Enriching project organisations with formal change agents

Abstract

Design/methodology/approach

The paper is based on a literature review and a multiple-case study in which six organisations participated in an actionoriented research project. The aim for the organisations was to obtain a better health status among the employees by accomplishing an internal change project that enhanced physical activity at the workplace and in leisure time. Change agents in the form of peer health ambassadors were selected by middle management and hereafter trained by the project representatives.

Findings

The findings suggest that the selection of change agents and middle and top management support are major determinants of success within change projects. To select change agents that the employees respect and can identify with, combined with top management prioritisation, is important in order for the project organisation to benefit from the additional role.

Practical implications

Selecting the “wrong” change agents can jeopardise a change project, even when the project is supported by top management and the target group members at the starting point are highly motivated to change.

Originality/value

The research contributes to the understanding of project organising by building theory on how formal peer change agents can enhance project success in change projects.

Introduction

It is a well-established truth within project management literature that so-called “project product success” (Baccarini, 1999) in terms of aimed-for value creation and benefits for the project investor and (selected) project stakeholders may be difficult to obtain even though proper project management is carried out (Dupont and Eskerod, 2016). More authors, e.g. Burnes and Jackson (2011), and Charles and Dawson (2011), state that substantial evidence exists to the statement that about 70 per cent of all change initiatives fail. The so-called “project management success” (Munns and Bjeirmi, 1996) in terms of agreed upon deliverables delivered on time and within budget does not automatically lead to project product success, e.g. increased profit or reduced sick leave in all projects (Ward and Daniel, 2006; Andersen, 2008; Creasy and Anantatmula, 2013; Serrador and Turner, 2015). Thus, it seems logical to expect that project product success may be even more challenging when dealing with change projects in which people involved are required to change behaviour (e.g., safety-related behaviour) and sustain this behaviour in the future in order to achieve the benefits.

Even though new tendencies can be observed, a major part of the current project management literature concerning change projects focuses narrowly on the role of the project manager and the project team members within project planning, executing and controlling – and on the information and communication needs and personal traits of project managers (e.g. Thomas et al., 2012; Creasy and Anantatmula, 2013; Müller, 2015; Lee et al., 2015; Sun et al., 2015). Another stream of literature offers strong documentation on the positive influence of the top management’s role in change projects (e.g. Kotter, 1996; Gemünden, 2014; Packendorff et al., 2014; Sun et al., 2015). However, there is a gap in the literature when it comes to other supporting formal roles within change projects.

Andersen (2008) states that for an internal project, e.g. a change project, the responsibility for harvesting the benefits rests with the base organisation, i.e. the permanent organisation undertaking the project. This implies that in order to implement and sustain ownership of necessary changes, additional actions to the ones done by the members of the project organisation should be undertaken by people outside the project organisation (Fernandes et al., 2014). Burnes and Jackson (2011) add to this argumentation when they state that a lack of alignment between the value system of the change intervention and the value system of the members of the organisation who undergo the change, i.e. the base organisation, is critical for success.

This paper discusses how project success can be facilitated by establishing an additional formal role to support the members of the project organisation. Building on literature within change management (Burnes, 2009; Dawson, 2003; and Jones, 2006), it seems reasonable to assume that behavioural changes may be easier to reach if formal change agents are appointed, i.e. “internal or external individual[s] or team[s] responsible for initiating, sponsoring, directing, managing or implementing a specific change initiative, project or complete change programme” (Caldwell, 2003, pp 139- 140). The project manager and his/her project organisation may then focus on project planning and controlling, whereas the change agents focus entirely on the needed behavioural changes of the people involved, i.e. “the adopters of change” (McMaster et al., 2005) or in other words, the target group.

Change agent to support the project team

A classical role to support the project team would be a middle manager in the form of a line manager (Dupont and Eskerod, 2016), and a major part of the existing literature deals with the middle manager as a change agent (Andrews et al., 2008; Alfes et al., 2010; Bryant and Stensaker, 2011; Burnes, 2011; Charles and Dawson, 2011; Thomas et al., 2012; Pollack and Algeo, 2014; Hornstein, 2015; Packendorff et al., 2014; Ekrot et al., 2016). This paper, however, adds to the literature by investigating peers as formal change agents that can support the project team. Researchers have identified the following justifications for the use of formally appointed peers: cost-effectiveness (in comparison with external consultants); credibility of the formally appointed within the peer group; empowerment of the peers formally appointed; utilisation of already existing means for information sharing and advice giving; more successful than professionals on passing information; positive role models; beneficial to those who are formally appointed; more accepted as educators; can reach hardto- reach target group members; and reinforcement possible through ongoing contact (Turner and Shepherd, 1999). Furthermore, recent calls for a more ethical approach to change management (By, 2005; Burnes and By, 2012; By et al., 2012) highlight the importance of involving all stakeholder groups in the management of the change process to avoid changes driven by unethical motives.

The major reasons for selecting peers as change agents to support project teams are various (Eskerod et al., 2015). The most important thing is that the peers are familiar with the company culture in the focal organisation, and they may know the target group colleagues in advance, so they share a common past and a future. Also, they may be able to combine the task as change agent with their regular job so that they can be flexible in terms of change agent efforts (number of hours, duration etc.) due to needs. Typically, they cost less than an external consultant. They meet the target group in various forums on a regular basis (like work situations, lunch and coffee breaks, in the corridors or in the company’s fitness facilities). Due to their knowledge about the focal organisation they can act as efficient and effective liaisons between the target group and other individuals, for instance, groups or units within the permanent organisation, e.g. top managers, middle managers and the HR department as well as between the target group and members of the project organisation, e.g. the project manager.

It is characteristic of peers to not have any leadership authority toward the members of the target group. Another characteristic is that they may have a deeper understanding and empathy of the members of the target group than an external person would. Furthermore, the research of Pollack and Algeo (2014) reveals that there is currently a lack of clarity as to how the roles of change managers and project managers should interact in an organisational setting, especially when it comes to reporting functions. It could be argued that tasking peers with change agent efforts avoids this confusion because their reporting line is still clear as they are not given a project team function per se but keep reporting to their manager in the line organisation.

Even though peers within an organisation seem to have great potential as supplements to the members of the project organisation, this type of change agent has to the authors’ knowledge neither been discussed thoroughly in the literature (an exception is Dupont and Eskerod, 2016), nor has it been an object of empirical studies when it comes to undertaking change projects.

Research questions

This generates a number of questions: How can formal peer change agents enhance project success in the form of compliance with aimed-for behavioural changes of the target group in change projects? Which skills do change agents need in order to succeed? How do companies select the right change agents? What is needed from the context to enhance success?

This paper answers the following overall question:

• RQ1. How can formal peer change agents enhance project succes in organisational change projects?
  And the sub-research question:
• RQ2. What is needed from the context to enhance project sucess based on formal peer change agents?

The research questions were addressed by an exploratory study nested within a larger study in which the overall purpose was to investigate the effects on health and productivity of implementing and embedding physical exercise training among office workers at the workplace in Denmark. The larger study involved a representative two-year randomised control trial (RCT) study that was performed in Denmark. The aim was to study whether individually tailored worksite-based physical training among workers with inactive job categories would: improve cardiorespiratory fitness and/or individual health risk indicators; improve muscle strength and decrease musculoskeletal disorders; succeed in regular adherence to worksite and leisure physical activity (PA) training; and reduce sickness absence and productivity losses (presenteeism) in office workers. The results of the larger study are referred elsewhere (e.g. Sjøgaard and Justesen, 2014; Dalager et al., 2016). This paper focuses instead on the efforts of the formal peer agents to enhance the target group members to undertake more physical training during the project course.

Contribution and outline

The research contributes to the project management field by building theory on how formal peer change agents can enhance aimed-for changed behaviour in projects. Being the result of a collaboration between clinical biomechanics and project management researchers, this paper also addresses what Gemünden (2014) names as a “plea for more cooperation with researchers outside our comparatively small project management research community” (p. 4).

The research represents an innovative approach to project management research in two ways: it builds a bridge between the project management literature and the change management literature by investigating the role of formal peer change agents as a supplement to the classical roles within the project organisation; and it is based on an action research approach in which comparative analyses are made possible due to an explicit and strict set-up of the study design in six case companies (see methodology section).

Outline of the paper: first, the theoretical frame of reference underlying the research is presented. This is based on literature studies of streams of literature within change management and peer agency. Second, the research design is presented. Third, there is a presentation and discussion of the findings from analysing the experiences of appointing peers as change agents in projects run by the six case companies. In this section, theoretical implications of the findings are also suggested. Finally, the conclusion provides a short summary of the research and offers suggestions for future paths to be taken.

Theoretical frame of reference

The purpose of this section is to establish the conceptual framework for understanding the assumptions underlying the belief that peers as formal change agents can be used to enhance project success in terms of both project management success (on time, within budget and to specification) and project product success (in this case, behavioural changes that are sustained over time).

Peer education can be defined as “ sharing our experiences and learning from others like us” (Robins, 1994, p. 2). This type of education has been popular within schools and youth services since the 1960s, starting out within drug prevention efforts and gaining especially intense attention in practice and in the literature of the 1990s (Milburn, 1995; Norman, 1998). The use of peers is also well-known in the academic world, where peer reviewing is a common phenomenon.

A rationale for using peers is tied to the notion of identity (Turner and Shepherd, 1999). The peer status in health promotion can be seen as an alternative to the “expert” status of a health education professional (Shiner, 1999). According to the rationale of identity, which builds on social learning theory, the peers are consciously or unconsciously perceived as role models by the target group members (Bandura, 1977; Dupont and Eskerod, 2016). A role model should be seen as attractive by the target group member, and attractiveness should be understood in broad terms (Eskerod and Jepsen, 2013). It may be that the target group member and the role model share certain characteristics which the target group member can identify with, or that the role model has characteristics which the target group member aspires to have him- or herself. “People are more likely to hear and personalise messages resulting in changing attitudes and behaviours if they believe the messenger is similar to themselves and faces the same concerns and pressures” (Gartner and Riessman, 1999, pp. 5-6).

Another argument for the potential of peer change agents is that peers can act as influential spokespersons for a change initiative if they are seen as trustworthy and as someone the target group members would like to be on good terms with (Eskerod and Jepsen, 2013).

Peer change ambassadors can be seen as a structural supplement to the classical roles within a project organisation, e.g. project manager, project owner, project team member, project contributor and reference group member (ref after review). On the other hand, they can be seen as alternatives to internal or external consultants (Grima and Trépo, 2011). The peer change agents become the link between those who want the change (e.g. top management) and the adopters of the change (employees) (McMaster et al., 2005).

Competence development (education and involvement) of change agents is a core element when it comes to successful implementation of change (Burnes, 2009; Gareis, 2010; Andersson, 2016).

To enhance the understanding of the nature of peer involvement, this paper builds on the classical conceptual work of Shiner (1999) who offers the distinction between “peer delivery” and “peer development.” According to Shiner (1999), the approach “peer delivery”, i.e. providing knowledge, facts, practical pieces of advice as well as formal sessions to the target group members, is the most common way of understanding the peer education concept. Whereas the usage of the approach “peer development” as an inspirational source, i.e. making the peer change agents develop their own situation in a positive way by identifying their own training needs, negotiating these with trainers and advancing their own interests concerning the peer activities is less commonly reported in literature. “Peer delivery” calls for project management and communication skills, whereas “peer development” relates to the potentials of being a role model who also changes health behaviour and develops him- or herself based on the health promotion initiative.

Drawing on the work of Pettigrew (2000), Weick (2000) and Burnes and By (2012), the change agent is not a neutral facilitator but a facilitator with an agenda which he or she at all times tries to fulfil. To follow the characterisation of the facilitator with an agenda, Buchanan and Boddy (1992) argue that the change agent has to work with the three logics of change: problem-solving, ownership and legitimacy.

However, existing change models focusing on the leadership and management of change leave out a thorough description of the role of the change agent (Burnes, 2009) even though Dawson (2003, p. 33) has “identified [the change agent] as a central element in the power plays and political manoeuvring of individuals and groups during programmes of change.”

When working with culture change projects, middle managers are the only group that can initiate change (Barton and Abrosini, 2013; Burnes, 2009). Middle managers’ job is to execute strategies and deliver results in time to fulfil the goals (vision) of the company.

To elaborate on the work of Caldwell (2003) and Jones (2006), the middle managers are in charge of organisational change at the various organisational levels, and in cultural change projects where employees need to change behaviour, change agents are selected by the managers and are in charge of the change on individual level as facilitators of change. The selected peers are trained on peer level in order to succeed as change agents. The change agents then become the link between those who want the change (steering committee) and the adopters of the change (employees) (McMaster et al., 2005).

In Abraham and Michie’s taxonomy (2008) on behaviour change techniques, three of their 26 techniques relate particularly well when using formal peer change agents. These three are as follows:

1. The usage of follow-up prompts, i.e. the target group member is contacted by the peer change agent during the course of a change programme.
2. Provision of opportunities for social comparison, i.e. the target group member can observe a non-expert’s, i.e. the peer change agent’s, performance related to the aimed-for changes.
3. Provision of social support, i.e. prompting consideration of how the target group member can change his/her behaviour by the peer change agent offering the member help or social support.

Institutional theory (Scott, 1995, 2014) is another relevant theoretical framework for this research. Scott (1995, 2014) suggests that institution building and maintaining (as well as destruction) is based on three pillars: rules and regulations (i.e. governance issues), norms (i.e. what is suitable and legitimate in the situation at hand) and cognitive-cultural (i.e. how people make sense of rules in terms of the prevailing (organisational) culture’s (and their own) norms).

In other words, “institutional theory and analysis address the processes by which social structures, including both normative and behavioural systems, are established, become stable and undergo changes over time” (Scott, 2012, p. 29).

In the research project at hand, it seems important that all three pillars should be supportive in order to create the commitment of the target group to trust and comply with the inputs of the peer change agents and that both top management and middle management have roles in enhancing the existence of the pillars.

Institutional theory has already been used by a number of other project management researchers (see e.g. Engwall, 2003; Morris and Geraldi, 2011; Müller et al., 2014) and is as such an established theoretical frame within the field.

Research methodology

The empirical part of the research was based on a longitudinal multiple-case study carried out in an action research study design. As mentioned in the introduction section, this study was embedded within a larger study involving a representative two-year RCT study. The aim was to develop theory on change agents’ influence on project success by applying case-study research methodology inspired by the work of Eisenhardt and Graebner (2007). The action research design was chosen because the authors believed that only through a long-term collaboration and partnership with the companies and the change agents would it be possible to better understand the role of change agents in a project setting as well as the inherent challenges.

Danish, Dutch and American studies have documented an increase in activity (Proper et al., 2006), increase in efficiency, increase in quality of work and a decrease in sick leave when PA is implemented at the workplace (Davis et al., 1987; Galinsky et al., 2007; Pronk et al., 2004; Bredahl et al., 2015). Furthermore, Danish studies have shown that active employees have a higher energy surplus and are less stressed (Hansen et al., 2010). Worldwide, many companies invest a considerable amount of money in health promotion activities. However, the benefits (for the employees personally as well as for the companies) cannot be achieved unless the employees change behaviour, i.e. start to exercise more. For some of the employees, substantial behavioural changes are needed, as they may not be physically active at all in their daily life. In addition, the increased PA level needs to be sustained in order to harvest benefits on a continuous basis. The request for changed behaviour is also in a grey zone between working life and private life. Despite the stipulated positive effects of health promotion efforts for the employees, it can be questioned whether it is ethically acceptable for employers to ask for the needed behavioural changes or it should be outside the scope of the organisation’s interferences and left to the private life. This question is especially relevant if physical activities (PA) are not directly job-related (in opposition to the case for policemen or pilots where it seems logical to require fulfilment of certain health performance standards). While there are many discussions on the topic of ethical change and the motivations behind organisational change (By, 2005; Burnes and By, 2012; By et al., 2012), the question of ethics falls outside the scope of this paper.

Based on the above argumentation, health promotion projects at the workplace (implementing health at the workplace) are found to be very suitable as an empirical setting for investigating formal peer change agents. The study was part of six health intervention projects (lasting two years each) at six Danish companies. The final aim of the projects was to investigate whether two years of exercise at the workplace had a positive effect on office workers’ productivity, sick leave, individual health and retention of employees. The additional aim of the project was to investigate how to successfully implement and embed healthcare at the workplace by combining project management and the usage of peer change agents in the form of selected and trained employees. The project was an interdisciplinary study in health science and social science.

Recruiting case companies

In may 2010, 103 companies across Denmark were contracted by email to determine their interest in this study. The project manager (who is also one of the authors) had a previous business relationship with each of the companies. The nature of this relationship was that the author had either previously taught project management to their employees or had acted as a health promotion consultant. A total of 17 companies expressed their interest, and six of these agreed to participate in the study. Moreover, ten of the remaining 11 companies interested in the project wanted to join at a later stage but were not included in this study. The six companies were located across Denmark. Two were private companies (a telecommunications company and a food company), two public municipalities and two National Boards of Social Services. The employees at all six companies included in the study were office workers according to the inclusion criteria, and none of them had jobs with a specific focus on health.

Recruiting participants

The target group members, i.e. the office workers, worked at least 25 hours a week, were not pregnant and did not suffer from certain specified serious diseases. Following baseline measurement, participants were randomised to either a PA training group (TG) (n¼194) or a control group (n¼195). The study underlying this paper on the peers as formal change agents at the workplaces was a sub-part of the randomised study and includes only findings from the participants randomised to the TG as well as those appointed as change agents called health ambassadors (HAs). This part is inspired by Eisenhardt and Graebner’s work (2007) on longitudinal multiple-case studies. The HAs’ task was to motivate colleagues to become and remain physically active during the research project and to co-ordinate and implement health promotion activities in the workplace in general. Among the eligible employees, middle managers appointed peers to act as HAs. Middle managers were asked to select HAs among employees with at least five years of seniority within the organisation. As criteria for selection, the middle managers were asked to choose employees they believed would be able to motivate their colleagues and have the drive to take initiatives, such as motivating colleagues in the TG to undertake PA, 30 minutes of moderate intensity workout six days a week, during the course of the research project. Furthermore, the persons appointed for the HA role had to be able and to have permission to allocate two hours a week during work time for two years. In addition, the middle managers were instructed not to appoint “fanatic” sports persons, i.e. persons who already did extraordinary many hours of PA prior to the project. Instead, the HAs were supposed to represent the “average” employee when it came to PA, thus providing a better basis for role modelling. The HAs trained together with the PA training group but were not part of the randomised TG. In total, 17 HAs were appointed, each covering 10-15 of the target group members, i.e. office colleagues in each HA’s office.

All participants gave their written informed consent, and the local ethics committee approved the study protocol.

Interventions

The training intervention and the theoretical framework of Intelligent Physical Exercise Training has been described in other publications (e.g. Sjøgaard and Justesen, 2014; Dalager et al., 2016). In short, all sessions were one hour long (50-minute training sessions – allowing 10 minutes for getting to and from the training area). Each employee received an individually tailored training programme based on the outcome measures of a health check and a questionnaire performed at baseline. For each measurement, cut-off points were identified to allocate individual training duration and intensity within cardio, strength and/or functional training.

The purpose of the HA training was to educate HAs to motivate and implement PA for their colleagues in the workplace. The training was based on evidence-based principles and built on the concepts of both “peer delivery” and “peer development” (Shiner, 1999) (see theoretical framework). All training sessions had a practical focus in which the HAs tried out the theories in practice using the learning and training principles described by Brinkerhoff and Mooney (2008). Table I describes the contents of the four days of training of the peer HAs.

Follow-up meetings

The purpose of the follow-up meetings was to support the HAs and collect data for the study by addressing the challenges they were facing as well as getting insight into their experiences with and thoughts about undertaking the role as change agent. All meetings had the same agenda (everybody presented (1) good stories (2) challenges and (3) how to deal with the challenges until the next meeting). The meetings were facilitated by one of the researchers. The researchers’ role was not to be an expert in the field of implementing health in the workplace but to help the HAs with their job by facilitating their own reflections and problem-solving. Follow-up meetings were held every quarter during the two-year intervention period, and all HAs participated in all the meetings.

Data collection

More means of data collection were applied in order to increase credibility and validity of the results by data triangulation (Cohen and Manion, 2000), and both quantitative and qualitative data were collected. Surveys (using SurveyExact), individual semi-structured interviews and focus group interviews were conducted. The surveys, interviews and focus group interviews were constructed and conducted together with graduate students, and e-mail was used for surveys.

One survey was aimed at the TG and concerned their views on the HAs’ influence on health promotion and health behaviour at the workplace. In total, 137 employees participated, giving an 80 per cent response rate. The employees answered 85 questions all together, and two of the questions were specifically focused on the work of the HAs and the HAs’ influence on health promotion and health behaviour in the workplace. The two questions were as follows:

To what extent did the HAs influence the health promotion activities at the workplace during the research period?
To what extent did the HAs influence your own health behaviour?
The two questions were scaled questions from 0 to 10, where 0 represented “no influence” and 10 “strong influence.”

Another survey and all interviews were aimed at the HAs. These means of data collection concerned the HAs’ views on the training as well as on their activities and perceived challenges in the role as HAs. All 17 HAs participated in the survey, giving a 100 per cent response rate. Focus group interviews were held at four of the six companies, interviewing ten HAs, i.e. three employees in each of two companies and two employees in each of two other companies, and individual semi-structured interviews were held in the last two companies, interviewing the remaining seven HAs.

Data analysis

In both the quantitative and the qualitative data analyses, a considerable number of comparison studies were made due to the potential differences and similarities across the TG and the control group, the HAs and the TG and the various case companies.

All quantitative data analyses were performed using SPSS statistical software, version 21. In our data reporting, data are shown as means and standard deviation (SD). Results were considered statistically significant if the two-tailed p-value was <0.05. In addition, quantitative analyses on the demographics of the HAs and the employees in the TG were undertaken in order to investigate whether the HAs’ demographics (as intended) were similar to those of TG.

As the source for the qualitative data analyses, we used written transcripts of the interviews/focus group meetings, as they were all recorded. Two of the authors as well as master students who undertook interviews coded the data individually and manually (inspired by Ramian, 2012; Brinkmann and Kvale, 2014) by the use of the method “systemic text condensation” (Malterud, 2012). The purpose was to identify categories of actions and opinions related to the change agents, i.e. the HAs, first as a within-case analysis in each of the case companies, and afterward as a cross-case analysis (Eisenhardt and Graebner, 2007). The cross-case analysis of the emergent categories formed the basis for the propositions developed. Afterward, meta-matrices that linked quotes from the cases to each proposition were constructed. Direct quotes were selected for the presentation of each proposition in this paper.

Findings and discussions

To investigate the potentials of the selected peer change agents for being a role model due to demographics, a comparison study of the two groups was undertaken. The study showed that the demographics of the two groups, i.e. the peer change agents and the employees in the target group, were similar, as intended in the research design. Participants were on average 44±10.4 years old, 75 per cent being female. The participants had an average body mass index of 25.4±5.1 kg/m2, an average percentage of body fat of 29.1±8.8 per cent and average steady state HR during the submaximal bicycle test for estimating VO2 max of 146±12.7 bpm. The results imply that a potential lack of identification with the selected peer change agents cannot be attributed to differences in demographics.

The quantitative study

For Question 1 in the quantitative study, “To what extent did the health ambassadors influence the health promotion activities at the workplace during the research period?”, the mean answer was 5.2 (SD¼2.9). For Question 2, “To what extent did the health ambassadors influence your own health behaviour?”, the mean answer was 3.9 (SD¼2.8). For neither of the questions, significant differences between men and women or across organisations were identified. See Table II.

The answers to both questions in the quantitative study imply that the HAs have had some influence on both the health promotion activities and the individual behaviour of the target group members. However, the numbers clearly show that there is room for improvement, especially when it comes to influencing health behaviour.

The qualitative study

The qualitative study revealed a number of issues related to selecting and educating persons for the role as formal peer change agent, i.e. health ambassador, at the workplace. The issues are discussed in the sections below. Based on the empirical study, we claim that three areas seem especially important in order to enhance project success, in the form of both project management success and project product success, when appointing peers as change agents in change projects. The areas are as follows: selection of change agents; role, tasks and skills of the change agent; and support needs of the change agent, managerial and educational.

Selection of change agents

A key finding from the literature review was that success of a peer change agent would rely heavily on the change agent being perceived as a role model or a spokesperson that the TG member would like to resemble or to be on good terms.

The qualitative study revealed that tension and conflict between some of the HAs and some of the TG members was present due to lack of respect of the HAs as well as envy. Even though the quantitative study showed similarities between the demographics of the HAs and the TG members, some of the TG members thought that middle management had appointed the wrong people for the role and felt they could be better HAs themselves. In order to promote identification between the HAs and the TG members, the middle managers were asked to select the ambassadors on the basis of their lifestyle in terms of being interested in health but not “fanatic.” Some of the TG members responded positively to this selection criteria, while others responded negatively. Those with negative responses could not identify with the appointed HAs and would not seek or take any of their advice. This can be seen in the following statements by some of the HAs:

"There was a lot of criticism in the beginning because I had been selected [by middle management] for the role and not been elected [by the employees]. A lot of colleagues thought that they would be better suited [for the role as HA]. They couldn’t understand why it was me [who had been selected]."

"Some active [colleagues] thought that I, as an HA, should be a front-runner. So the fact that I had difficulties when it came to meeting at 8 am [the time at which a health activity started] resulted in many negative [comments]. I was supposed to be “the holy” and “the good” who worked hard – and preferably did better than the others."

"[In my role as peer HA] I wanted to be the way [the colleagues] wanted me to be, but that was difficult as they didn’t all want the same thing."

"I’m a bit afraid of the comments [about me not being the right person to be HA] […]. If too many of these [comments] pop up, it pains me, and therefore I chose not to [take on a certain activity related to the project]."

The findings show that the peer change agent’s own attitude towards the change, in this case improved health through increased PA, combined with “not being fanatic” was not enough to select effective change agents. It seems important that the peer change agent had a natural interest in health and was enthusiastic about it, but it was even more crucial that colleagues could relate to and compare themselves with the change agent and that the person had a high informal rank in both the social and professional networks at the workplace.

The findings relate to Walker and Avis’ (1999) statement that being involved in peer education does not automatically make the person participating empowered. To be successful, the person must be perceived as a role model, be a spokesperson and be thought of as resembling the target group (Eskerod and Jepsen, 2013). Furthermore, the findings relate to the work of Nikolaou et al. (2007), who point to the importance of a relationship based on trust between the peer change agent and the target group, as well as to Buchanan and Boddy (1992), who point to the importance of legitimacy, and Bandura (1977), who states that to be a credible role model the selected peer must have a high status within the peer group. In addition, it relates to Ryan and Kossek (2008), who discuss the need for considering whether the implementation of the change initiative breaks down or creates barriers when it comes to inclusion of all the employees (and e.g. their different health conditions and lifestyles), and to Ekrot et al. (2016), who explore the influence that organisational-based self-esteem (OBSE) has on the voice and success of change managers. Finally, it relates to the two of the three pillars in the institutional theory (Scott, 1995, 2012, 2014), i.e. the norms pillar and the cognitive-cultural pillar. The target group members could simply not accept being guided by persons they believed had a lower physical training status than themselves. They thought this was not complying with the cultural norms on role modelling of the society or of their own understanding on what made sense.

Based on these insights, the first proposition is the following:

• P1. In order for a formal peer change agent to be influential on peers’ behaviour, the target group must be able to identify with the peer, i.e. see the peer as a role model, or at least respect the peer as a spokesperson of the change project.

Role, tasks and skills of the change agent

Shiner (1999) points to the fact that the definition of peer education in change management literature is not clear, and that various peer approaches can be chosen. Tucker et al. (2015) also focus on the importance of properly communicating role expectations to change agents. This was confirmed in the empirical study. Even though it was thought that a clear description of the role as peer health ambassador had been made, many different understandings of the role among the 17 appointed ambassadors were observed. Shiner (1999) claims that it especially is the nature of peer involvement that needs to be scrutinised, investigated and defined more clearly. He contributes to this by offering a distinction between “peer delivery” and “peer development.” “Peer delivery,” which according to Shiner (1999) is the most common way of understanding the peer education concept, is about providing knowledge, facts, practical pieces of advice as well as formal sessions to the TG members. On the other hand, “peer development” is about the peer change agents developing their own situation in a positive way by identifying their own training needs, negotiate with trainers for that and go on with their own interests concerning the peer activities. The former approach calls for project management and communication skills, whereas the latter relates to the potentials of being a role model (as already presented in the former section). Both were seen in the empirical study, and here are the examples delivered by the peer HAs:

Peer delivery activities:

"I have made all the material about our project and posted it on the net. Purpose, background, who is participating […] and then people can read it if they like."

"In the beginning, we put up posters with information on how much [extra] exercise you need to do when you drink cola or eat candy."

"We need all the time to make people aware that we still are here. [This is done by taking initiatives, e.g. through the intra-/internet or by arranging events]."

"They [the TG members] often ask about various practical issues. Then I send an e-mail to all of them."

"My primary role is to be primus motor and liaison between [the project manager], the [training] instructor, the participants in the TG and own management. ONE local coordinator […]. My secondary role is to initiate health-promoting activities at the workplace which are beneficial for all employees, like elastic band exercises, registration for various company sport campaigns etc. […]."

"If I can see that something is about to be derailed, it is me who takes it up […]. It is also me who takes initiative to [collect] training stories from the members of the [target] group. I go out and ask them about their stories on what they think about it [being part of the health promotion project], write them down and post them on the intranet so that we, all the time, can make people aware about what is happening and keep the pot boiling. It is me who is the driver."

Peer development activities:

"I could have [intervened more directly with the individual participants]. I expect this will come [during the project]. It is not natural for me [and my personality], it is something I have to learn."

Based on the above findings, the following is suggested:

• P2. In order to be successful, the peer change agent must be able to be a driving force with proper project management and communications skills.

This relates to the current literature on change management, e.g., Rogers (1995) who points to the need for being able to diagnose problems and to establish information exchanges as well as to create an intent to change and translate that intent into action. It further relates to Buchanan and Boddy (1992) who emphasise the importance of problemsolving skills and Pollack et al. (2013) who focus on the role of stable networks of association and interaction through action network theory.

On top of the mentioned activities and attitudes, there were additional activities not covered by the “peer delivery” and “peer development” concepts offered by Shiner (1999). It is suggested that these activities be classified under a new heading: “peer attentiveness.” Peer attentiveness means that the peer change agent is observant on the TG members’ compliance with the agreed upon behavioural changes and dare to ask if compliance is not shown, as well as the peer change agent is able to feel empathy with the TG members and therefore supports them in a positive way instead of getting annoying with the non-compliance. This may be easier when people know each other (due to being colleagues and having a present and a past together) and when they expect to interact in the future (also due to being colleagues in the future to come). Based on the feedback from the 17 HAs in the survey and the six HAs in the interviews, it appears that skills related to “peer attentiveness” are important in order to be a successful peer change ambassador:

"The times when she [the TG member] has not been attending [for weekly training offered for the TG and HAs], I have approached her [in the office] and said that we missed her Thursday and asked her where she had been."

"If you can see they don’t attend [training], you approach them and talk about it."

"You don’t have to be negative even though you think it is not good enough that the person the last 10 times not has been attending training. Instead, you can turn it around and talk with the person about why and whether I [as the peer HA] can do something to make the person come to the training."

"[When some TG members and the HA were on a bicycling trip] and I could see that we approached some hills that were not so good [because she was not able to cope with the physical strain], I put a hand on her back and helped her up the hill."

"If you meet them in the city, you can ask them whether it hurts in “the grandmother arms” today, and you would never do this if you were not participating in the project."

"When asking why [concerning noncompliance], you can get closer to some of [the colleagues] than others."

"I think that you need to know “your people” in order to know what to do to motivate them."

The above statements relate to Nikolaou et al. (2007) who point to the importance of a relationship based on trust between the peer change agent and the TG members, as well as to Buchanan and Boddy (1992) who point to the importance of legitimacy. It also relates to Ryan and Kossek (2008) who discuss the need for considering whether the implementation of the change initiative break down or create barriers when it comes to inclusiveness of all the employees (e.g. their different health conditions and lifestyle). More recent work by Müller (2015) investigates knowledge sharing within an organisation, stating that it is higher when there is trust from both the managers and their employees and between employees.

Based on the findings and literature, another proposition is made:

• P3. In order to be successful, the peer change agent needs to be attentive to and explicit about lack of compliance as well as feel and show empathy with the TG members and build up trustful relationships.

Related to the above observations about the nature of the peer involvement, the interviews clearly stated that the HAs understand their role as change agents very differently when it came to the extent of “pushiness” when being attentive. This divided them into two groups – as the following statements show:

"I only do something if my colleagues come to me with problems."

"It is evident for me that I contact my colleagues in need of motivation and guidance – I want to be proactive."

"I could probably be more “pushy” towards people. But I don’t feel comfortable in the role as the one with a whip or a bitch who constantly has to remind grown-ups to do this or that."

It is a crucial finding that half of the HAs found that their job was to wait for colleagues to come to them with problems or issues regarding health at the workplace. However, the data did not show whether these HAs were less successful in terms of compliance of their target group members than their more “pushy” colleagues. This would be an important subject for additional research activities in this or other research projects. The empirical studies showed that an important part of the peer change agent task is to negotiate and influence within the organisation so that room is created for the changed behaviour, i.e. increasing PA at the workplace and during work hours, to take place. Examples of related statements are as follows:

"Planning is important in a department with many meetings."

"Sometimes I have had to “play hard against hard” because [middle managers mentioned] their schedules etc. I am not willing to accept that when the managing director has said that it is possible [to include training during the workdays]."

"If I have seen that [the TG members’] calendar was booked with all sorts of other activities than training, I have approached them and asked for the reasons for not attending training. If they have had [jobrelated] excuses I have said that they have full commitment from top management, and that no other meetings should be scheduled at the time when they are supposed to train. This has been helpful to some of the participants, and then it hasn’t been a problem to reschedule the meetings."

Based on these findings, it can be claimed that it is important that the peer change agent understands the culture of the workplace and is able to negotiate and influence the target group members as well as middle management. This may be easier if the change agent has been employed at the workplace for a certain amount of time. To support these skills, a selection criterion for the research project was that the change agents should have been with the company at least five years. This was fulfilled for two-thirds of the change agents. This leads to the following:

• P4. In order to be successful, the peer change agent needs to have negotiation and influence skills (backstage skills) as well as in-depth knowledge about the company culture.

This relates to Parkin and Mc.Keagany (2000) who point to the importance of differentiating between initiatives aimed at changing community norms (like making managers and employees stop scheduling meetings at the time where the training is scheduled in the case companies) and initiatives targeted at the level of the individuals (like discussed in P3).

Support needs of the change agent, managerial and educational

In terms of managerial support, literature shows that change initiatives can be supported from various management levels (Caldwell, 2003; Venugopal and Rao, 2011; Hornstein, 2015). In all six case companies, top management support was observed. However, the analysis showed big variations in middle management support. To the question “What is needed in order for you to obtain success in your role as health ambassador?”, many of the peer HAs pointed to the need of more (middle) management support:

"Commitment from my boss – who absolutely does not think that it is a good idea that we use our work time for healthrelated training."

"That we, in reality, have time for it – also the employees [not only the health ambassadors]."

"Management commitment – [incl.] participation in different activities. So that it is visible that management prioritises it and finds time for it. That work time will be scheduled to activity planning."

"Full commitment from management."

"That we prioritise it at the highest level, also resource-wise."

"That [middle] managers get a thorough instruction from top management about the importance of this."

The data show that change agents cannot stand alone in order to succeed with the aimed-for change over time. Change agents implement individual change in co-operation with their middle managers, and top management implements organisational change with a project manager in charge of the process (not a neutral facilitator). Furthermore, for HAs to succeed middle managers, they need tools to follow up on their activities and measure the results. Turning back to P1, middle management also has a major task in selecting proper peer change ambassadors. In addition, institutional theory (Scott, 1995, 2012, 2014) would suggest that supporting rules and regulations (one of the pillars mentioned in the theoretical framework) and top and middle management enhancing supporting norms (another of the pillars) would be helpful. Therefore, the following are proposed:

• P5. In order to be successful, the peer change agent needs middle management support, including willingness to do a careful selection of peer change agents and to schedule for the change-related activities during working hours as well as participation in the activities.
• P6. In order to be successful, the peer change agent needs top management support, including explicitly stating the importance and priority of the change project – for both employees and middle managers – as well as participation in the related activities.

This relates to Caldwell (2003) who points to the need to differentiate between leadership change agents (understood as top managers) and management change agents (understood as middle managers) to support successful change implementation.

In terms of educational support, the aim of the four-day training programme in the research project was to give the HAs knowledge of and tools for how to motivate their colleagues and initiate health activities at the workplace. The findings show that the HAs found the education helpful. One stated the following:

"I can’t remember the name of the model, but the wheel with the different processes [i.e. stages of change] a person has to go through in order to change behaviour has worked for me. In the beginning, it was hard to use the models and theories in practice, but I looked at the material from the training programme, and that helped me getting started."

However, the ambassadors would like to have more theory, practice tools and skills training on two issues: (1) how to give feedback concerning observed changed health behaviour as well as observed non-compliance and (2) dealing with other stakeholders within the organisation, e.g., middle management.

Concerning (1), i.e. how to give feedback concerning negatively changed behaviour and non-compliance, the degree of “pushiness” toward the target group members, i.e. the usage of personal follow-up prompts, was an issue more ambassadors were struggling with, as can be seen from these statements (and was also mentioned in the section above):

"I could have [intervened more directly with the individual participants]. I expect this will come [during the project]. It’s not natural for me [to do personal prompting], it’s something I have to learn."

"I could probably be more 'pushy' towards people. But I don’t feel comfortable in the role as the one with a whip or a bitch who constantly has to remind grown-ups to do this or that."

"I only do something if my colleagues come to me with problems."

Concerning (2), i.e. dealing with other stakeholders within the organisation, the ambassadors expressed that they needed skills on how to negotiate and influence management levels within the organisation to create time for PA during work hours:

"Sometimes I have had to take a hard line because [middle managers pointed to] their schedules etc. I’m not willing to accept that, when the managing director has said that it is possible [to include training during the workdays]."

Based on the findings, it can be said that it is important that the HA education includes theory, practice tools and skills training on how to deal with resistance in the form of non-compliance from target group members as well as from middle managers. This relates to literature on change management, e.g. Rogers (1995) who points to the need to be able to diagnose problems as well as to create an intention to change and translate that intention into action; Buchanan and Boddy (1992), who emphasise the importance of problem-solving skills; and Ekrot et al. (2016), who stress the need for enhancing organisational skills in employees to enhance their OBSE.

Finally, the ambassadors pointed to the need for follow-up on their training:

"We need a follow-up every six months to stay focused. [Such a follow-up will] at the same time […] motivate us to do more."

This relates to the work of Gareis (2010) and Michel et al. (2013) who point out that education is a core element when it comes to successful implementation of change and sustaining momentum. It also relates to the work of Walker and Avis (1999) who state that many projects involving peers underestimate the training required.

Based on the above, the following is the final proposition:

• P7. In order to be influential, the HA needs theory, practice tools and skills training on how to give feedback on changed health behaviour as well as non-compliance and how to negotiate with management in order to create room for workplace health promotion. This should not only be during the project course but also as post-project follow-up training.

All the propositions developed can be seen in Table III.

Concluding remarks

The aim of this paper was to provide answers for these two research questions:

• RQ1. How can formal peer change agents enhance project success in change projects?
• RQ1a. What is needed from the context to ehance project success based on formal peer change agents?

An action-oriented case study involving six case companies was carried out. Peers were appointed as formal change agents called HAs, and their job was to enhance behavioural changes in the form of more PA at the workplace and at home during the project course.

The research shows that formal peer change agents can enhance project success in the form of sustained behavioural change among the target group members during the change projects. This happens through being role models for their peers and by approaching the peers assigned to them in the project in many arenas and in many ways when they observe non-compliance with the agreed-upon behavioural changes, e.g. participation in fitness training during working hours. Furthermore, the peer change agents can support external consultants, the project manager and project team due to their knowledge and insights of the company culture, managers and target group members. For formal peer change agents to be successful in enhancing changed behaviour of their peers, proper selection of the change agents is important. The change agent must be accepted as a role model for the target group members or at least as a spokesperson of the project. The analysis shows that selecting the “wrong” change agents can heavily jeopardise a change project, even when the project is supported by top management and the target group members at the starting point are highly motivated to change their health behaviour. The findings point to the relevance of including concepts like informal status and power in a theoretical framework on formal peer change agents.

In addition to the acceptance by the peers as role model or spokesperson for the project, the change agents should possess a variety of skills, including project management and communication skills combined with negotiation and influence skills, so that they are able to work not only on the individual level (peer-to-peer) but also on the organisational level, influencing company norms. At the same time, they should dare approaching non-compliant participants as well as managers showing resistance to or ignorance of the project. But also feeling empathy with other people is an important characteristic in order to support the target group members sufficiently. Furthermore, the findings show that education of the change agents is needed. This may include theory, practice tools and skills training on giving feedback to target group members, not least in the case of non-compliance, as well as skills training on how to deal with other stakeholders in the workplace, e.g. middle managers, so that they are not only able to work on the peer-to-peer level but also on the management level.

Management support from both middle and top managers also seems to be a significant context condition in order for the change agents to be successful. This is in line with institutional theory, which has been broadly accepted as an important framework to understand phenomena within the field of project management. The research contributes to both the field of project management and the field of change management. Formal change agents in the form of peer change ambassadors represent an innovative approach to helping middle managers and project managers succeed with organisational change projects, such as implementing health promotion activities at the workplace.

The findings increase the understanding of both researchers and practitioners on the challenges of undertaking projects related to implementation of behavioural changes as well as point to means to deal with them. The findings point to potentials of supplementing the classical project roles with a new role, the formal peer change agent, and thereby increase the likelihood that behavioural changes will be undertaken and sustained, whether this is related to specific job tasks or, e.g. as in this case, health-related behaviour.

This paper did not touch upon goal setting and follow-up activities on the department or company level, but this would be the subject for further research activities.

In addition, future research may include a similar study where relevant skills and competences for the peer HAs from this study are used in the selection process together with an even deeper investigation of middle and top management’s role in supporting peer change agents’ implementation efforts on health management at the workplace.