Budgets for upskilling are rising while enthusiasm for work is not. Global employee engagement slid to 20 percent in 2025, the lowest since 2020, with Gallup estimating roughly 9 percent of world GDP in lost productivity, about 10 trillion dollars (Gallup, 2025). Manager engagement fell sharply too, from 27 to 22 percent in a year and down nine points since 2022. Many firms responded with longer course catalogs and new compliance clocks. That treats learning as attendance. The better question is what would make people choose to learn, and keep choosing when the novelty fades.
What self-direction actually requires
Self-directed education is not a mood or a menu. It is an operating design that meets three basic psychological needs: autonomy, competence, and relatedness. The point is not theoretical purity. The evidence base has matured enough to test design at scale. A recent meta-analysis synthesized 36 educational intervention studies with 11,976 participants and, for the first time in over a decade, reviewed interventions that target all three needs in combination, not just autonomy support (Wang et al., 2024). The authors explicitly examine whether self-determination theory interventions move core outcomes such as motivation. Even without overstating results, the direction of travel is clear. The field is not arguing about slogans. It is testing mechanisms.
For business leaders, the translation is practical. Autonomy is not the absence of guardrails. It is meaningful choice inside clear constraints. Competence is not praise. It is fast, specific feedback and work that gets harder as you master it. Relatedness is not a social wall. It is a community that notices your effort and invites contribution. If your learning offer ignores any one of these, self-direction becomes a euphemism for neglect.
Safety is the gate to learning behavior
Even the best-designed curriculum will stall if people worry that asking a basic question will mark them as slow. The construct for that worry has a name. Psychological safety is a shared belief that a team is safe for interpersonal risk taking, and it predicts whether people engage in the behaviors that drive learning on the job, such as admitting a mistake or requesting help (Edmondson, 1999). In a field study of 51 teams within a single manufacturing company, team psychological safety was associated with learning behaviors. Those behaviors, in turn, mediated the relationship between safety and performance. Team efficacy alone did not explain learning once safety was accounted for. The study is cross-sectional and limited to one firm, so it is not a causal proof across industries. It is, however, the classic case that put a crucial variable on the table.
If your upskilling program does not change the risk calculus for speaking up, it will change very little. The implication is simple. Treat learning conversations like code reviews or safety huddles. Make error reporting routine. Reward the first person to surface a gap, not the last person to memorize a slide.
Structure, then freedom
Self-direction fails in organizations that confuse fewer managers with less management. The research on self-managing organizations draws the line clearly. These are systems that decentralize authority in a formal, systematic way across the whole company. Their defining feature is the elimination of the manager–subordinate relationship, replaced by explicit processes for who decides what and how conflicts are resolved (Lee and Edmondson, 2017). The literature they synthesize argues that classical hierarchy works well for reliable execution in stable settings, but tends toward rigidity and status dynamics that stifle cross-boundary problem solving and people development. The paper is a theoretical synthesis rather than an experiment, yet it is the canonical definition and a useful map of the limits of traditional design.
The lesson for learning leaders is straightforward. If you want adults to own their education, give them real authority over the path and the pace, then codify the interfaces. Publish decision rights for projects, not just for budgets. Make peer review the default, not a favor. Document how someone earns the right to attempt higher-stakes work, and what evidence counts as readiness. Freedom sits on top of process, not in place of it.
The manager’s new job
Gallup’s numbers hint at a more basic problem than content mismatch. With only 22 percent of managers engaged in 2025, and a 79 percent engagement rate in best-practice organizations, the gap looks like design, not fate (Gallup, 2025). That matters, because managers are the de facto learning environment. If they mostly assign courses, the signal is compliance. If they design conditions that satisfy autonomy, competence, and relatedness, the signal is ownership. The Wang et al. meta-analysis is not a license to romanticize choice. It is a reminder to operationalize it: expand the menu of worthy problems, legitimize different paths to mastery, and tie recognition to demonstrated capability rather than seat time (Wang et al., 2024).
In practice this looks less like a training calendar and more like an apprenticeship market. Managers publish backlogs of real business challenges with the skills they demand. Employees opt in, form learning circles, and commit to deliverables. Leaders coach, clear blockers, and hold the quality bar. The psychological safety play is continuous. Leaders narrate what they are learning in public, normalize the phrase “I do not know yet,” and treat early errors as data. The self-management play is to make decision rights transparent, so contributors understand when and how they can move without waiting for permission (Lee and Edmondson, 2017; Edmondson, 1999).
The practical bar for success is not attendance. It is throughput of solved problems and visible growth in capability. In a world where engagement has fallen and the cost of that drift is measured at the scale of national economies, treating learning as a design problem is both humane and hard-nosed (Gallup, 2025). Build for autonomy, competence, and relatedness. Make safety explicit. Give people real authority and codify the edges. Then measure what they ship. That is how self-directed education becomes an operating advantage rather than a slogan.
Sources
- Wang et al. (2024). SDT-based educational interventions: a meta-analysis. Synthesizes 36 interventions with 11,976 participants and, for the first time in over a decade, reviews designs targeting autonomy, competence, and relatedness while explicitly testing effects on motivation. https://www.sciencedirect.com/science/article/abs/pii/S0023969024000572
- Lee and Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Defines self-managing organizations as formal, systematic decentralization that removes manager–subordinate relationships, clarifying why structure is needed for autonomy to work. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams. Introduces team psychological safety and shows, in a field study of 51 teams in one firm, that safety is associated with learning behaviors that mediate performance. https://journals.sagepub.com/doi/10.2307/2666999
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20 percent, an estimated cost near 10 trillion dollars or 9 percent of GDP, and a sharp drop in manager engagement, underscoring the design problem facing learning at work. https://www.gallup.com/workplace/659610/state-global-workplace.aspx