Why Trainings Don't Stick
And what the research says actually works.
It’s the beginning of the year. Maven courses are in full swing. Professionals are resolving to improve their skills, joining cohorts, and paying thousands of dollars. I’m one of them.
Maven instructors have made over $20M combined. Some earn $1M+ a year working 10-15 hours a week.
The course economy is booming. But here’s what I keep thinking about:
I’ve taken dozens of courses in my career. As a coach, I run workshops with teams on product strategy and mindset. And I’ve watched the same pattern play out over and over.
Years ago, my company brought in a well-known product consultant for a two-day workshop. Someone who literally wrote the book. The leadership team showed up. They participated. They asked good questions.
I left that hotel ballroom hopeful. They get it. We’re going to work differently.
A colleague talked about “delighters” for weeks. It felt like something had shifted.
Within the year, those same leaders introduced a heavyweight planning process. Quarterly ceremonies. Hardened commitments. The opposite of what we’d just learned.
I don’t think it was malicious. They genuinely believed they wanted transformation. But when transformation meant giving up control, they chose a framework that looked modern while keeping the old power dynamics.
They didn’t want product thinking. They wanted to have done product thinking.
In the courses I’ve taken and the ones I’ve led, one thing keeps happening: people show up, feel energized, and the trainer feels accomplished. Then nothing changes.
This is why I started one-on-one coaching. I kept seeing this pattern with my teams. The workshops were well received, but when people went back to their day jobs, nothing stuck. I have the advantage of being embedded with the teams I coach. I see what happens after. An outside consultant doesn’t have that view. They have to believe the training worked. I’m not sure how many go back to check.
So I’ve been thinking about why trainings don’t stick. That led me to some research.
The stat everyone cites
You’ve probably heard it: “Only 10% of training transfers to the job.”
It sounds rigorous. It gets cited everywhere. It makes you nod and think, yeah, that tracks.
Here’s the thing: no one can find where it came from. It’s been repeated so many times that everyone assumes it’s real. But the original source doesn’t exist.
For a field obsessed with data, we’ve been running transformations on vibes.
So what does the research actually say?
What the research says
In 1988, two researchers named Baldwin and Ford asked a simple question: Why doesn’t training transfer to the job?
They found three factors that matter:
Who’s in the room — trainee characteristics
How good the course is — training design
Whether you get to use it — work environment
All three matter. But the research found that one matters more than the others.
It’s not who’s in the room. It’s not how good the course is.
The biggest factor is whether your environment gives you the chance to apply what you learned.
Give someone a framework and no authority to use it, and you’ve wasted everyone’s time.
This isn’t just one study. In 2010, a meta-analysis of 89 studies confirmed the same thing: work environment and supervisor support are among the strongest predictors of whether training sticks.
Looking back at that workshop, I get it now. It was never going to stick. Not because the training was bad. Because we never created an environment for it to survive.
So what actually works?
1. Find other believers
One person modeling different behavior converts almost no one. Research suggests you need about 25% of a group to believe in something before it tips.
The question isn’t “how do I change the culture?” The question is “how do I get to 25%?”
You can’t do it alone.
2. Build operational autonomy
Psychological safety matters. But it’s not enough. Safety means you can speak up without fear. Autonomy means you can actually do something about it.
And here’s the harder truth: some leaders create the perception of psychological safety without the real thing. They say “my door is always open,” but punish the people who walk through it. They ask for feedback but get defensive when it comes. The language of safety without the behavior.
Even when safety is real, most orgs stop there. They think “we created a safe space” is the finish line. But if people can raise concerns but can’t act on them, you’ve given them a voice with no power.
Operational autonomy means: clear goals that don’t shift constantly. Outcomes-driven teams figure out how. Frank conversations tied to real goals. “Inform, don’t ask permission.” More on that in the next post.
But even structure is fragile. It needs density to be sticky, or it reverts.
3. Create artifacts that outlast you
When I say artifacts, I don’t mean posters with values on the wall.
I mean: What question gets asked in every meeting? What document is required before a decision?
Amazon’s 6-pager is a good example: a narrative memo, no bullet points. You can’t hide shallow thinking in prose. The format shapes the questions. Change the format, change what gets asked.
Decision logs. Pre-mortems. Rituals. These continue long after you’ve left the room.
The course wasn’t the variable
A colleague and I took the same course. Same content, same investment—$2,000 each. We both believed in it.
I applied the learnings immediately. It stuck.
My colleague couldn’t. Not because of their motivation. Because the system didn’t give them the chance. No space to experiment. No authority to change how decisions were made.
Same course. Same belief. Different environments. Different outcomes.
The course wasn’t the variable. The environment was.
You don’t rise to the level of your training. You fall to the level of your environment.