Standard advice for a major career change follows a planning sequence that most professionals recognize. Identify what you are good at, choose a target role or field that fits, map the gap between where you are and where you want to be, and work your way across it. The sequence is intuitive, consistent with how most organizations approach development, and widely used in coaching and career counseling. It is also, according to the empirical research on how professionals actually move between careers, a reasonable description of only the easy transitions.
Ibarra's (1999, 2003) multi-year interview studies of consultants, investment bankers, and other professionals in major work role transitions found that the planning-first sequence failed most of the people who tried it. The reason was not that they chose the wrong target or failed to plan rigorously enough. It was that the method assumes identity is stable while the plan is executed, when in fact the identity itself is what the transition is changing. A professional cannot plan toward a target self they do not yet have access to, because the preferences, instincts, and relationships that would allow that self to feel coherent have not formed yet.
What Ibarra and later researchers documented instead is a process that looks the opposite of planning. Professionals act in small ways that are inconsistent with their current identity, observe the response from themselves and others, and gradually accumulate the material that lets a new identity become plausible. The planning that advice literature treats as the beginning of a career change tends in practice to come near the end: a narrative constructed retrospectively, once something new has taken shape, that organizes the preceding experiments into what sounds like a plan.
This difference is not semantic. If identity reconstruction precedes planning rather than following it, the useful activities during a transition are different from the ones that career advice typically prescribes. The sections that follow review what the research has shown about the activities that make such identity change possible, the stages at which they concentrate, and the specific transitions where this has been studied most closely. The planning-first sequence does work well for the smaller class of transitions in which the identity does not have to change, such as a lateral move between similar roles.
Ibarra's (1999) study of consultants and investment bankers in career transitions, published in Administrative Science Quarterly, introduced the concept of provisional selves. Her interviews revealed that professionals undergoing significant role changes did not try to adopt a fully formed new identity. They tried out partial versions: speaking the way a more senior consultant might speak, taking on a project that resembled what a target role would involve, modeling their behavior on a specific person who already held the identity they were considering. Each experiment produced information from two sources. The internal source was the person's own sense of whether the behavior felt coherent, which Ibarra treated not as a vague feeling but as practical data about what the person could sustain. The external source was the response of others, particularly those whose judgment the person already weighted, whose reactions supplied evidence about whether the trial version was legible to the audience it would need to persuade. Some of the trial behaviors felt accurate and were kept. Others felt forced and were discarded. The eventual identity was assembled from what survived.
Working Identity (Ibarra, 2003) extended the research base through additional interviews with thirty-nine professionals in significant career change and consolidated the findings into three practices. The first was crafting experiments: small-scale, time-limited attempts at the new work, often outside the current job, that generated real information rather than hypothetical analysis. The second was shifting connections: deliberately changing the people whose feedback mattered, usually by reducing contact with those who held the current identity in place and increasing contact with people embodying the possible one. The third was making sense through narrative: constructing and revising a story about who the person was becoming, often in conversation, until a version held together under questioning.
Ibarra and Barbulescu (2010), in a conceptual paper in the Academy of Management Review, argued that the narrative practice is not optional ornamentation on the other two. It is the mechanism by which the experiments and connections cohere into something a person can act from. Without a plausible story, the material produced by experiments stays as unrelated episodes. With it, the material organizes into an identity the person can carry into the next situation. The three practices are therefore interdependent. Experiments without new connections produce findings nobody responds to. Connections without experiments produce admiration for identities a person has no way to try. And both without narrative produce experience that never becomes identity.
Petriglieri, Petriglieri, and Wood (2018), in a longitudinal qualitative study of fifty-five managers pursuing a one-year international MBA at INSEAD and published in Administrative Science Quarterly, extended the findings into a setting Ibarra had not studied. Participants were in the middle of mobile careers across organizations, industries, and countries. The authors documented identity work that produced what they termed portable selves: identities loosely bound to any single role, assembled to survive repeated transitions. Read alongside Ibarra's work, the study suggests that identity reconstruction occurs inside structured programs as well as in open career markets, though Petriglieri and colleagues frame their contribution as the emergence of portable selves, not as a test of Ibarra's three practices.
Nicholson (1984), in a theory paper in Administrative Science Quarterly, offered a framework that locates where identity work happens within the broader arc of a role change. Transitions, he argued, proceed through four stages. Preparation covers the period before the change, when expectations and anticipatory adjustments form. Encounter is the first experience of the new role, where expectations meet reality. Adjustment is the longer phase in which the person modifies the role, is modified by it, or both. Stabilization is the period in which the resulting arrangement becomes the new normal.
Nicholson's contribution was partly to name what psychologists and sociologists had been describing in scattered form, and partly to argue that each stage has characteristic risks. Preparation can produce expectations so detailed that encounter is experienced as rupture. Encounter can be short enough that the person retreats to their prior identity before adjustment has begun. Adjustment is where identity reconstruction actually happens, and it is also where organizations most often fail to provide the space for it, treating a new role occupant as if they are already stabilized within weeks of starting. Stabilization carries its own risks, including premature closure, in which the person settles into a workable version of the role that forecloses the better one they might have reached with more time.
Nicholson distinguished among modes of adjustment: replication, in which the person conforms to the role as given; absorption, in which the role transforms the person; determination, in which the person transforms the role; and exploration, in which both change. Which mode dominates depends on the individual's prior socialization, the degree to which the role is loosely or tightly defined, and the discretion the organization allows. The practical implication is that adjustment is not a single process. Transition quality depends on whether the mode that emerges fits the circumstances, and on whether the organization and the person can tell which mode they are in.
Subsequent work (Ashforth, 2001; Ashforth, Harrison, and Corley, 2008) extended the framework by examining role boundaries, the micro-transitions people make into and out of roles each day, and the broader dynamics of identification in organizations across more than four decades of research. Nicholson's four-stage structure remains the standard reference for the arc of a major transition and the point at which identity work concentrates within it.
Hill's (2003) longitudinal interview study of first-time managers, originally published in 1992 and reissued with new material, follows nineteen high-performing individual contributors across their first year in a management role at four US corporations. Hill interviewed each manager roughly monthly over the first year, supplemented by observation and documents. The research question is not whether these transitions succeeded, since Hill selected people who were considered strong candidates, but what the transition actually involved. The answer was that it involved reconstructing much of what the person had understood about work.
The managers Hill interviewed had generally succeeded as individual contributors by developing deep technical expertise and completing their own work to a high standard. The new role rewarded none of this directly. It rewarded getting work done through others, which required a different set of behaviors, a different source of satisfaction, and in many cases a different understanding of what counted as a contribution. The managers who struggled were not those who lacked management skills on arrival. They were those who kept trying to add value through the technical work they already knew how to do, because the identity that produced their previous success did not transfer to the new role.
Hill traced four common misconceptions new managers brought with them. They expected the role to give them more authority, when in practice it gave them more constraints. They expected to be the technical expert their team relied on, when in practice the team's technical questions moved beyond them within months. They expected relationships with their former peers to continue, when in practice the relationships were restructured by the hierarchy, whether the new manager wanted this or not. They expected to focus on individual team members, when in practice the actual leverage was in group dynamics, which many first-time managers had never consciously attended to. Each misconception, Hill noted, had to be revised through experience. None could be revised by being told about it in advance, because the revision was an identity change, not an information update.
The practical consequence for how first-time manager programs are structured follows from this. A classroom module on, for example, situational leadership can transmit the concept, but it cannot by itself change the sense of self that keeps the new manager reaching for the technical work. What appears to help, in Hill's data and in subsequent research on managerial transitions, is a combination of structured exposure to the demands of the new role, access to more experienced managers whose identity has already made the shift, and time for the new manager to construct a working account of what they are now responsible for.
If working identity research focused on people leaving one career for another, and Hill's work focused on a role transition within a career, Pratt, Rockmann, and Kaufmann's (2006) ethnography of medical residents, published in the Academy of Management Journal, extended the picture into a different case: people inside a long, highly structured socialization process whose eventual identity is in principle well specified. Over six years of observation and interviews with fifty-five residents across three specialties, the authors documented identity work that resembled Ibarra's findings in structure while differing in content.
The trigger for identity work in Pratt and colleagues' setting was what they termed work-identity integrity violations: discrepancies between what the residents were doing and the sense of self they had entered the specialty with. A pediatrics resident, for example, might find that most of her days consisted of insurance paperwork and parent reassurance rather than the clinical reasoning she had trained for. The resulting gap produced three identity-customization patterns the authors labelled enriching, patching, and splinting.
Enriching involved adding to the current identity to accommodate the actual work. A resident might reframe parent reassurance as a central skill of pediatric practice rather than a distraction from it. Patching involved substituting an adjacent identity for parts of the role that fit poorly, such as treating the paperwork portion as administrative work rather than medical work, drawing on an identity the resident had held in a prior career. Splinting involved temporarily holding two incompatible identities together, usually under acute strain, until one could be relinquished. The three were not fixed types. Residents moved among them as the work and their understanding of it changed over the course of training.
What Pratt and colleagues' study established, and what the earlier research had only implied, was that identity work is a continuous feature of professional life, not an episodic response to career change. The authors described this as a learning cycle in which the resident's current identity shaped how new work was interpreted, the interpretation produced new experience, and the experience in turn modified the identity. The cycle continued for the full six years of observation, with no evidence that it closed once the formal socialization was complete. Even in a role whose destination is prescribed, the person inhabiting it is doing constant customization work to keep sense of self and daily activity in a workable relationship.
Caza, Moss, and Vough (2018), analyzing qualitative data gathered over five years from forty-eight plural careerists (people holding more than one work-role identity simultaneously, such as actor-waiters and researcher-consultants) and published in Administrative Science Quarterly, extended the picture further. They described a progression from synchronizing individual work-role identities to harmonizing them into a broader work self, and argued that authenticity does not require being one consistent identity across contexts. The practices resemble Ibarra's and Pratt's in structure, but their object is the integration of multiple identities rather than the reconstruction of one.
Several caveats constrain how far this synthesis can be pressed. The first concerns sample composition. Ibarra's (1999, 2003) studies drew heavily on management consultants, investment bankers, and other professionals in high-discretion knowledge work. Hill's (2003) managers worked in US corporations in the 1990s. The identity reconstruction described in both may look different among workers with less role autonomy, in different national contexts, or in gig and platform work where the organizational structures that frame a transition are weaker. Petriglieri, Ashford, and Wrzesniewski (2019), in an interview study of sixty-five independent workers published in Administrative Science Quarterly, found that identity work in gig arrangements requires self-constructed holding environments to replace what employers used to provide. The applicability of the earlier frameworks is therefore uneven across employment types.
The second concerns the role of narrative in the methodology. Much of the evidence is retrospective interview data, in which professionals describe transitions they have already made. Retrospective narrative is itself one of the practices under study (Ibarra and Barbulescu, 2010), which raises the question of whether what interviewers hear reflects the actual process or the coherent story the participant has since constructed. The design is appropriate for mapping how people make sense of transitions. It is less so for establishing exactly which activities caused which outcomes.
The third concerns demonstrated intervention. The literature reviewed here describes patterns and names practices. It does not provide experimental evidence that teaching the practices to people in transition produces better outcomes than the planning-first approach it critiques. Caza, Vough, and Puranik's (2018) review of identity work research in the Journal of Organizational Behavior noted that intervention studies lag the descriptive literature by a wide margin. The frameworks here are best treated as a grounded account of what happens during career identity change, not as a demonstrated model of how to make such change go well. A separate concern is that different authors in this literature define professional identity differently, so findings do not combine cleanly across different operationalizations of the construct.
Professional identity does not stay fixed across a career, and it does not change through planning alone. It changes through small experiments with new versions of the self, through deliberate shifts in the relationships that confirm or question who one is, and through the stories that organize this material into something coherent. These practices concentrate in the adjustment phase of a transition, which the research consistently identifies as the stage organizations most often fail to protect.
The useful implications are different from what career advice typically offers. For individuals, the question during a transition is less what the target role is and more what experiments would generate real information about possible selves, which connections to tend, and what narrative is beginning to feel accurate. For organizations, the implication is that the people most at risk during a transition are not those who need more training but those whose adjustment phase is being compressed into weeks. Protecting the time and conditions in which identity work can happen is a less visible investment than skill development, and based on the evidence reviewed here, it is at least as consequential.