| Phase 1: Familiarisation with the data | All 19 interviews were transcribed and read multiple times, with detailed notes taken in NVivo following thematic immersion and reflexive engagement (Nowell et al., 2017). The focus was on how FMs described their daily work, decision-making boundaries and responses to organisational conditions, providing insights into role enactment in complex environments (Gioia et al., 2013). Emotional tone, language and implicit meanings were noted, underscoring the importance of interpretive sensitivity in thematic analysis. Reflexive notes and early analytical memos were recorded to document initial interpretations and emerging assumptions |
| Phase 2: Generating initial codes | Using NVivo, transcripts were inductively coded line-by-line, following thematic analysis recommendations (Braun and Clarke, 2022). Codes were assigned to in vivo quotes representing behaviours, beliefs, and organisational dynamics, forming first-order codes in participants' own words (Gioia et al., 2013). During the inductive coding, care was taken to reflect participants' lived experiences (Wimpenny and Gass, 2000) |
| Phase 3: Searching for themes | Coded data were reviewed to be grouped into clusters of meaning (Braun and Clarke, 2022). First-order codes were grouped into second-order themes, e.g. “boundary spanner”, “opportunity identification” and “translating knowledge”, reflecting conceptual patterns across interviews (Gioia et al., 2013). Merging constructs began to reflect or resonate with key CE concepts, including environmental triggers, organisational structures and individual agency (Urbano et al., 2022) |
| Phase 4: Reviewing themes | A constant comparison process refined emergent patterns through iterative analysis (Fereday and Muir-Cochrane, 2006). Overlapping categories were refined or collapsed into broader themes (Gioia et al., 2013). For example, “organisational constraints” and “barriers” were merged into the “organisational inhibitors” theme. Transcripts were reread to ensure data segments captured meaning and were appropriately coded (Braun and Clarke, 2006). Themes were iteratively compared across interviews to assess consistency, variation and analytical fit |
| Phase 5: Defining and naming themes | Themes were named to reflect how FMs engage in entrepreneurial activities in their organisations. Themes were iteratively refined to ensure coherence with participants' experiences (Fereday and Muir-Cochrane, 2006, Nowell et al., 2017). Second-order themes such as “management support”, “autonomy” and “role negotiation” were mapped onto CE constructs, linking antecedents (environmental, organisational, individual) with CE activities (e.g. opportunity recognition and role adaptation) and innovation outcomes (Ireland et al., 2009) |
| Phase 6: Final data structure with aggregate themes | Themes were synthesised into a coherent narrative about FMs' engagement in CE within high-tech manufacturing. The final data structure depicts first-order codes, second-order themes and aggregate dimensions (Gioia et al., 2013). ORT is used as an interpretive lens to gain insight into FMs' expected and enacted roles in CE (Anglin et al., 2022). An audit trail was maintained throughout the analysis to document coding decisions, theme development, and refinements to the emerging data structure. Rigour was further strengthened through peer debriefing with two senior qualitative scholars (Miles et al., 2020) and member checking to confirm transcript accuracy (Saunders et al., 2019), and data triangulation (Miles et al., 2020) |