Many connection programs fail by targeting the wrong level of the problem, not because of poor execution.

Disconnection at work is structural, relational, or individual—and each requires a different response.

Diagnosis before design is the step that determines whether any intervention will work.

This post is co-authored by Nathaniel Sabater (Annecy Behavioral Science Lab), Mariangel Maldonado (former head of well-being at Booking.com), and Hans Rocha IJzerman.

Every year, organizations roll out empathy training, wellness programs, mentoring schemes, and team rituals. Engagement scores shift in some teams. In others, nothing moves. Leadership concludes that the initiative needed stronger buy-in, better facilitation, or more budget. They try again next quarter with a slightly different version of the same thing.

Here is a harder explanation: The problem was often never the program. It was the problem selection.

Organizations frequently invest in solutions before identifying the actual source of their disconnection. The result is not just wasted resources. It is something more damaging: employees who engage with initiatives that leave their core frustrations untouched and who eventually stop believing that leadership understands what they are actually experiencing. Over time, responsibility tends to shift toward individuals—people are encouraged to adapt to conditions that are better addressed at the system or team level. This is not a failure of intent. It is a challenge of problem selection.

This is why diagnosis must come before design.

Why Symptoms Are Misleading

Burnout, disengagement, quiet quitting, and high turnover are symptoms. They tell you something is wrong, but they do not tell you where the problem lives. And location determines everything about how to address it.

Structural disconnection looks different from team-level disconnection, which looks different again from individual disconnection. When a hybrid policy creates unequal access to leadership, that is a structural problem. When a manager's communication style makes people reluctant to ask for help, that is a team problem. When an employee struggles with belonging after a major life transition, that is an individual problem. All three might register identically as disengagement on your annual survey. They require completely different responses.

Applying the wrong intervention is not neutral. It can actively reinforce the conditions causing the problem. Encouraging employees to build personal resilience in response to structural overload, for instance, shifts system-level problems onto individuals who did not create them. As Gartner's research across 426 chief human resources officers (CHROs) in 23 industries highlights, organizations are far more likely to achieve lasting change when it is embedded into the daily rhythms of work—not added on top of an unchanged environment through campaigns. Connection efforts lose traction precisely when they are designed before leaders are clear on what is disrupting connection in the first place.

A Three-Level Diagnostic Framework

Drawing on Burke and Litwin's causal model of organizational performance and change, our multi-country research on social connection, and qualitative interviews with people leaders across industries, we summarized a structured approach to help organizations identify where disconnection is actually generated—before designing any intervention.

The framework distinguishes three levels, and the order in which you assess them matters.

The first is the structural and organizational level. The key question here is: Are the conditions we have designed quietly disconnecting people? This level covers workload distribution, staffing ratios, meeting loads, hybrid and remote policy design, role clarity, decision authority, and how restructuring transitions are managed. Structural factors shape whether meaningful interaction can even occur—the frequency of contact, modes of communication, and whether employees have access to the relationships and support they need to do their work. If this level is broken, no amount of team programming or individual support will compensate for it.

The second is the interpersonal and team level. If the structure seems sound, the question becomes: Is the team environment making connection feel unsafe? This covers psychological safety, manager behavior, team norms, collaboration quality, communication reciprocity, and how conflict and support—emotional, informational, and practical—are handled. Team factors shape not whether interaction occurs but how safe and supportive those interactions feel. A team can have every structural condition in place and still be a place where people feel unable to speak honestly or ask for help.

The third is the individual level. When both structure and team dynamics are genuinely supportive, the question becomes: Do individual factors explain persistent disconnection? This level covers personality, beliefs, mental health, personal life circumstances such as caregiving responsibilities or health challenges, social skills, coping strategies, and the impact of life transitions. Individual factors shape subjective experience—whether a person feels they belong, are appreciated, and can be authentic. This is the appropriate level to address last, not first.

The sequence matters because each level sets the conditions for the next. Structural constraints define the boundaries within which teams operate. Team dynamics shape what individuals experience day to day. Beginning with individual-level solutions—which is where most programs default—when the problem is structural or relational is not just ineffective. It misattributes the source of disconnection and makes it harder to address the real issue later.

What Diagnosis Makes Possible

When organizations identify the right level before designing their response, our analysis shows that return on investment can be two to three times higher than when generic approaches are used. Tailored interventions that address root causes consistently outperform mismatched solutions that treat surface patterns while leaving underlying conditions unchanged.

The practical implication for people leaders is clear, if uncomfortable: Slow down before the next initiative launch. If your data points to structural overload, a mindfulness program will not move the dial. If it points to unsafe team dynamics, a belonging campaign will not reach the core issue. If individual circumstances are the driver, system-level redesign is the wrong tool.

Diagnosis is not a delay. It is the precondition for doing anything that actually works.

A culture of connection cannot be built through programs alone. It must be engineered into systems, relationships, and the daily architecture of work. That starts with knowing which problem you are actually trying to solve.

A version of this post also appears on the Annecy Behavioral Science Lab website.

© Psychology Today