarding session, and every product update your company ships, and the compounding cost is enormous.

There is a compounding effect inside the learner too. Multitasking during training does not just cost the minutes spent looking away. It fragments the encoding of everything around those moments, which is why half-watched training so often produces confident learners who cannot actually do the task.

Why passive video invites multitasking in the first place

Passive video is uniquely easy to multitask through because it never notices you left. A video plays at the same rate whether one person or nobody is watching, and it has no mechanism to detect, interrupt, or respond to a wandering learner.

Compare that with a conversation. You cannot check Slack in the middle of a one-on-one with your manager, because the format demands a response within seconds and the absence of one is obvious to everyone.

This is the design insight most L&D teams miss. Attention is not primarily a willpower problem to be solved with policy. It is a format problem. A format that requires nothing gets nothing.

The strongest evidence for the alternative comes from higher education. Freeman and colleagues' 2014 meta-analysis in PNAS synthesized 225 studies and found that active learning raised exam performance by 0.47 standard deviations over traditional lecturing, while failure rates fell from 33.8% under lecture to 21.8% under active learning. Same content, same instructors, different demand on the learner.

Encouragingly, employees already know this about themselves. 86% say they learn by doing, per the TalentLMS data. The formats they are given simply do not let them.

How to design training that earns attention back

Design for response, not for runtime. Every technique below shares one property: it makes the learner's silence visible, so drifting away has an immediate cost.

1. Open with a question, not an agenda. Ask something the learner has to answer before any content plays. It sets the expectation that this session talks back.
2. Cap uninterrupted playback at roughly two minutes. Not because attention spans are short, but because a two minute ceiling forces you to insert a genuine interaction rather than a decorative one.
3. Make the interaction consequential. A "click next to continue" button is theater. A question whose answer changes what comes next is not.
4. Use free response over multiple choice where you can. Recognizing the right answer from four options is a much weaker demand than producing one, and it is far easier to fake while distracted.
5. Let learners set the pace and the depth. Someone who already knows the material should be able to prove it and move on. Forcing them through content they have mastered is the fastest way to teach them that multitasking is safe.
6. Give feedback in the moment. Feedback delivered at the end of a module lands on a learner who has already mentally left.

Adaptive interactive training videos build this in structurally: an AI tutor asks questions mid-lesson, waits for a real answer, and branches based on what the learner actually said. The learner cannot idle through it, because the session does not advance without them.

How to tell whether attention actually improved

Stop treating completion as an engagement metric and start measuring response behavior. Completion tells you a video reached its end, which is precisely the number multitasking learners keep looking healthy.

Track these instead:

• Response rate per prompt. What share of interaction points got a real answer rather than a skip or a one word dismissal?
• Median response latency. Answers that arrive suspiciously fast or suspiciously slow both signal a learner who is elsewhere.
• Answer quality over the session. Open response depth that degrades from chapter one to chapter five is attention draining in real time, visible while you can still act on it.
• Post-training task performance at 30 days. The only measure that matters in the end, and the one that exposes courses everybody "completed."

One caution on interpretation. A drop in your completion rate after you make training more demanding is not necessarily a failure. It often means the course finally stopped rewarding people for sitting through it.

FAQ

Is multitasking during training always a problem, or does it depend on the content?

It depends on the demand the content places on working memory. A learner can reasonably skim a policy update while answering email, because nothing complex is being encoded. But anything involving a new skill, a judgment call, or a procedure they will have to reproduce later deserves undivided attention, and the research suggests you will not get it unless the format requires it.

Should we just ban phones and laptops during training sessions?

Bans treat the symptom and tend to fail with adults who have real work waiting. The TalentLMS data shows the underlying cause is workload, not indiscipline: 53% of employees say there is genuinely little room for training. A better move is to shorten the session, make every minute demand a response, and get explicit manager agreement on protected learning time.

How short does training need to be to hold attention?

Length is the wrong variable to optimize. A 30 minute conversation that keeps asking you questions holds attention better than a 5 minute video that asks nothing. Focus on interaction density, meaning how often the learner has to actually produce something, and let the runtime follow from the content.

Attention is the scarcest resource in corporate learning, and the 70% multitasking figure says the current formats are losing the competition for it. The fix is not more policy or shorter runtimes, it is training that notices when a learner stops participating and does something about it. Build learning that talks back, and attention stops being something you have to ask for.