9. Balogun & Johnson (2005).pdf

01_Balogun_correxs 18/10/05 3:29 pm Page 1573 1573 Authors name From Intended Strategies to Unintended Outcomes: The Impact of Change Recipient Sensemaking...

01_Balogun_correxs 18/10/05 3:29 pm Page 1573 1573 Authors name From Intended Strategies to Unintended Outcomes: The Impact of Change Recipient Sensemaking Julia Balogun and Gerry Johnson Abstract Julia Balogun The tendency for intended strategies to lead to unintended consequences is well City University, documented. This longitudinal, real-time analysis of planned change implementation London, UK provides an explanation for this phenomenon. We focus on the social processes of Gerry Johnson interaction between middle managers as change recipients as they try to make sense University of of the change interventions. We show the extent to which lateral, informal processes Strathclyde, UK of inter-recipient sensemaking contribute to both intended and unintended change outcomes, and therefore the unpredictable, emergent nature of strategic change. The findings raise the issue of the extent to which it is possible to manage evolving recipient interpretations during change implementation. Keywords: strategic change, sensemaking, middle managers, unintended consequences Organizational change is a context-dependent, unpredictable, non-linear process, in which intended strategies often lead to unintended outcomes. Empirical research (Gouldner 1954; Harris and Ogbonna 2002; Johnson 1987; Mintzberg 1978; Ogbonna and Harris 1998; Pettigrew 1985; Pettigrew and Whipp 1991) shows that strategy development and change should be viewed as an emergent process. The rationalistic, objective tradition within the strategy field, which views strategic action primarily in top-down terms with an implication that practice flows naturally from policy, is questionable. Calls to understand better the role of micro organizational social processes and to acknowledge the impact of those outside the senior management team on strategy formation (Jarzabkowski 2004; Johnson et al. 2003; Balogun et al. 2003), go hand in hand with increasing evidence that middle managers play a significant role in strategizing (Burgelman 1983; Floyd and Wooldridge 1997; Huy 2002). Research that builds on these insights is growing, but there remains much to be understood about the strategic role of organizational actors outside Organization of top management teams. Nonetheless, in practice, many organizations Studies attempt to implement shifts in strategic direction through top-down change 26(11): 1573–1601 ISSN 0170–8406 initiatives. This paper therefore contributes to these developing insights by Copyright © 2005 examining how middle managers make sense of such top-down initiatives and SAGE Publications how their sensemaking affects resulting organizational changes. (London, Thousand Oaks, When organizations attempt to implement change through top-down CA & New Delhi) initiatives middle managers become key, as they are both recipients and www.egosnet.org/os DOI: 10.1177/0170840605054624 01_Balogun_correxs 18/10/05 3:29 pm Page 1574 1574 Organization Studies 26(11) deployers of the plans designed by their seniors (Fenton-O’Creevy 1998; Floyd and Wooldridge 1997). The acknowledgement of the agency of those outside the senior management team alerts us to the need to understand how these managers react and respond to such top-down change plans if we are to understand how change develops. Calls to recognize people as agents who construct their work environments (for example, Brown 1998; Tsoukas and Chia 2002) reinforces the focus of others on interpretations and meanings within organizations (Bartunek 1984; Giddens 1984; Ranson et al. 1980; Smircich and Stubbart 1985; Willmott 1987). We know from research on sensemaking and cognition (Balogun and Johnson 2004; Bartunek and Moch 1987; Brown 1995; Brown and Humphreys 2003; Poole et al. 1989; Isabella 1990; Barr 1998; Gioia and Chittipeddi 1991; Gioia et al. 1994; Labianca et al. 2000) that recipient interpretations of change plans, and how these interpretations are mediated by their existing context of action, ways of thinking, and interactions with others, are likely to be key. The content of new information is mediated for recipients by their existing knowledge, creating scope for both intended and unintended meanings (Smircich and Morgan 1982). Structure is both the medium and outcome of action (Giddens 1984), and all actors have the capability to do otherwise. Control is not one- way since it involves negotiation between different groups to produce particular outcomes (Coombs et al. 1992; Brown 1998). Through our focus on how middle managers make sense of top-down change plans we are able to explore acknowledged links between recipient cognition and the emergent and unpredictable nature of strategic change. We report on the findings from a real-time, longitudinal, qualitative study on how the implementation of an intended strategic change initiative progresses from the perspective of middle managers as change recipients. The findings show how intended and unintended change become inextricably linked as implementation progresses. Consistent with an interpretive approach on organizations, a sensemaking perspective (Brown 2000; Weick 1995) is used to develop a framework from the empirical data to account for these findings. Through this framework we are able to build on what we already know about cognition and change to show that the interpretations middle managers arrive at concerning how they should respond to the change interventions arise primarily from lateral, informal social processes of interaction between themselves. As they seek to put the change plans of their seniors into action, their everyday experiences of the actions and behaviours of others, and the stories, gossip, jokes, conversations and discussions they share with their peers about these experiences, shape their interpretations of what they should be doing. Change interventions and plans are translated into action through the medium of these inter-recipient processes, turning top-down intended change into an emergent and unpredictable process. In this paper, after reviewing the literature on change, cognition and sense- making, and describing our methodology, we present a first-order analysis (Van Maanen 1979), a thick description that describes the change process through the eyes of the research participants. We then present a second-order analysis to develop our model. Finally, we discuss implications for future research and practice. 01_Balogun_correxs 18/10/05 3:29 pm Page 1575 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1575 Change, Cognition and Sensemaking We position our consideration of top-down management of change within the increasingly well-articulated strategy process research and sensemaking literatures that bring to the fore the agency of diverse organizational groups, and the need to address both structure and action (Pettigrew 1985; Ranson et al. 1980; Bartunek 1984). While different authors use different theoretical perspectives to examine these issues, as we argue in the introduction, the acknowledgement of the need to understand the agency of those outside the senior management team suggests a need to study their interpretations of change. Therefore, we adopt a sensemaking perspective. Research on cognition and change more explicitly focuses on shared meanings as an aspect of structure. Such research is based on the premise that to achieve strategic change it is necessary for a change to occur in organi- zational interpretive schemes (Bartunek 1984; Ranson et al. 1980) — the fundamental shared assumptions that determine the way the members of an organization currently conceive of their organization and their environmental context and how they act in different situations. The research shows sense- making to play a central role in change (Isabella 1990; Gioia and Chittipeddi 1991; Gioia et al. 1994; Gioia and Thomas 1996), examines how shifts occur in individuals’ schemata during organizational transformation (Balogun and Johnson 2004; Bartunek and Moch 1987; Poole et al. 1989; Labianca et al. 2000) and how this leads to change in organizational interpretive schemes. Schemata are the mental models held by individuals that affect the events individuals respond to and how (Schutz 1964; Fiske and Taylor 1991; Labianca et al. 2000). They act as ‘templates against which members can match organizational experiences and thus determine what they mean’ (Poole et al. 1989: 272). While the notion of organizational or group schemata is contentious (Labianca et al. 2000; Klimoski and Mohamed 1994) since cognition is an individual phenomenon (Larson and Christensen 1993), during times of stability, when existing schemata and, for example, routines of interaction are not challenged, some level of shared understanding needs to exist for co- ordinated activity to occur (Barr and Huff 1997; Langfield-Smith 1992). The commonality between individuals’ schemata leads to an enacted reality (Weick 1979, 1995) at group level in the form of routines, rituals, systems, norms, assumptions and beliefs. Weick (1995) drawing on Wiley (1988) refers to this as the level of generic subjectivity — the level of constructed, taken-for-granted reality, which can be associated with ‘organization’. Weick argues that at times of stability the generic subjective takes on many forms, such as scripts, interlocking routines and habituated action patterns, and shared understanding. The ongoing re-enactment of the generic subjective reinforces individual schemata. When individuals face change, however, existing cognitions are likely to surface as individuals experience surprise. Their expectations differ from their experience (Louis 1980; Louis and Sutton 1991). They start to act in a more conscious sensemaking mode (Brown 2000; Johnson et al. 2000; Weick 1995) as generic subjectivity breaks down to make sense of what is going on around 01_Balogun_correxs 18/10/05 3:29 pm Page 1576 1576 Organization Studies 26(11) them. This more conscious sensemaking mode is a social process which Weick (1995), again drawing on Wiley (1988), refers to as the level of intersubjective sensemaking. It is where ‘Individual representations (thoughts, feelings, “intentions”) become merged or synthesized... not directly into supra-interactive structures, but into face-to-face conversations and inter- actions’ (Wiley 1988: 258). Sensemaking is primarily a conversational and narrative process (Brown 2000; Gephart 1993, 1997) involving a variety of communication genre (Watson and Bargiela-Chiappini 1998), both spoken and written, and formal and informal. However, more specifically, sensemaking involves ‘conver- sational and social practices’ (Gephart 1993: 1469). It occurs through both verbal and non-verbal means (Gioia and Chittipeddi 1991; Gioia et al. 1994). Individuals engage in gossip and negotiations, exchange stories, rumours and past experiences, seek information, and take note of physical representations, or non-verbal signs and signals, like behaviours and actions, to infer and give meaning (Isabella 1990; Gioia and Chittipeddi 1991; Gioia et al. 1994; Gioia and Thomas 1996; Poole et al. 1989; Labianca et al. 2000). Change comes about through shifts in conversations and language (Barrett et al. 1995; Brown and Humphreys 2003; Ford and Ford 1995; Heracleous and Barrett 2001). Although we know from studies that examine the way interpretations of change initiatives develop through time (Balogun and Johnson 2004; Brown and Humphreys 2003; Isabella 1990; Labianca et al. 2000) that these recipient processes of interaction are key during change, they remain an under-explored research avenue. Through this real-time study with its focus on middle managers, we therefore build on these earlier studies to explore the links between recipient cognition and the acknowledged emergent and unpre- dictable nature of change implementation. We seek to understand (1) the nature of the processes of interaction between recipients, (2) how they contribute to a new negotiated group understanding and (3) the impact of this negotiated understanding on the change outcomes. Research Study Methodology and Data Collection Like other research on processes of strategic change, this research takes a real-time, case-based, processual approach (Pettigrew 1992), using a qualita- tive embedded single case study design (Yin 1994), to capture contextual richness and complexity (Dyer et al. 1991). Since the research focus was on change recipients and how they make sense of change events, the research was conducted from the interpretive perspective of inquiry from the inside (Brown 1994, 1995; Isabella 1990). The case study site was the core business division of a recently privatized utility, Anonco, implementing planned strategic change (Romanelli and Tushman 1994; Pettigrew 1985) initiated by senior managers in response to forthcoming regulatory change. The formal divide of the old division into three new divisions with the appointment of three divisional directors marked the start of implementation in March 1993. Middle managers were selected 01_Balogun_correxs 18/10/05 3:29 pm Page 1577 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1577 as the recipient group for the research since they play a key role in making change happen (Floyd and Wooldridge 1997), yet their role, particularly in planned radical change, remains under-researched (Huy 2002). They can also, as at Anonco, be the target and agent of change (Fenton-O’Creevy 1998). The new strategy required the middle managers to undertake significant personal change. Furthermore, they had to implement the plans devised by their seniors, with little involvement in the up-front decision-making. Tracking of implementation progress from the perspective of the middle managers started from August 1993 once they had taken up their appointments in the new structure. The tracking continued until July 1994. Access was granted to the organization in return for feedback from the researcher to senior managers on the progress of change. Diaries, a form of personal document (Denzin 1989), were used as the primary data collection mechanism, since diaries provide an insider’s account of a situation (Burgess 1984). In all, 26 middle managers were recruited as diarists from a group of about 90 managers at this level. Potential diarists were identified from the new organization charts by the three divisional change managers appointed to support the divisional directors during the transition. At the request of the researcher, the change managers selected representatives (many not personally known to them) to cover most departments and interfaces, and also maintained diaries themselves. In order to track the managers’ interpretations of the change process, the diarists were given individual diaries, which contained separate entries for each time period with five questions — What is going well and why; What is going badly and why; What problems do you foresee; What have been the significant events; and What rumours are circulating? The diarists initially completed their diaries fortnightly, but moved to monthly reporting in November as the pace of change slowed. Frequent contact was maintained with the diarists by telephone. These phone interviews were used to follow up on queries raised by the diary entries and to add detail. Review meetings on a divisional basis were held with the change managers and diarists every 6–8 weeks. These meetings enabled the researcher to visit the site, to observe how the diarists interacted with each other and to hear the stories and gossip they exchanged. The researcher also maintained close contact with the three divisional change managers, who worked closely with their respective divisional directors, providing insight into the senior manager perspective. The researcher interviewed the change managers individually prior to the start of the research, and had individual meetings with them whenever she visited the site. One-to-one in-depth, relatively unstructured interviews with all diarists took place at the start of the research and towards the end. The interviews enabled the researcher to probe for additional information about the change process and to gain information on the different perspectives and responsibilities of the diarists. Each divisional director was interviewed once. There were also regular progress meetings with the Core division director. Additional data were collected from focus groups with the diarists as the research drew to a close. Documentation such as copies of briefings and workshop documents were also collected, providing additional data triangulation (Jick 1979). 01_Balogun_correxs 18/10/05 3:29 pm Page 1578 1578 Organization Studies 26(11) The Case Study The purpose of the radical restructuring at Anonco (a pseudonym) was the creation of an internal market for the provision of services to drive down costs while maintaining quality and improving service levels. Three new divisions were created from the old business division — a small Core division and two support divisions, Maintenance and Services. The Core division became the owner of the company’s asset base, responsible for the strategic development of the business. The two support divisions were to provide services to the core on a contractual basis, creating internal divisional customer supplier relationships. Maintenance was to focus on the maintenance and development of Anonco’s distribution network, and Services was to focus on end customer services. As the largest division, Maintenance was then split into three businesses. The Core division could be allowed to contract out services should the support divisions fail to offer competitive levels of servicing. The overall vision was for the three divisions to realize the potential of the business, maintaining high standards of customer supply and services and good industrial relations. The planned changes also involved redundancies, delayering, new working practices and a change from a technical, risk-averse culture to a customer-focused organization. A team of consultants and senior managers (the design team) prepared the rationale for, and outline of, the proposed new structure, and job roles and responsibilities. Implementation was based on the roll-out of the change plans developed by the design team. The middle managers had to work out the detail of their job roles and responsibilities themselves once they were in place. The implementation was monitored through the transition phase, which was initially to last about a year until the contracts came into operation in April 1994. Following the appointment of the directors, all staff were subse- quently reappointed to the new structure, or accepted exit terms. For each staff level there were appointment briefing meetings with, for managers, details of the different positions staff could apply for, followed by counselling on the options available. The planned interventions were primarily hard levers to do with new structures, systems, roles and responsibilities. ‘Business as usual’ kept the business running during the transition. Staff continued to do the work they used to do, while gradually taking up their new responsibilities. There was an extensive communications campaign to ensure that everyone understood the new business vision. All staff saw a video as part of a question- and-answer session at the start of implementation and subsequently attended ‘vision workshops’. Data Analysis: Intended and Unintended Change Outcomes An inductive data analysis and theory development approach was used, based on the ethnographic approach of first-order and second-order analysis (Gioia and Chittipeddi 1991; Van Maanen 1979). At the first level of analysis it was necessary to identify the change interventions perceived by the middle 01_Balogun_correxs 18/10/05 3:29 pm Page 1579 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1579 managers to have been put in place, any other events they perceived to have an impact, and the change outcomes they perceived to occur. At the second level of analysis, the research sought to understand how the interventions and events were interpreted and why, and how these interpretations led to the perceived outcomes. Early analysis focused on the construction of narrative (Langley 1999; Pettigrew 1990). A change story was written for each division from the participants’ perspective by pulling together the accounts from the different diarists and the documentary evidence, to create a thick description. In parallel, key themes and patterns were evolved from early and detailed coding (Miles and Huberman 1994), which utilized codes based on the language of the diarists (Van Maanen 1979; Strauss and Corbin 1990), such as ‘walls between businesses’, ‘who does what’ and ‘black holes’. Categories for change interventions, other events and change outcomes perceived by the middle managers were derived from their diaries, and their comments at meetings and interviews, as data collection progressed. This stage was equivalent to open coding (Strauss and Corbin 1990) in grounded theory. All data from interviews were transcribed, and entered into and coded in NUD.IST, software which aids the coding, searching and theorizing of non-numerical data, along with all diary entries and notes and observations from meetings. The detailed codings were gradually grouped into a set of broader categories to do with:  Designed change goals and interventions: actions taken to deliver the planned changes, such as the new structure with new roles and responsibilities.  Design flaws: aspects of the designed change process perceived by the recipients to hinder rather than help the change process, such as inadequate attention to transition management.  Congruent change consequences: outcomes consistent with the change goals desired by those leading the change process, and which facilitate intended change.  Counteracting change consequences: outcomes that differ from those intended by the leaders of the change process, and which block intended change progress.  Existing ways of thinking: the mental models individuals hold about the organization and the way individuals work together. Increasingly it became clear that the diarists perceived complex patterns of interaction between the different change interventions and events and the existing organizational context leading to both the counteracting and congruent change consequences. This then turned the planned implementation into a more unpredictable process with both intended and unintended emer- gent change outcomes. A ‘visual mapping strategy’ (Langley 1999) based on a type of causal network (Miles and Huberman 1994) was used to map the perceived interactions. A supporting narrative was developed for each map to explain the linkages perceived by the diarists. This formed the first-order analysis that follows. These maps and narratives then became the focus for the subsequent second-order analysis.  01_Balogun_correxs 1580 Figure 1. Development of Inter-divisional Relationships 18/10/05 TIME Organization Studies 26(11) 3:29 pm Contract training (12) Design team New Reinforced output (1) structure (2) Page 1580 negative Ongoing i n t e r p r e t a ti ons inter-divisional Protection of the new tensions (6b) Old working of turf structure (4b) patterns (5) (7) Negative interpretations Inter-divisional of the new tensions (6a) Prolonged structure (4a) business as Launch usual (11) communications (3) (Reducing) Arguments over black holes (8) Inter-divisional liaison & co-operation (14) No contract Little inter- details (9) divisional transition management (10) Contracts (15) Interpretation of need for co-operation (Core) (13) Designed change Design flaws Counteracting change Existing ways goals and consequences of thinking interventions Congruent change consequences 01_Balogun_correxs 18/10/05 3:29 pm Page 1581 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1581 First-Level Findings The analysis identified a wide range of change consequences. To show how they developed we need to link the organizational level of analysis with the individual level, by examining how the middle managers as change recipients were responding to the change initiatives and how this led to the emergent outcomes. Therefore, for illustrative purposes, in this paper we focus on a particular series of interlinked counteracting and congruent change conse- quences that occur as new working relationships between the divisions are put in place. This example not only illustrates how the interpretations which the change recipients form through their ongoing social interactions create both intended and unintended outcomes, but also the ebb and flow of change consequences as change develops. The following presents a causal network for this example (Figure 1) and the accompanying narrative description. The narrative is cross-referenced to Figure 1 through numbers to show the linkages the diarists perceived between the different components. The example is then re-examined in the second-order analysis to show the role played by inter- recipient social processes of interaction in creating the interpretations that underpin the different change consequences. The Development of a New Working Relationship between the Divisions The design team developed a designed change goal (1) of a new structure (2) with three divisions working together on a commercial basis within a framework of customer–supplier contracts. Separate divisional identities had to develop to enable the divisions to perform effectively what were essentially different types of work and responsibilities. Structural change was also to drive cultural change: ‘We think that by changing the structure, by changing the working relationship and the systems and locations and responsibilities, that the culture change occurs by that’ (Core Division change manager). The launch communications (3) talked of the intent for staff in the three divisions together to establish new co-operative working relationships. The middle manager vision workshops, led by their senior managers, were completed before the start of the tracking of change progress, and provided information on the change rationale, the working of the new structure and the different divisional goals and visions. There was also ongoing communication via team briefs and memos. However, even in the early diary entries, there was evidence of negative interpretations of the new structure (4a) and future work- ing relationships between the divisions, particularly for staff in Maintenance: ‘Maintenance staff have again expressed the feeling that the Core staff all feel as if they will be alright and “calling all the shots”, with no worries and the destiny of Maintenance in their hands.’ ‘Core Division is a small group of people, and the feeling out there is that everybody else is going to be floated off, they are not going to be tied to this company, and this company is my company, I worked 27 years here, and they don’t like losing it.’ ‘(Maintenance staff) now feel like contractors.’ Others in Maintenance felt the new structure created a perception of a core elite and the rest, leading to bad feelings: 01_Balogun_correxs 18/10/05 3:29 pm Page 1582 1582 Organization Studies 26(11) ‘I think there is still the impression that if you made it through to Core you are one of the elite. I get the feeling there is a lot of sour grapes, backbiting between Maintenance and Core on the basis that people that didn’t get there, think they should have.’ ‘It will be Maintenance that achieve the targets at the end of the year, but it won’t be us that get any accolade, it will be the Core people... all Maintenance people see is Core staff getting in the way if anything.’ The new structure challenged traditional working relationships and the old ways of thinking (5). Staff in all three divisions were used to working together co-operatively as colleagues of equal status (the new structure had ‘divided a united pack’), with equal ownership of the company assets that the staff jointly maintained: ‘There was a lot of work done on the “I owe you one principle” where you have done something extra for someone... and you go to the guy that you have done a favour for and basically call the favour back.’ (Core) ‘They have owned the records, and have taken a great deal of pride... and suddenly they are not their records... and some people find that really tough... we created them, we’ve worked on them, we’ve looked after them... (now) they are owned by Core.’ (Maintenance) Inter-divisional tensions (6a) were developing: ‘Strong feelings between divisions of competition. Behaviour of some individuals (managers) almost threatening.’ Services, however, had only recently been incorporated into the Core business division. Historically they had not worked as closely with the Core division staff, and indeed, in terms of customer service, saw them- selves as ahead of the other divisions. Thus, while experiencing the tensions, they did not develop the same interpretations about vulnerability as Maintenance. Tensions were fuelled across all three divisions by the many rumours and stories circulating about the working of the new structure, and disputes among senior managers: ‘I think even the people that I get on well with... they tell me things... and say you know... you’re going to be the contractors working for us.’ (Maintenance) ‘The Quality Assurance Team (Core division) are being referred to as the “Hit Squad” — purpose to knock Maintenance down.’ (Core) ‘Rumour: The three Directors are fighting among themselves.’ (Core) ‘Maintenance Director is not giving the same messages as the Core Director — very aggressive at Maintenance Briefing... “Core division think they will tie us down — we’ll be the ones doing the tying” — does little to improve inter-divisional relationships.’ (Core) As change moved into the autumn, appointments for all staff were finalized. The divisions started to focus their attention on developing their new working practices and systems. The Core division needed to establish new working practices for the commissioning and monitoring of work to Maintenance and Services. Maintenance needed to re-engineer their working practices to significantly reduce their cost base. Services similarly were looking at the introduction of flexible working through the use of new technology to enable manpower reductions. In response to the developing tensions between the divisions, the divisional directors went on a weekend team-building exercise in October, and following 01_Balogun_correxs 18/10/05 3:29 pm Page 1583 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1583 that initiated quarterly inter-divisional senior manager meetings. However, these events passed by without comment by the middle managers (probing at the final interviews revealed that they either didn’t know about these meetings or had seen no output from them). Meanwhile stories and rumours continued to circulate about inter-divisional tensions, with ongoing comments in all three divisions about the negative behaviours and attitudes middle managers were encountering from those in the other divisions. In particular, there were comments about protection of turf (7) between divisions: ‘Core division wanted to obtain some fairly trivial information from Maintenance. Maintenance would not release the information without a manager’s consent in case it could be used “against them” in future negotiations.’ (Core) ‘Everybody is trying to hold their budgets and hold their corner, and it is really getting a bit unhealthy.’ (Services) There was also a perception that Core division staff were not working as hard as the staff in the other divisions: ‘And they (Core) are starting to feedback their findings (from an investigation)... I would love for our people to go out and monitor our own staff never mind look at somebody else’s... There is a total imbalance of work.’ Other inter-divisional issues developed. There was confusion over ‘who does what’. Contracts had not been issued yet. The boundaries of inter- divisional responsibilities remained fuzzy. The diarists talked of ‘black holes’ (8) — responsibilities not clearly specified by the design team in the change documentation. There were disagreements over these issues between staff when trying to work across inter-divisional boundaries: ‘There are still some black holes that we seem to be arguing about, whose job is which. People are reluctant to take on the bits in the middle, the grey areas, and they are saying no that’s not mine that’s yours.’ (Maintenance) ‘There seems to be a clear indication of Services taking an uncompromising attitude to keeping things going. For example... “we’re not getting paid for it, so we don’t need to do it any longer”.’ (Core) The lack of detail on roles and responsibilities (1, 3) and the lack of contrac- tual information (9) perpetuated the black holes (8) and the disagreements and negotiations: ‘We need to see the contracts between the divisions a.s.a.p. so we know where we are going. Managers are continuing to have to work in the dark.’ (Maintenance) In addition, as change progressed through the autumn and into the next year, there was an extension of business as usual (11), partly due to the disagreements over black holes, but also driven by the fact that the Core division had not yet taken over all the duties they were ultimately to be responsible for: ‘People have been appointed into Core and numbers have significantly reduced in Maintenance without a reduction in workload because of the “business as usual” thing... (This is) seen to be imposing additional work onto our staff who have difficulty coping with their existing workload because of significantly reduced numbers.’ (Maintenance) 01_Balogun_correxs 18/10/05 3:29 pm Page 1584 1584 Organization Studies 26(11) ‘I think we’ve given up hope, and we’re accepting that if a job arrives for us, we’re doing it under “business as usual”.’ (Services) The slow resolution of many of these problems was blamed on a lack of adequate inter-divisional transition management (10) — problems had to be passed up to Director level: ‘The design team just finished and that was it... I think we needed someone there to actually see that what the design team said we should do actually happened. There is just no plan... we are just given “business as usual” and there was nobody there to resolve real interface issues.’ (Services) ‘I think where it went wrong was, because there are so many issues that you come across as you go along... people then tending to use “business as usual” as an excuse, you know if they are presented with a new problem that actually is their responsibility or will be, and they haven’t thought about that yet, it’s just “business as usual”.’ (Services) As the end of the first year of change approached, there was also specula- tion over the impact of the contracts on the inter-divisional relationships following the contract training (12). It was feared that the need for the Core division to enforce the contracts on Maintenance and Services might increase the tensions. For example, the researcher noted from one Maintenance feedback meeting, ‘Widely agreed that the introduction of the contracts would re-surface much inter-divisional friction. Diarists commented that it hadn’t gone away: it was there just below the surface.’ Earlier negative perceptions of the new structure (4b) and inter-divisional tensions (6b) were therefore reinforced. On the other hand, managers in the Core division saw a need to develop co-operative working relationships (13) across the divisions so that the divisions could ‘succeed together’, and were liaising (14) to achieve this: ‘On an individual basis we are liaising and working together as much as possible to reduce barriers.’ In November the Core division diarists organized a joint diarists’ meeting across the three divisions, facilitated by the change managers as a gesture of goodwill. There was a lot of debate at the meeting about the inter-divisional issues, but nothing was resolved. The Maintenance diarists in particular suggested that little could be done to resolve the feelings of vulnerability within their division. The Core division diarists persisted in their liaison efforts however: ‘They’ve got the expertise that we need, it is as simple as that... I had this yesterday with one of the engineers... and I said... “who’s got the expertise, who’s got the skills... you have”.’ These liaison efforts triggered some positive responses in Maintenance by the end of the first year of change, if not in Services: ‘Core division have set up a department that mirrors our setup which is helpful because we have got a contact person.’ At the end of the first year of change, the contracts (15) were gradually introduced by the Core division. The contracts specified contractual duties of the various Maintenance and Services departments, but were ‘shadow’ 01_Balogun_correxs 18/10/05 3:29 pm Page 1585 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1585 contracts only. For the first year of operation the contracts were not to be strictly enforced. They were to be used as a benchmark to identify black holes and other problems, so that these issues could be dealt with prior to the contracts becoming live in a year’s time. Once the contracts (15) were implemented, they were used by middle managers to liaise (14) more and resolve outstanding issues where possible: ‘I’ve got to comment, because I’ve been highly critical all the way through every meeting because the contracts have not been issued and that we are working to “business as usual”. Well, I’ve got to report that the contracts have been issued and that has got to be a step forward, and we are starting to work out what people’s definitions are, and working better with Core division.’ (Maintenance) As problems subsided, so did concerns about the impact of the contracts, and some of the antagonism. However, a degree of tension between divisions remained, particularly over ‘business as usual’: ‘People that won’t accept responsibility for saying yeah, yeah, that’s my problem, you’re right, leave it with me.’ Core division departments still had not taken on all their new duties. Furthermore, Maintenance and Services wanted the contracts to mark an end to ‘business as usual’. The Core division diarists disagreed: ‘People have forgotten that the contract and the “business as usual” are supposed to be running in parallel... they have these tablets of stone and they are going to abide by them come hell or high water.’ (Core) In addition there were concerns over the decline in some of the customer response time service measures that appeared to result from the issues discussed here, such as ‘who does what’, the lack of co-operation, and the workloads. Running beneath the tensions between the divisions was a recognition that the three divisions were working towards different goals, yet needed to co-evolve: ‘I think it is unfair to say all the inter-divisional attitudes are down to Core, because they are not. They are generated internal to Maintenance... We feel very vulnerable.’ (Maintenance) ‘I think the root cause is really that there is a lack of guidance as to where the three divisions are actually moving towards... the relationship in terms of talking to people is fine, but the understanding of where we are going a few months from now, I think I know and they think they know and I think it is a different opinion.’ (Maintenance) Second-Order Findings: Linking Change Recipient Sensemaking to Emergent Change Outcomes Contained within the first-order story are findings about how and why the many interlinked change consequences arose. A key factor is the social processes of interaction occurring between the middle managers in response to the different change initiatives and events as they try to work with their peers to establish the new structure, and the interpretations they develop as a result of this interaction. To identify the significance of these processes and interpretations we moved to second-order analysis to establish an explanation for the first-level findings. We examined the maps and supporting descriptions 01_Balogun_correxs 18/10/05 3:29 pm Page 1586 1586 Organization Studies 26(11) Figure 2. First- and Second-Order Themes Sensemaking triggers Designed change goals Designed change interventions New structure, roles, relationships Launch communications New working practices & systems Ongoing communications Contractual working & contracts Design team output Customer-focused culture Old schemata Relocation offices/depots Business as usual Team building Cost reduction Existing ways of thinking Change plans Job for life Total quality training We work for one company as equals Working parties / process redesign We own assets we maintain Re-appointments (briefing & counselling) Work done on basis of goodwill / I owe you one Redundancies Control / blame management Technical excellence Old schemata Sensemaking inform… triggers lead Design flaws to… Lack of inter-divisional transition management Business as usual getting in the way No detailed design team output No contract details early on Level/pace redundancies Social processes of interaction Inadequate communication Stories & rumours/gossip Sharing of experiences Sharing of interpretations Discussions Negotiations Non-verbal signs and signals Social processes of interaction result in developing schemata leading to emergent change outcomes Developing schemata Emergent change outcomes We need to make changes work (C) Positive staff attitudes (C) We have done well (C) Congruent change We need to develop our own job roles (C) Staff adaptation (C) consequences We need to work together (C) Inter-divisional liaison (C) Inter-divisional relationships improving We can work together through the contracts Inter-divisional co-operation Core division different to other divisions (C) More open management Less blame / control in Core division(C) style (C) Them and us Inter-divisional tensions We are contractors/Core the elite (M) Protection of turf Business as usual an excuse Disagreements over black holes New structure = competition Contracts will make things worse Counteracting We don’t know who does/pays for what Inter-business barriers (M) & change Them and us/unfair business differences (M) protection of turf (M) consequences We are not empowered (M & S) Prolonged business as usual Blamed for making mistakes (M & S) Lack of empowerment (M & S) Review group will design detail of structure (C) False expectations of output (C) Key C = Core Headings in italics are first-order themes M = Maintenance Headings in bold are second-order themes S = Services 01_Balogun_correxs 18/10/05 3:29 pm Page 1587 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1587 of how the counteracting and congruent change consequences developed: the first-level findings became the data for the second-order analysis, and a set of ‘second-order thematic labels’ were inductively derived from these descriptions (Gioia et al. 1994). Five additional thematic concepts were identified — social processes of interaction, developing schemata, emergent change outcomes, old schemata, and sensemaking triggers. Figure 2 shows how the first-level themes were amalgamated and extended in the second level of analysis to build a deeper explanation for the emergent change patterns identified in the first-order analysis (Gioia et al. 1994; Labianca et al. 2000). Figure 2 also contains details of all the change consequences identified from our analysis, and not just those discussed here, to show the full extent of the first- and second-order coding. Social processes of interaction are the conversational and social practices identified from the first-order analysis that the middle managers engaged in as they attempted to make sense of the new structure and their new roles and responsibilities. Old schemata are the existing ways of thinking which individuals hold about their organization, and provide the context within which change is initially made sense of. Sensemaking triggers are the events and happenings identified as triggering intersubjective sensemaking during change, and include the various designed change goals and interventions, the encountered behaviour of other organizational actors, and the design flaws. Developing schemata are the interpretations that change recipients arrive at, through their social processes of interaction, of what change is about, such as the negative interpretations of the new structure. In the early tentative stages of change they can be equated with what Isabella (1990) refers to as ‘in progress frames of reference’. These developing schemata underpin the emergent change outcomes that become visible through actions and behav- iours in the form of counteracting and congruent change consequences. The interpretations arrived at by individuals through the social processes are, therefore, important. If the meanings and interpretations individuals develop are consistent with those intended by the instigators of the changes, they are likely to behave in a way consistent with intended outcomes, leading to con- gruent change consequences. If the meanings and interpretations developed by the recipients differ from those intended, the result may be counteracting change consequences. Figure 3 shows how these different concepts link together to create an ongoing cyclical sensemaking process that turns change implementation into an emergent and unpredictable process, as change recipients develop particular interpretations about the imposed changes through their social processes of interaction. These interpretations then lead to both intended and unintended change outcomes. To illustrate the role of the different social processes of interaction the middle managers engaged in during change, we draw on the concept of the intersubjective level of analysis (Weick 1995; Wiley 1988). The intersubjective level looks at the processes involved in the interplay between the individual and generic subjective levels during times of change. Whenever change recipients encounter sensemaking triggers they cannot account for in terms of their existing mental models or schemata, they 01_Balogun_correxs 18/10/05 3:29 pm Page 1588 1588 Organization Studies 26(11) Figure 3. The Role of the Intersubjective level Intersubjective Level during Social processes of Change interaction In d ivid ual interp re tatio ns S e n s e m a k i n g tr i g g e r s : Developing schemata (from current schemata) D es ign ed cha n ge goals & lea ding to … i nt e r ve n t i o n s , e v e r y d a y ev ents, desig n f la ws Emergent change outcomes: C o u n te ra c t i n g a n d c o n g r u e n t c h a n ge c o n s e q u e n c e s Individual level G e n e r i c s u bj e c t i v e level Imposed Schemata changes Routines, rituals, symbols, s t or i es, s y st e m s, n o r ms , b e li ef s engage in more conscious social processes of interaction to attempt, within the context of their current schemata, to resolve their ambiguity and uncer- tainty (Brown 2000; Weick 1995; Louis and Sutton 1991). As a result of these processes old schemata and interpretations are amended and/or new schemata develop, leading to emergent change outcomes in the form of counteracting and congruent change consequences. The individual schemata or interpre- tations drawn on at any point in time constitute a mix between old schemata that have not been challenged, schemata in the process of transition, and schemata that have already been changed. Only as change subsides can individuals return to a less conscious sensemaking mode. Through this cyclical sensemaking process, earlier schemata and outcomes become the ground for subsequent sensemaking through the mediation of social processes of interaction, in a pattern similar to the structuration process of institutionalization proposed by Barley and Tolbert (1997). New sense- making triggers are constantly encountered as recipients interact with each other and attempt to arrive at some shared new meaning to enable them to once more operate together in a more taken-for-granted manner. These sensemaking triggers result from and are then interpreted within the context of (diverse) existing interpretations, so that the schemata become the context and outcome of action. Change consequences develop through time as a result of this sensemaking process, and can appear and disappear as alternative interpretations are arrived at. This is consistent with the macro-level change patterns described by the strategy process research. Explaining Emergent Change Outcomes Figures 4, 5 and 6 rework the example of the development of the new working relationships in terms of the model in Figure 1 to show how the second-order 01_Balogun_correxs 18/10/05 3:29 pm Page 1589 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1589 concepts enabled an explanation to be derived from the data for both the intended and unintended change outcomes. While clock time may provide a socially constructed tool for co-ordination and social interaction, it can obscure other temporal patterns within processes of change (for example, Whipp 1988). Adam (1990), building on Mead (1959), places the main emphasis on the emergent event in time, since the past and future are constantly created and recreated in present events. The past is irrevocable, yet it changes as we assign it different meanings and interpretations based on our experiences in the present. In line with this, we have split our analysis into three time frames based on sensemaking cycles, within which meanings, interpretations and events remain relatively constant, rather than on elapsed or clock time. This shows there are cycles of sense- making, in which the change consequences (the change outcomes) that develop in one period of sensemaking, then become part of the inputs to the next time period. Earlier schemata and outcomes become the ground for subsequent ones (Barley and Tolbert 1997). Furthermore, conceiving of a change process in terms of cycles of sensemaking, rather than elapsed time, better reveals the uneven pace of interpretation development, and the unsteady pace of change. Figures 4, 5 and 6 also split the social processes of interaction into vertical processes of interaction occurring between senior and middle managers, and lateral processes of interaction occurring between the middle managers themselves. These social processes of interaction are further broken down into verbal, behavioural and other non-verbal signals to capture the diversity of the conversational and social practices occurring. This analysis reveals that, as change progresses, most of the reported interactions that contribute to the emergent change outcomes occur informally between middle managers as they go about their everyday work. In time T1, sensemaking triggers for the change recipients included the new structure, the designed change interventions, such as the appointments, and various launch communication events/information (see time T1, Figure 4). These triggers led to a variety of lateral social processes of interactions. Many comments and stories were exchanged between middle managers, such as those about the Core division as an elite, and Maintenance staff becoming vulnerable contractors. These comments often indicated strong feelings of resentment, particularly between the Core division and Maintenance. Members of a group can also interact in ways other than the spoken and written word (Gioia and Chittipeddi 1991). The new imposed structure was a signal in itself, to which ‘meaning was attributed beyond its substantive impact’ (Brown 1994) and in combination with middle manager discussion produced a developing schema of ‘them and us’. In addition, there was the behaviour of the senior managers and the rumours and stories about this, and the stories about the negative attitudes of peers. Interpretation of events in T1 was affected by individuals’ existing schemata, their ‘context specific dictionaries’ (Berger and Luckman 1966: 138) to do with one company, equal status and ownership, and co-operative working; whereas contracts meant only doing what you got paid for, leading 01_Balogun_correxs 18/10/05 3:29 pm Page 1590 1590 Organization Studies 26(11) Figure 4. Triggers Development of Designed change goals / New structure, appointments to new structure, Inter-divisional briefings meetings on appointments Relationships, Time Interventions Launch communications (roadshow & video, T1: April– vision workshops, special team brief) September 1993 Change documentation Day-to-day events Ongoing team briefings, meetings, phone calls, visits, chance encounters, letters/memos Design flaws Vertical social processes interaction Verbal Content of team briefs, change documentation Content of workshops and appointment briefing presentations, video, question and answer sessions at events Behavioural Senior manager behaviour Non-verbal signals Shape of new structure and appointments Lateral social processes interaction Verbal Discussions implications new structure, stories & rumours senior and peer behaviours and attitudes Behavioural Early indications of lack of co-operation Non-verbal signals Staff moving jobs and locations Existing schemata We work for one company, equality of colleagues, co-operation by ‘I owe you one’, joint ownership of assets Developing schemata Them and us, we are vulnerable/contractors/no longer own assets (Maintenance), new structure = competition Emergent change outcomes Counteracting Inter-divisional tensions consequences Congruent consequences to additional interpretations of the ‘new structure = competition’. Everyone had heard that the three new divisions needed to work together to make the new structure work, but shared interpretations had developed as staff swapped stories and experiences that the new structure was more to do with ‘building walls’ between the divisions. By the end of T1, therefore, a new set of schemata to do with ‘them and us’, ‘new structure as competition’, and in Maintenance ‘we are vulnerable contractors’, had developed primarily through the middle manager lateral social processes of interaction as they sought to make sense of the various change interventions within the context of their old schemata to do with equality and co-operation, leading to outcomes of inter-divisional tensions and protection of turf. The stream of events had led to interpretations of the past ways of working as irrelevant, within a projected future of greater commercial awareness — and with it the need for each division to protect their own interests. By October 1993 the flurry of formal interventions relating to the new structure had come to an end (see time T2, Figure 5). Change more or less stood still, with few interventions occurring between September 1993 and 01_Balogun_correxs 18/10/05 3:29 pm Page 1591 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1591 Figure 5. Triggers Development of Designed change goals / Contract training Inter-divisional interventions Director weekend team building event, quarterly Relationships, Time senior manager meetings, inter-divisional diarists’ T2: October meeting 1993–March 1994 Day-to-day events Ongoing team briefings, meetings, phone calls, visits, chance encounters, letters/memos Design flaws No detailed design of responsibilities/roles, no contract details, lack of transition management (no resolution mechanisms) Vertical social processes interaction Verbal Content of team briefs Behavioural Senior manager behaviour Non-verbal signals Shape of new structure Lateral social processes interaction Verbal Ongoing discussions implications new structure, stories, rumours, discussions about unco-operative inter-divisional behaviour but some stories re liaison behaviour Questions, discussion, disagreements, negotiations over who does what (black holes) and business as usual Stories and discussions about unresolved problems, lack of information and resolution mechanisms Discussion of impact of contracts Behavioural Negative attitudes, protection of turf, lack of co- operation... but Core division attempts at liaison Non-verbal signals Core division seen to be working less hard Existing schemata Them and us, we are vulnerable/contractors/ (from T1) no longer own assets (Maintenance), new structure = competition Developing schemata We don’t know who does what or who pays for what, business as usual an excuse, contracts will make things worse, we need to work together (Core division) Emergent change outcomes Counteracting Continuing inter-divisional tensions consequences Protection of turf Disagreements, black holes and prolonged business as usual Congruent consequences Core staff liaison April 1994 to alter the nature of the working relationships developing between the three new divisions. Middle managers were trying to make the new struc- ture work, through working out the boundaries of their new responsibilities and their interfaces with other managers and divisions. The sensemaking triggers occurring were based on the developing schemata from time T1 to do with ‘them and us’, ‘competition’ and ‘vulnerability’. As the middle managers tried to go about their everyday work, they encountered negative attitudes in others, and protection of turf. This unco-operative behaviour led to more lateral social processes between colleagues in the form of rumours 01_Balogun_correxs 18/10/05 3:29 pm Page 1592 1592 Organization Studies 26(11) and shared experiences about the non-verbal signals from this behaviour, and also discussions and stories about the tensions and defence of turf. The middle manager interactions reinforced earlier interpretations of ‘new structure = competition’ and ‘them and us’, and the emergent change outcome of inter- divisional tensions with, now, protection of turf. The middle managers also engaged in many social processes of interaction triggered by the absence of sensemaking triggers. For example, when staff tried to establish the extent of their responsibilities in the new structure, they found they did not have sufficient information in the change documentation. There were black holes. Furthermore, there were no resolution mechanisms unless problems were passed up to director level. In addition, the contract details were not available yet. In situations of rapid change, with high levels of uncertainty and equivocality (Daft and Weick 1984), individuals seek to clarify their situation to enable them to return to non-conscious responding (Weick 1995). There were, therefore, discussions, disagreements and negotia- tions which led to a developing schema that ‘we don’t know who does what or who pays for what’, and ultimately, given existing schemata to do with ‘them and us’, and ‘new structure = competition’, additional counteracting change consequences with disagreements over black holes and prolonged ‘business as usual’. The design flaws of no detailed design/contracts and no resolution mechanisms illustrate how emergence leads to re-creation and interpretation of the past. Diarists openly criticized the level of design detail, that contracts were not planned for completion until March 1994, and that there were no resolution mechanisms. Their experiences in the present were leading them to attach different meanings to past events. The growing resentment over ‘business as usual’ in Maintenance and Services was also affected by the different temporal patterning (Hassard 1996) of change actions within and between divisions. There were multiple and overlapping cycles of change. Although there was little change in the development of inter-divisional relationships during time T2, change was occurring within the individual divisions as each developed their independent new working practices and systems. However, as these changes were not enabling the Core division to take on board their new duties as fast as redundancies were progressing in Maintenance and Services, they were perceived to be changing more slowly. The inter-divisional conflict was heightened by the need to synchronize, temporally, the different cycles of change in the different divisions to achieve the new inter-divisional working relationships. Meanwhile, a growing concern was developing among Core division middle managers in particular about the inter-divisional tensions. They were of the opinion that ‘we need to work together’, leading to further lateral social processes of interaction as Core division managers engaged in liaison behaviours with the other divisions. In addition, diarists in the other divisions shared their experiences of efforts by Core division managers to improve the inter-divisional relationships, through stories about the initiatives taken. However, these stories were still infrequent in comparison to stories about problems between divisions. 01_Balogun_correxs 18/10/05 3:29 pm Page 1593 Balogun & Johnson: Intended Strategies to Unintended Outcomes 1593 Figure 6. Triggers Development of Designed change goals / Contracts issued, contract management Inter-divisional interventions procedures Relationships, Day-to-day events Ongoing team briefings, meetings, phone calls, Time T3: visits, chance encounters, letters/memos April–July 1994 Design flaws No detailed design of responsibilities/roles, lack of inter-divisional transition management (no resolution mechanisms) Vertical social processes interaction Verbal Content of team briefs (declining performance indicators), contract details Behavioural (Nothing reported*) Non-verbal signals Shape of new structure, physical contract documents Lateral social processes interaction Verbal Reducing stories about unco-operative inter- divisional behaviour. Increasing stories about liaison/contractual co-operation Discussion and negotiations to resolve who does what (black holes) with contract information and to clarify contract problems Stories and discussions/disagreements about ongoing business as usual Discussions about declining performance Behavioural Decreasing lack of co-operation, but still defence of turf. Co-operation through the contracts Non-verbal signals Non-appearance anticipated contractual problems, positive effects of contractual liaison (problem resolution, fewer black holes) Existing schemata Them and us, we are vulnerable/contractors/no (from T2) longer own assets (Maintenance), new structure = competition, we don’t know who does what or who pays for what, business as usual an excuse, contracts will make things worse, we need to work together (Core) Developing schemata We can work together through the contracts, inter-divisional rel