Your Safety Training Isn't Failing. It's Fading.

In May, LUMA1 published The Three Principles, an industry study by John Hudson that brings together decades of research on what makes safety training work. Its first principle is the simplest of the three, and probably the one most programs are furthest from: short and frequent beats long and rare. 

Here's what that means in practice, and why the way most teams schedule training works against it. 

Memory Doesn't Follow a Training Calendar 

People forget. That isn't a criticism of workers. It's how memory works, and it has been studied for more than a century in fields from medicine to aviation to manufacturing. The finding is the same everywhere: information that isn't revisited fades. 

The study puts it plainly. A two-hour orientation on day one fades meaningfully within weeks. An annual refresher fades within months. Across the research on adult learning, spaced repetition (short sessions repeated over time) outperforms one-time delivery on every outcome that has been tested. 

That's a problem for any program built around one big session up front and a refresher once a year. On paper, everyone is trained. In practice, much of what they learned has faded by the time they need it. 

Orientation and Toolbox Talks Are One System, Not Two 

One of the study's sharpest points is about how training gets organized. In most companies, orientation and toolbox talks are separate line items, with separate owners, separate budgets and separate schedules. 

The research says they're two halves of the same thing. Orientation sets the baseline when someone arrives on a new site or starts a new scope of work. Toolbox talks keep that baseline from fading. Without orientation, the talks have nothing to build on. Without the talks, the orientation doesn't last. Treating them separately is one of the most consistent operating errors seen in this industry. 

Weekly Safety Training Isn't Often Enough 

This is the uncomfortable part. Many teams run toolbox talks weekly, every other week, or monthly. The study's view is that memory doesn't respect those schedules: a hazard that isn't reinforced within the working week isn't front of mind when the work actually happens. 

The approach with the strongest evidence behind it is daily, task-specific reinforcement lasting two to five minutes, usually a supervisor briefing at the start of the shift supported by a short digital prompt. That doesn't mean more training overall. It means smaller pieces, more often, tied to the work people are doing that day. 

It matters most for people who are new to a site. According to the US Bureau of Labor Statistics, more than 60% of construction accidents happen within a worker's first year with their current employer. Those are the workers whose orientation is most recent, and who have had the least time on site to reinforce it. 

Making Small and Often the Default 

LUMA1 is built around this kind of cadence. Training can be sequenced automatically. For example, a site-specific briefing on day one, a reinforcement a few days later, and a targeted briefing when someone moves to a new area or scope. Frequent reinforcement doesn't depend on someone remembering to send it. It runs on any device, in more than 40 languages, and records whether each worker understood the content, not just whether they opened it. 

To see how it works in practice, book a demo at luma1.com. You can read the full study here: The Three Principles