‘Burned out’ millennials! Can job crafting promote employee well-being in the SMEs?

The basic tenet of the COR theory posits that individuals invest efforts to gain, foster, preserve, and conserve resources they centrally value (Hobfoll 1989; Hobfoll et al. 2018), highlighting humans’ natural tendency to safeguard existing resources and acquire for additional resources (Halbesleben et al. 2014). As a critical aspect in the COR theory, resources are elements that individuals perceive as beneficial in achieving their goals (Halbesleben et al. 2014). Applying COR to the organisational context, employees will strive to foster and protect their well-being by increasing and safeguarding the resource pool central to work goal attainment (Harju et al. 2021; Singh and Singh 2018).

According to the COR theory, psychological strain in the form of burnout and stress emerges when individuals face an actual or anticipated resource loss (Halbesleben et al. 2014; Wallace and Coughlan 2023). Hence, individuals will rationally engage in behaviour seeking to maximise resource gain and minimise resource loss (e.g., job crafting) to mitigate strain because resource loss leaves more severe impacts on individuals than resource gain (Hobfoll et al. 2018). As job crafting allows employees to build and acquire resources to fulfil their personal needs and preferences, employees will gain greater job satisfaction through job crafting (Kim et al. 2018).

The gain paradox principle embedded in the COR theory states that when individuals are in situations of high resource loss, resource gain brings greater value to them (Hobfoll et al. 2018). Burnout generally happens when employees fail to replenish their resource pool or do not obtain the expected returns from their resource investment (Halbesleben et al. 2014). Applying the gain paradox principle, allowing employees to craft their jobs will help them derive greater job satisfaction, indicating a resource gain to them. Such a gain effect will override the loss effect of burnout, leading to lower burnout experience.

This study differentiates job crafting into four independent dimensions (i.e., increasing structural job resources, increasing social job resources, increasing challenging job demands, and decreasing hindering job demands) because these dimensions comprehensively capture the aspects of job crafting (Ok and Lim 2022). Increasing structural resources involves behaviours aimed at gaining autonomy, task variety, and development opportunities, whereas increasing social resources involves behaviours to find social support, such as coaching or feedback (Luu 2020; Tims et al. 2015). Employees who voluntarily expand their job scopes or take on advanced tasks are seen as increasing challenging demands, while those who reduce their workload and minimise emotionally, cognitively, or physically taxing aspects are viewed as decreasing hindering demands (Wang et al. 2017).

Prior studies have shown that job crafting is negatively related to burnout (Cheng and O-Yang 2018; Singh and Singh 2018). Paradoxically, job crafting may also impair employees’ well-being as it increases job complexity and workload (Harju et al. 2021; Li et al. 2022). Drawing on the COR theory, this study expects job crafting to mitigate burnout among millennials working in SMEs for two reasons. The first reason pertains to millennials’ self-centredness (Gong et al. 2018) and strong desire for decision-making power (Magni and Manzoni 2020). With these distinct characteristics, it is believed that millennials will prioritise building a resource base to achieve their work goals and find opportunities to control their job characteristics. Essentially, millennials will perceive job crafting as a resource-seeking strategy instead of a resource-depleting behaviour (Gong et al. 2018). Secondly, to survive in SMEs’ scarce resources environment (Woschke et al. 2017), job crafting is one optimal strategy for millennials to achieve work goals and reduce strain. The different job crafting dimensions function similarly to resource accumulation and conservation behaviour (Harju et al. 2021; Ruparel et al. 2022), helping millennials working in SMEs to maintain their well-being. Hence, we propose the following:

The literature views job satisfaction in different ways but this construct generally reflects employees’ affective responses or positive feelings toward their jobs (Cheng and O-Yang 2018; Tsigilis et al. 2004). Job satisfaction critically influences employees’ productivity and organisational success (Ekmekcioglu and Nabawanuka 2023). Utilising samples of millennials, Moulik and Giri (2023) have confirmed that job crafting in the form of increasing structural and social resources leads to higher job satisfaction. Such findings align with the notion of the COR theory, which emphasises that gaining social and task resources keeps individuals away from resource depletion and enables them to deal with demands (Nielsen and Abildgaard 2012). Although the dimension increasing challenging job demands requires employees to invest extra effort, it promotes the accumulation of resources to one’s capabilities (Ruparel et al. 2022). Arguably, this job crafting dimension is particularly important for the millennial cohort, given their tendency to venture into challenges or seek new knowledge and crave continuous learning (Ng and Feldman 2010; Yap and Badri 2020). Given that employees feel stressed and need to channel extra resources to resolve hindering demands (Harju et al. 2021), having the opportunity to decrease hindering demands through job crafting will logically increase job satisfaction (Nielsen and Abildgaard 2012). Previous studies have also found that job crafting is positively related to job satisfaction (Cheng and O-Yang 2018; Ekmekcioglu and Nabawanuka 2023). Therefore, we suggest the following:

Scholars have defined burnout in different ways (Maslach et al. 2001; Tsigilis et al. 2004). Debates on how burnout should be conceptualised and measured are still ongoing (Demerouti et al. 2021). In this study, burnout refers to a psychological job-related syndrome caused by exposure to chronic job stressors, resulting in a physically, emotionally, and mentally drained state (Maslach et al. 2001). Burnout is notoriously associated with negative attitudinal and behavioural outcomes, such as higher turnover intention and counterproductive work behaviours (Lu and Gursoy 2016; Wallace and Coughlan 2023). While previous studies have established a significant negative relationship between burnout and job satisfaction (Ayachit and Chitta 2022; Kim et al. 2017), we posit that satisfied employees are less likely to experience burnout with the guidance of the COR theory. The resource-draining conditions resulting from burnout are especially salient in the absence of resource gain (Hobfoll 2001). Although job satisfaction is an intangible feeling, it generates positive and pleasurable emotional resources to weaken the stress and strain resulting from burnout (Li et al. 2023). Utilising the COR theory, Wallace and Coughlan (2023) also highlighted that constructs representing positive emotions are resources that can buffer against burnout. Thus, we posit that:

Following the discussions above, it stands to reason that job satisfaction mediates the relationship between the different job crafting dimensions and burnout. By actively engaging in different job crafting dimensions, employees expand their resource reservoir and possess increased control over their work and future outcomes, resulting in higher job satisfaction (Cheng and O-Yang 2018; Wang et al. 2017). Subsequently, the burnout level among employees with higher job satisfaction is lower because they perceive their work as meaningful and fulfilling (Ekmekcioglu and Nabawanuka 2023). The proposed mediating relationship in this study will be particularly effective because millennials value challenging and meaningful work (Lu and Gursoy 2016; Magni and Manzoni 2020). If restricted job crafting opportunities are given to them, they will feel less satisfied with their jobs and more stressed about surviving in the inherently limited resources environment of SMEs. As such, we propose the following:

Figure 1 shows the proposed conceptual model in this study.

This study recruited participants through Prolific, an online platform offering high-quality data from real respondents (Palan and Schitter 2018). To ensure we reach the relevant samples, the following selection criteria were set on Prolific: (1) Participants currently working full-time in SMEs; (2) They must fall within the millennial age bracket (i.e., born between 1981 and 1996). Qualified respondents were directed to complete a 7-minute survey and received a compensation of $1.05 upon completion.

While recognising the potential of facing common method bias (Kock 2015), this study collected cross-sectional, single-wave, and self-report data based on two considerations. First, self-report data was appropriate because the constructs under investigation evaluated the subjective perceptions and feelings of the respondents (Wang et al. 2023). Second, the survey was not lengthy, as all questions were worded based on established scales with clear and concise statements. Therefore, we decided to collect the data at a single time point to prevent a low response rate. Respondents’ confidentiality and anonymity were ensured during data collection to reduce potential common method bias-related issues. Since a forced response was set for each question, there was no missing data. We collected 352 valid responses from individuals residing in the UK. The respondents consist of 42.6% (n = 150) male and 57.4% female (n = 202). A majority of them were single (54.3%, n = 191) and reported holding intermediate (38.1%, n = 134) and mid-level (39.5%, n = 139) positions in their respective companies. This information is included in Table 1.

Job crafting. Job crafting was assessed using Tims et al.’s (2012) 21-item scale with increasing structural job resources, increasing social job resources, and increasing challenging job demands consisting of five items respectively, while decreasing hindering job demands consists of six items. All items were measured using a five-point scale, ranging from 1 = Never to 5 = Very often. Example items included: “I try to develop my capabilities” and “I ask my supervisor to coach me”. One item from the dimension increasing structural job resources was deleted due to low factor loadings. The dimensions increasing structural job resources and increasing social job resources yielded the same Cronbach’s alpha values of 0.88, while the dimensions increasing challenging job demands and decreasing hindering job demands yielded Cronbach’s alpha of 0.86. As this study seeks to reveal how specific forms of job crafting affect millennials’ burnout experience, job crafting was treated as distinct single-order measures instead of a higher-order construct.

Job satisfaction. Lee and Bruvold’s (2003) three-item scale was adopted to measure job satisfaction. Answers were rated using a five-point scale, ranging from 1 = Strongly disagree to 5 = Strongly agree. A sample item included was “Generally speaking, I am satisfied with my job”. Job satisfaction had a Cronbach’s alpha of 0.84.

Burnout. Burnout was assessed using four items from the Copenhagen Psychosocial Questionnaire (Llorens et al. 2019), which required respondents to rate their burnout frequency on a five-point scale, ranging from 1 = All the time to 5 = Not at all. One sample item was “How often have you felt worn out?”. Burnout had a Cronbach’s alpha of 0.91.

Control variables. Gender (coded 1 = male, 2 = female) and marital status (coded 1 = single, 2 = married) were included as control variables due to their potentially confounding effects towards job satisfaction and burnout (Cañadas-De la Fuente et al. 2018; Muskat and Reitsamer 2020).

SPSS v.28 and AMOS v.28 were used to analyse data in this study. SPSS was used to analyse basic statistics, while AMOS was used to conduct confirmatory factor analysis (CFA) to examine the discriminant validity of the study measures (Hair et al. 2019), followed by assessment of structural models and hypotheses testing.

As presented in Table 2, the skewness and kurtosis values of the data fell within the range of ±2 and ±7, suggesting normally distributed data (Hair et al. 2019). There was no multicollinearity and common method bias (CMB) with variance inflation factor scores (VIF) all below 3.3 (Kock 2015). Further ensuring the non-existence of CMB, Harman’s single-factor test was conducted (Podsakoff et al. 2003). This test suggests that a data set has the potential risk of CMB if a single latent factor explains most of the variance (Kim et al. 2018). Our analysis showed that the single factor explained only approximately 33% of the variance. Thus, this alternative CMB test signalled little concern about CMB. Correlation results indicated all three job crafting dimensions of increasing structural job resources (0.410**, p <.01), increasing social job resources (0.290**, p <.01) and increasing challenging job demands were positively linked to job satisfaction (0.322**, p <.01). Meanwhile, decreasing hindering job demands was negatively linked to job satisfaction, but the association was not significant (−0.006, p >.05).

CFA was performed to determine the fit of the factor structure with the study sample based on the recommendation by Byrne (2016), including re-specification to covariate the error terms to improve the model fitness based on modification index value > 4 (Hair et al. 2019). Fit statistics, such as the Chi-square goodness-of-fit test, root mean square residual (RMR), Tucker-Lewis index (TLI), comparative fit index (CFI) and root mean square error of approximation (RMSEA) were used to evaluate the model fit. The model derived a good fit if the values for RMR and RMSEA were below 0.08, and TLI and CFI were beyond 0.90 (Hair et al. 2019). We also assessed the standardised root mean square residual (SRMR), following the recommendation of Mai et al. (2021).

The visual representation of our CFA model is shown in Fig. 2 in the Appendix. The overall model fit of the hypothesised six-factor model was acceptable (X²/df (306) = 1.87, RMR = 0.067, TLI = 0.946, CFI = 0.953, RMSEA = 0.050, SRMR = 0.031). The factor loadings were all above 0.50. As shown in Table 3, all scales had Cronbach’s alpha value exceeded 0.70, suggesting good internal consistency. Similarly, the measurements obtained composite reliability (CR) values above 0.70, demonstrating adequate reliability. We further assessed convergent validity through average variance extracted (AVE) values, and all AVE values exceeded the threshold of 0.50 (Hair et al. 2019). To examine discriminant validity, we used heterotrait-monotrait ratio of correlations (HTMT) with a cut-off value of 0.85 following Henseler et al. (2015). Given that all HTMT values did not exceed 0.85, the measurement model did not have the issue of discriminant validity. As depicted in Fig. 3 in the Appendix, fitness indices of the structural model were evaluated prior to hypotheses testing.

The proposed structural model showed a good fit with (X²/df (348) = 1.81, RMR = 0.063, TLI = 0.942, CFI = 0.950, RMSEA = 0.048, SRMR = 0.033). It is worth noting that the CFI and SRMR values were well within the flexible cut of criteria set by Mai et al. (2021) for studies with sample sizes of more than 200, indicating that the empirical data collected fit the structural model. We employed bias-corrected bootstrapping with 95% confidence intervals, including lower bound (LB) and upper bound (UB), to test the hypotheses. As indicated in Table 4, the studied model was free from the confounding effects of the control variables, gender and marital status towards job satisfaction and burnout. We found increasing structural job resources (β = 0.048, p =.647) and increasing social job resources (β = − 0.076, p =.333) with a mini-scale effect on burnout, thus not supporting H1a and H1b. Likewise, H1c and H1d were not supported with both increasing challenging job demands (β = 0.108, p =.313) and decreasing hindering job demands (β = 0.069, p =.347) having non-significant association with burnout. These findings imply that the four job crafting dimensions did not reduce millennials’ burnout levels.

Furthermore, a positive and significant relationship was found between increasing structural job resources (β = 0.330, p =.001) and increasing social job resources (β = 0.187, p =.022) with job satisfaction, lending support to H2a and H2b. Put differently, millennials feel more satisfied with their jobs with more developmental opportunities, autonomy, as well as co-worker and leader support. The effects of increasing challenging job demands (β = 0.041, p =.653) and decreasing hindering job demands (β = − 0.079, p =.216) on job satisfaction were not statistically significant, failing to support H2c and H2d. These results reveal that the flexibility to adjust challenging and hindering demands did not derive a sense of satisfaction among millennials. H3 was supported with a significant effect found between job satisfaction and burnout (β = − 0.440, p =.001). This means that millennials with higher levels of job satisfaction reported lower burnout.

Regarding the proposed mediation effects, it was found that the direct effect of increasing structural job resources on burnout was not significant, but the indirect path was significant (β = − 0.161, LLCI = −0.274, ULCI = −0.079), suggesting that job satisfaction fully mediates the relationship between increasing structural job resources and burnout. Hence, H4a is supported. Similarly, the direct effect of increasing social job resources on burnout was not significant, but the indirect path was significant (β = − 0.073, LLCI = −0.149, ULCI = −0.015), indicating the full mediation role of job satisfaction and H4b is also supported. Contrastingly, the direct and indirect paths for increasing challenging job demands and decreasing hindering job demands to burnout were not significant. These findings show that job satisfaction did not mediate the relationships and thus failed to support H4c and H4d. To confirm that the mediation effects in H4a and H4b are statistically significant, Sobel’s test (1982) was carried out on both pathways. The results also indicated significant indirect effects of increasing structural job resources on burnout via job satisfaction (z = −2.16, p =.002) and increasing social job resources on burnout via job satisfaction (z = −2.34, p =.019). The mediation analysis result is presented in Table 5.

This paper explores the connections and mediating mechanisms that elucidate the relationship between job crafting and burnout experience among millennials in SMEs. Our findings build upon the empirical research conducted by Moulik and Giri (2023) on job crafting’s impact on millennials’ employee well-being and happiness. However, we shift to the perspectives of burnout, examining whether millennials’ job satisfaction mediates this relationship.

We hypothesised that job crafting dimensions increasing structural job resources, increasing social job resources, increasing challenging job demands, and decreasing hindering job demands would reduce millennial burnout. Interestingly, we found no evidence of a direct link between job crafting dimensions and burnout, suggesting that millennials perceive job crafting as a constructive method to improve their overall satisfaction via creativity and exploring different ways of working. Our finding confirms the popular COR premise emphasising employee motivation to conserve and acquire resources to maintain work outcomes (Halbesleben et al. 2014). Another possible reason to explain these findings may be that the data was collected during the post-peak pandemic period, in which many SMEs still suffer from the severe resource crunch of the COVID-19 pandemic (Dejardin et al. 2023). Under such conditions, employees may feel particularly challenged to engage in job crafting, as organisational crises likely inhibit their ability to do so. Consequently, job crafting showed no impact on their burnout experiences.

Unlike previous studies that examine job crafting as an aggregate construct of the different dimensions (Cheng and O-Yang 2018; Singh and Singh 2018), our findings underscore the importance of structural and social resources in enhancing millennials’ job satisfaction. Once again, we ascribe these findings to the turbulent conditions SMEs faced during our data collection period. As posited in the COR theory, individuals require not only physical but also psychological resources to engage in job crafting (Hur et al. 2024). Hence, the lag in performance and the general financial difficulties faced across most SMEs may take a toll on employees’ psychological resources (Zutshi et al. 2021), causing them to gain job satisfaction only through increasing job resources. As increasing challenging job demands and decreasing hindering job demands may aggravate the already precarious situations faced by SMEs, employees did not report job satisfaction by engaging in these job crafting dimensions.

Although burnout has been recognised as a significant factor that can negatively affect employees’ job satisfaction (Ayachit and Chitta 2022; Cheng and O-Yang 2018), little is known about whether job satisfaction can mitigate burnout. By confirming that job satisfaction had a negative impact on millennials’ burnout experience, we highlight the importance of instigating affective emotions towards work among employees to buffer against burnout (Tsigilis et al. 2004; Wallace and Coughlan 2023). The full mediation effects of job satisfaction were only found between increasing structural job resources–burnout and increasing social job resources–burnout relationships. Such findings underscore that in times of crisis and environments of resource scarcity, millennials’ engagement in resource-seeking behaviour reduces their burnout experience primarily because of the sense of satisfaction gained. Ultimately, this study reveals the significance of job satisfaction as a key psychological resource that reduces millennials’ burnout experience, responding to the call for research to examine factors to mitigate burnout (Wallace and Coughlan 2023).

The findings yielded from this study offer practical implications for SMEs in managing the millennial cohort. Despite the challenges SMEs face due to resource constraints and structural limitations associated with their smaller size (Ali et al. 2020; Maden-Eyiusta and Alten 2023), these organisations should cultivate environments with a high acceptance level for job crafting. Employees may be reluctant to engage in job crafting if they are concerned that the changes they make could introduce organisational risks, potentially jeopardising their job security (Bizzi 2017). Hence, SMEs should focus on training and equipping employees with the knowledge and skills necessary to identify job crafting opportunities and effectively modify job resources and demands (Cheng and O-Yang 2018; Singh and Singh 2018). A strategy for SMEs to impart job crafting skills pertains to delegating temporary assignments to employees to expand their competencies and sense of identity outside of their stipulated roles (Rogiers et al. 2021). This strategy fosters a mindset that it is safe to explore alternative job competencies and identities, encouraging employees to seek resources and adjust demands proactively (Rogiers et al. 2021).

Drawing on the result that job satisfaction negatively impacts burnout, SMEs should also pay close attention to initiatives that can increase employee satisfaction. Implementing support programmes aimed at helping millennials with stress management could be a strategy to elevate their job satisfaction (Hammond et al. 2019). Considering the resource constraints of SMEs, they may not need to allocate large budgets to establish a formal stress management and wellness department. Leaders’ initiatives, such as offering coaching, informational support or holding regular conversations to assess employees’ needs for guidance or support (Abukhait et al. 2023; Chan et al. 2023), can foster a sense of care from managers, resulting in increased job satisfaction among employees. Most importantly, leaders need to emphasise the confidential nature of such initiatives so that employees feel psychologically safe to share their concerns (Hammond et al. 2019).

As with any study, our study has several limitations. First, our proposed model solely examines the impact of job crafting on job satisfaction and burnout among millennials. Job crafting, being an individual factor, is limited to individual efforts. However, it is crucial to acknowledge that relational factors (e.g., leader and co-worker support) and organisational factors (e.g., organisational culture and perceived organisational support) may also influence employees’ burnout experience (Ayachit and Chitta 2022; Slemp et al. 2015). Including these factors is necessary because job crafting results in proactive changes that can influence and be influenced by other stakeholders in the workplace (Bizzi 2017; Tims and Parker 2020). Therefore, future research should consider expanding the model to incorporate different levels of factors or moderators to better understand the mechanism of job crafting and burnout. Second, our exclusive emphasis on millennial samples employed in SMEs has the potential to undermine the generalisability of the findings. Empirical evidence suggests that generational identity influences employee attitudes and practices, affecting burnout experiences (Abate et al. 2018; Barboza-Wilkes et al. 2023). Hence, applying our findings to mitigate burnout among workers from other generations requires caution. To address this limitation, future studies should collect data from diverse generational cohorts and test the moderating effect of generation on models exploring employees’ burnout experience (Lu and Gursoy 2016). Furthermore, our study collected data solely from employees working in the SMEs context. Subsequently, these findings may not be generalisable to employees in large organisations or the public sector due to differences in available resource pools (Kim 2018). In light of this, we suggest future research replicating the model in different organisational contexts for comparison. Finally, the cross-sectional design employed in this study restricts the ability to establish causal relationships among the variables. Prior research has highlighted that job crafting has longer-term effects, which may deplete employees’ energy over time (Rudolph et al. 2017; Zhang and Parker 2019). Thus, future research may employ a longitudinal design to draw more robust inferences regarding the relationships between job crafting, job satisfaction, and burnout.