Constructing Spaces for Strategic Work: A Multimodal Perspective

Introduction

The recent emphasis on materiality in practice research generally (e.g. Leonardi and Barley, 2010), and strategy-as-practice research specifically, has emphasized that people do strategy with ‘stuff’ (Jarzabkowski, Spee and Smets, 2013). That is, strategists cannot be separated from the artefacts such as spreadsheets, flipcharts and computer screens (Vaara and Whittington, 2012), bodily performances or spatial arrangements through which they do strategic work (LeBaron and Whittington, 2011). Yet many studies in strategy-as-practice have focused on discursive practices (e.g. Balogun, Jarzabkowski and Vaara, 2011; Rouleau and Balogun, 2011; Samra-Fredericks, 2003), examining talk-based interactions and underplaying the bodily, material and spatial aspects of strategic work. Hence, while burgeoning research points to the critical role these additional resources play (e.g. Hodgkinson and Wright, 2002; Kaplan, 2011; Liu and Maitlis, 2014), we still have only partial understanding of the ways in which they are orchestrated or implicated in accomplishing strategic work. To address this blind spot, Vaara and Whittington (2012) have called for strategy-as-practice research to ‘go beyond discourse to consider how the material, in the form of both bodies and artefacts, is used to accomplish strategy work’ (2012, p. 316). Such calls encourage researchers to adopt a multimodal lens, to examine the interplay of bodily orientations, gestures, gazes, material artefacts and talk (Streeck, Goodwin and LeBaron, 2011), in order to enrich our understanding of how actors perform strategic work.

We respond to this call by exploring the orchestration of bodily, material and discursive resources in accomplishing strategic episodes. Our paper first highlights gaps in understanding about the way these resources come together in performing strategic work. We then develop an inductive analysis building on a video-ethnographic study of strategic episodes for making capital allocation decisions on reinsurance deals, which examines the everyday interactions through which reinsurers enact their firm's strategic portfolio within the particular spatial–material arrangements of a reinsurance trading desk. Our findings show that actors construct three distinct spaces through different orchestrations of bodily, material and discursive resources. We show that these spaces are consequential for the types of strategic work performed and for accomplishing the particular outcomes of the strategic episode. Drawing on these findings, we generate a conceptual framework for understanding strategic work as a spatial accomplishment and describe the specific contributions that our framework makes to the strategy-as-practice field.

Theoretical framework

This study builds on the foundations of strategy-as-practice research (Jarzabkowski, Balogun and Seidl, 2007; Johnson, Melin and Whittington, 2003; Johnson et al., 2007; Whittington, 2003, 2006) by focusing on strategizing as the ‘actions, interactions and negotiations of multiple actors and the situated practices that they draw upon’ (Jarzabkowski, Balogun and Seidl, 2007, pp. 7–8). While this research agenda exhorts us to focus on actual strategic work – what managers do when they enact strategy – (e.g. Jarzabkowski, 2005; Jarzabkowski and Whittington, 2008; Whittington, 2003, 2006) there are still only partial understandings of what this work comprises. In particular, existing research has focused on the discursive work of doing strategy (Balogun et al., 2014). This has provided valuable insights into the discursive practices through which managers at different levels of the firm participate in (Mantere and Vaara, 2008), make sense of (Balogun and Johnson, 2004; Rouleau, 2005) and enact their strategic roles (Mantere, 2008; Rouleau and Balogun, 2011). Such discursive work is undoubtedly consequential in developing, sharing and implementing strategy, including negotiating through the inevitable ambiguities and contradictions that are integral to strategizing (e.g. Abdallah and Langley, 2014; Dameron and Torset, 2014; Kwon, Clarke and Wodak, 2014; Sillince, Jarzabkowski and Shaw, 2012). Discursive practices are, for example, central to the way that middle managers make sense of and enact organizational restructuring during strategic change (e.g. Balogun and Johnson, 2004; Rouleau and Balogun, 2011). Discourses also inform what is considered strategic (Hardy and Thomas, 2014), fuel resistance (Balogun, Jarzabkowski and Vaara, 2011), shape the objectives included in a firm's strategy (Kaplan, 2011; Spee and Jarzabkowski, 2011), and enable and constrain the development of strategic orientations (Jarzabkowski and Seidl, 2008; Vaara, Sorsa and Palli, 2010). In a particularly illuminating study, Samra-Fredericks (2003) illustrated how a strategist shaped a firm's strategic investments through skilfully employing discursive resources during a management meeting to draw attention to two organizational weaknesses. Discursive resources are thus an important part of strategizing.

However, as recent calls to extend our focus to the sociomaterial elements of strategy indicate, a focus on discourse alone provides only partial understandings of how strategic work takes place (e.g. Balogun et al., 2014; Vaara and Whittington, 2012). In this vein, some recent contributions have begun to shed light on how PowerPoint, flipcharts and documents are constructed and mobilized during interactions (e.g. Spee and Jarzabkowski, 2011; Whittington et al., 2006). As they become part of strategic activities, these materials acquire meaning, even as they also afford particular types of strategic interactions. For example, Kaplan (2011) illustrates how the use of PowerPoint mediated the discursive practices of strategy making by enabling actors to collect, represent, share and edit ideas. Yet the spatial and bodily activities for accomplishing strategic work have received only limited attention within strategy-as-practice research (e.g. LeBaron and Whittington, 2011; Liu and Maitlis, 2014). Liu and Maitlis (2014), for instance, illustrate the importance of the physical body in their investigation of how emotional displays – using facial, physical and verbal cues – shape the strategizing process. Hodgkinson and Wright (2002) also illustrate the strategic consequentiality of bodily, spatial and material work with their reflective account of a failed strategy intervention that shows how powerful actors use their physical bodies (e.g. sitting remotely, pacing up and down) and material objects (e.g. marker pens and white boards) to control meetings. Despite these contributions, few studies have considered the role of material or bodily performances in strategic work. Thus, a blind spot remains around how bodily and material resources are employed, either singly or in concert, to accomplish strategic work.

In part, this gap may be attributed to the problem of identifying what and whose work is strategic. Strategy-as-practice has long called for research that extends beyond top managers (Johnson, Melin and Whittington, 2003; Rouleau and Balogun, 2011) or those specifically labelled strategists (Jarzabkowski and Spee, 2009; Whittington et al., 2003). Indeed, Floyd and Lane (2000) emphasize that strategizing occurs at multiple levels of the firm, while Mantere and Vaara (2008) examine how managers participate in strategy at different levels. Yet the ‘strategic’ nature of these actors' work is often not apparent until after the fact. For example, Salvato's (2003) study showed how the combination and recombination of micro-strategies shaped critical firm evolution in two firms, Alessi and Modafil. Similarly, Regnér (2003) shows how exploratory product experimentation by peripheral managers enabled Ericsson's strategic shift into the mobile telephony market. Other studies illustrate the importance of sales managers' micro-actions in translating and enacting firm strategy (e.g. Rouleau, 2005; Rouleau and Balogun, 2011), or orchestra members' political discourses in strategic failure (Maitlis and Lawrence, 2003). And yet, a priori, few would have identified these operational managers, middle managers, sales managers and orchestra members as strategic actors, or their talk, their tinkering and their everyday, often routinized, business actions as strategic work. Yet, it is through the everyday enactment of this work that strategies of firms are brought into being and accomplished. As Chia and Holt (2006, p. 637) note, ‘Strategy is not some transcendent property that a priori unifies independently conceived actions and decisions, but is something immanent – it unfolds through everyday practical coping actions’. The practical coping that Chia and Holt (2009) advocate encourages us to move beyond overtly strategic work as it is displayed by purposive actors and give more serious attention to how actors in situ engage with the spatial and material arrangements to hand (Heidegger, 1962; Schatzki, 2005).

In order to address calls to go beyond studies of the ‘overtly strategic’, such as top managers and strategic plans, we must examine the strategic episodes of managers outside the top team whose everyday actions enact a firm's strategy and are critical to a firm's ‘wealth-generating capability’ and ‘long term survival’ (Hendry, 2000, p. 970), wherever they take place in a firm. In this paper we focus on a particular group of professional actors who have high autonomy for making decisions that enact core aspects of their firms' strategies (Løwendahl and Revang, 1998). Specifically, we examine the everyday practice through which a group of reinsurance managers enact the capital allocation decisions that are necessary to fulfil their firm's strategy portfolio. The strategic consequences of these everyday decisions are critical – failure to fulfil the portfolio with appropriate deals inhibits the firm's wealth generating capability, while over-exposing the portfolio to risk can cause the firm to collapse – and yet such high-stake decisions are also part of these managers' routinized practice. We undertake an inductive ethnographic analysis of strategic episodes of these managers at their trading desks, in order to address the research question: how is strategic work accomplished through the orchestration of material, bodily and discursive resources?

Methods

Findings

Our findings are presented in three sections. First, we introduce two representative examples of strategic episodes to illustrate how speech, bodily and material resources were simultaneously employed during reinsurer–broker interactions. Second, based on our analysis of all ten episodes, we show how actors used constellations of semiotic resources to construct three distinct spaces. Finally, we explain how the specific activities in each of these spaces constituted three distinct kinds of strategic work that enabled the accomplishment of strategic outcomes.

Strategic episodes and patterns of semiotic performance

In strategic episodes we watched reinsurers work with brokers to generate business opportunities, arrive at decisions about deals and maintain strategic business relationships. Two examples of these strategic episodes are presented in Table 1. In episode 1, Nigel (reinsurer) and James (broker) constructed a novel solution which enabled Nigel's firm to place capital and generate a strategic opportunity for their sister company in Europe. In episode 2, Mike (reinsurer) considered some US business presented by Ben (broker) but decided not to trade due to strategic concerns about profitability and portfolio management. Such strategically consequential outcomes generated on the trading floor were not predetermined but were ‘accomplished’ through the mutual work of the actors involved.

Table 1. Two typical strategic episodes

Episode 1	Constructing a solution
19:28 minutes	Nigel (the reinsurer) eventually turns his head towards James (the broker), which opens the interaction, and they engage in some small talk before turning to the business at hand. James is returning with the pricing of a deal that was slightly lower than Nigel had expected. At the beginning of the episode, James and Nigel discuss the details of the deal, particularly some new insured dwellings that it contains. This development is a concern for Nigel. He is worried that this will give his firm exposure to additional risk he had not planned for when evaluating the deal initially. Thirteen minutes into the exchange, however, Nigel puts forward a potential solution which would enable his firm (ReinCo) to place capital on the deal, while also creating a new commercial opportunity for their sister company in Europe (EuroRe). This opportunity emerges as Nigel queries: ‘Do Europe write this as well?’ After some discussion about EuroRe's involvement, Nigel proposes that they could offer the same amount of capital, but that EuroRe could be introduced to take a split so that he places around half of the desired amount of capital while EuroRe offers the remaining capital to match the overall capital allocation that James would like. He explains his thinking to James: ‘From a group perspective we can handle around 50%. But it might be an opportunity for Europe. … I think they have tried to write this in the past but there hasn't been an opportunity for them. But if we take our line down it might give them an opportunity.’ Initially, James is not enthusiastic about Nigel reducing his line, particularly as he was hoping to wrap up the deal. However, they discuss how this might work and James eventually concedes there could be an opportunity for EuroRe to place an equivalent amount of capital on the deal. They agree Nigel will pursue this internally and that they will meet later in the week.
Episode 2	Deciding not to trade
10:03 minutes	The interaction begins when Ben (the broker) comes to the box with a major US wind programme he is trying to place. Mike (the reinsurer) states early in the interaction, ‘I have already placed a lot of capital on US wind … our book is virtually full’. However, he also leaves open the possibility of writing additional business if the right opportunity comes along: ‘I mean if the prices really did kick-on then I suppose there is an argument for opening up the book’. The deal itself covers four coastal zones in the USA and Ben is looking at a 15–20 million limit across the regions, but Mike tells Ben, ‘That's far too low for me’, meaning that the risk is too great for their portfolio. He explains: ‘I mean we are in pure defensive mode in terms of what we are writing now. You know the only deals we will write over the next 6 weeks will be middle and top end stuff – away from losses and paying a good price. We're not doing anything below 25 at this stage and most of our business is 40 million upwards.’ Even so, Ben encourages Mike to speculate on what kind of price he would be looking at for a 20 million limit. Mike gives him a rough price, but immediately qualifies this by saying, ‘But it's far too low for me … we have taken the book up and away’. There is some discussion about the pricing and Ben tries to persuade Mike to reconsider, which he thinks about, including doing some brief calculations, before restating his position, ‘We are full’, at which Ben thanks Mike ‘for your consideration’ and leaves. At a debriefing after this interaction, Mike explains his decision to the researcher: ‘I've filled my portfolio on the USA. We have a plan for capital allocation and, of course, I can go over it if there is a good reason – pricing is high – or it gets me something else I want. So I might have, you know … but this deal, there's nothing wrong with it, but it just wouldn't pay enough extra when we've met our targets for the USA already.’

These episodes can be partially understood by focusing on what was said during the interactions. However, closer inspection reveals that multiple semiotic resources were being employed throughout and were ‘crucial to how participants … built action together’ (Goodwin, 2003, p. 20).

Speech resources

Talk was used extensively as actors negotiated terms, exposures and capital allocations and engaged in dialogue to explore commercial positions. However, periods of silence were also a common and important feature of these interactions. In episode 1, for instance, we observed over 20 distinct ‘moments of silence’, equating to 2:20 minutes. Similarly, in episode 2 there were 13 distinct ‘moments of silence’ constituting 1:07 minutes. In other words, a relatively large proportion (over 10%) of these interactions was conducted without words being spoken. First, there were relatively long periods of silence when reinsurers suddenly stopped talking and worked on their computers, switching between spreadsheets, doing calculations and jotting down notes. Nigel, Mike and other reinsurers appeared to take what we referred to as ‘time-outs’ to read and consider information, contrast it with their own commercial analyses and think through options. Additionally, shorter silences frequently occurred as the actors explored options and negotiated decisions. For example, in episode 1 a number of short silences were observed after James told Nigel the new terms.

Nigel just looks at James and does not speak. After a few seconds he says, ‘It's too low’. There is then another 6-second silence before James starts to argue that there are some incorrect assumptions in the original exposure figures.

These kinds of short silences were common in all the episodes. In episode 2, for instance, they happened when Mike stated ‘No, we are full’ in response to Ben's petitions. This categorical statement triggered a prolonged silence in which Mike and Ben just looked at each other, seemingly each waiting for the other to fill the discursive void. Later, Ben used silences as if trying to invite Mike to elaborate or change his position.

At the end of the interaction Ben picks up his folders as if to leave and gives Mike a long final look without speaking. With no response from Mike, Ben eventually says, ‘Okay then … [3-second silence] … thank you’ and leaves.

Material resources

The use of material resources was also central to the work being accomplished. The compact nature of the space, and the density of materials located within it, provided both actors with an array of opportunities to interact with materials, such as highlighting pens, notepads, reports, information packs, information sheets, the desk, calculator, mouse, keyboard, the computer screens. The use of materials was partly determined by differential ownership and layout. For example, reinsurers ‘owned’ the desk area, the keyboard and the computer screens, while brokers controlled the information packs, notepads and reports that they carried into the space, and the orthogonal seating arrangement meant that reinsurers were positioned facing their computer screens, while brokers sat to the side facing the reinsurer.

However, these materials were used in some very proactive ways. On some occasions actors worked with material objects independently. In episode 1, for example, Nigel concentrated on his computer screen for 45 seconds, studying figures and switching between spreadsheets, while James sat quietly watching him (see Image 2). A similar pattern was observed in episode 2, when Ben (the broker) focused on his notepad (see Image 6).

On other occasions, however, both actors interacted collaboratively with the same object. In episode 2, for instance, Ben and Mike used a common object (e.g. a coastal map) to illustrate the risk (see Image 5).

Ben takes a US coastal map out of his briefcase and places it on the desk in front of Mike. He then borrows a highlighter pen and starts drawing circles on the map to illustrate the risk. Mike is immediately drawn to the document and it becomes a focal point for the ensuing discussion. Both Ben and Mike touch, point and write on the map. They number the zones, point to them, mutually assign values and discuss the differences between Florida and North Carolina.

Finally, peripheral objects (briefcase, telephones, pens, photographs) were sometimes enrolled in performing the episodes. An example of this occurred when Nigel focused on his computer, essentially ignoring James. In the field note below, we describe how a simple act of taking, opening and commenting on a peripheral object, such as a sweet, reengaged Nigel and allowed James to continue arguing the merits of the deal.

After a long period James starts to fidget, exhale breath and tap the table, seemingly to distract Nigel and reengage his attention. Eventually, James stretches out his right arm and takes a sweet out of a jar on Nigel's desk, keeping his eyes fixed on Nigel (see Image 2). As he unwraps the crackly plastic wrapping of the sweet he makes an incredible amount of noise. He places the sweet in his mouth, utters a sound of contentment (Mmmm) and comments on the taste of the sweet. Nigel suddenly looks up, smiles and says: ‘They're very good aren't they!’

Bodily resources

Bodily movements enabled the use of materials. When actors focused on the same material object – e.g. a figure on a screen, a calculation jotted down on a notepad, a piece of information in a document – they used their bodies (pointing, drawing etc.) to make points, direct attention and signal areas of agreement and disagreement. For example, in episode 1:

James sits down and places his information sheet on the desk in front of Nigel, directing his attention to the revised exposure figures he wants him to read (see Image 3). They both ‘lean in’ and look at the sheet together, discussing the information it contains.

A similar pattern occurred later in this episode when Nigel ‘turns his computer screen so that James can see it’ (field note), triggering a transition into a new bodily formation in which Nigel and James studied the exposure figures together and Nigel used the material to justify his EuroRe proposition. When actors focused on different objects in the space (see Image 2), by contrast, one actor interacted with the material(s) – e.g. turning pages, reading, typing, moving mouse, pushing buttons, writing etc. – while the other sat relatively still, often keeping his gaze fixed on his counterpart, although, as we noted above, brokers would sometimes fidget, exhale or tap tables to try to distract reinsurers to regain their attention. Actors did not, of course, always focus on materials. They sometimes focused on each other (see Images 1 and 4), using subtle facial expressions (e.g. smiling, frowning, rolling eyes, exhaling etc.) and bodily movements (e.g. nodding, looking left and right, leaning in).

The creation of spaces for work

The active use of speech resources (including silence), material resources and bodily resources enabled actors to co-create and work within distinct kinds of spaces associated with the connections illustrated in Figure 1. Each type of connection manifested a different kind of space, which we have categorized as ‘mutual space’, ‘dialogic space’ and ‘restricted space’. Within each of these spaces we found particular types of activities being performed, which are summarized in Table 2. We now look at each in turn.

Table 2. Representative data illustrating the spaces created, the activities performed and the strategic work accomplished

The construction of spaces within spaces	Types of activities	Representative data	Strategic work accomplished
1. Mutual space Constructed when actors use constellations of semiotic resources to create material connections wherein both actors focus on the same material object – e.g. a figure on a screen, a calculation jotted down on a notepad, a key piece of information in the information pack	1a. Directing each other's attention	The broker points at the page on the desk and explains the background behind the company data (‘That's sort of the general stuff’). The reinsurer says, ‘what's the combined ratio?’ The broker turns the pages and says, ‘If I take you to surplus first, and I haven't forgotten, we'll go on to combined ratio’. He points his index finger at some figures: ‘in 2008, 9.75. Then 11.51 so.’ The reinsurer says ‘yep’ and nods. He then picks up a pencil and starts jotting down notes while the broker holds the page open, points and talks. The reinsurer then points to a figure at the sheet and asks, ‘What is the difference between 11.5 and 10.2?’ He then uses both hands and points at the two different figures at the same time and says, ‘What of that is modelled and what of that is actual?’ The broker is not sure and says, ‘I will speak to the client and try to strip that out for you’ (episode 6) The broker is now instructing the reinsurer what to type into the search engine. He guides him to the webpage saying, ‘Click that. Now scroll down, stop.’ He points at a table on the screen but says nothing, leaving the reinsurer to study the table. After 10 seconds the broker says, ‘See, all around there is mainly gas’. The reinsurer says, ‘thanks for that’. He then reattaches his screen protector so the broker can no longer see his screen (episode 7) There are two open documents, a large open book and a notepad on the desk. The broker and reinsurer lean in with their heads and gaze, both focused on the open book. They are both pointing at figures on the page. The reinsurer copies some of the figures to a notepad with his fountain pen (episode 5)	Collaborative work Created the opportunity for a reinsurer and a broker to accomplish tasks together. The practitioners worked concertedly with common materials to align meanings and establish areas of common ground. They directed each other's attention to salient, critical and ambiguous information, enabling them to ‘zoom in’ on what was important. They sketched illustrations to co-define the parameters of risks and explored the possibility of allocating capital. By working collaboratively reinsurers made brokers ‘aware’ of the commercial underpinnings of their positions, providing the ground for negotiating work (below) and making it easier for brokers to accept decisions because they could see ‘how’ and ‘why’ a reinsurer got to a position
	1b. Sketching illustrations pertaining to the deal	The broker explains, ‘So the 10 over 5 would pay 11¼ … and the 25 over 15 is paying about 4½’. As the broker talks the reinsurer sketches a picture of the programme on his notepad with the figures. He then draws a square around one of the layers and tells the broker, ‘That's the one we'd probably look to write 6% in’. He has circled the 25 over 15 layer. As they talk the reinsurer adds some notes and they build up a fuller picture of what that layer entails. The broker then leans in, points with his pen and says, ‘would you be able to have a look at this one as well?’ He points his pen at the 10 over 5 layer represented on the notepad (observational note) The broker asks for a pen. He places a blank sheet of paper on the desk in front of the reinsurer and draws a large rectangular box with lines across to indicate the different layers on the programme (there are four). He writes figures in the top two rows and starts to explain the structure of the programme and areas where the reinsurer could write. The reinsurer is following the broker and nodding as he draws. He then starts to speculate about where he could write, pointing his pen at the different boxes as he talks (episode 10)
	1c. Mutual calculating	Graham (broker) said that it should be calculated as if it is only halfway into the layer. He points again to something on the screen and says, ‘It just adds these figures here’. Stephen (reinsurer) said ‘it is coming out at 808’. Graham said ‘no, it's 717’. Stephen said, ‘Well we're quite a bit off’. Stephen is playing with the figures on the screen and he shows Graham that these are the layers and these are the risks for the layers (observational note) The reinsurer says, ‘So that is 60’. The broker responds, ‘No it's 57.4’. The reinsurer writes on his pad and picks up his calculator. The broker leans in so they are in material connection. He says, ‘I tell you what, let's do a little trick here. Can we work it out without using a calculator?’ The reinsurer jokes, ‘Well, you probably can because you already know what the figure is’. Laughing he continues, ‘When you say little trick this is like all the other tricks’. They are now both laughing with the broker denying the accusation. They both work through the financial figures and percentage calculation together, pointing as they work. The broker estimates first and the reinsurer then confirms with his calculator (episode 10)
2. Dialogic space Constructed when actors use constellations of semiotic resources to create moments of face orientation wherein actors fix their gaze on each other and adopt a mirrored body position	2a. Discussing proposals and terms	‘This could be the worst business I've seen’, the reinsurer says, tongue-in-cheek with a grin. The broker retorts, ‘Oh come on, it can't be’. The reinsurer says, ‘Well, it's one of them! Look, it's really not of interest at all. I'm sorry we are not going to write it.’ The mood becomes a little more serious. The broker says, ‘Really? You're not going to write anything?’ The reinsurer says, ‘No, sorry, I really couldn't justify it’. The broker says, ‘Just do half a line?’ But the reinsurer is firm, ‘No, sorry, I can't. I can't see the first bit of logic.’ There is a momentary pause before the broker says, ‘Alright, I won't waste your time arguing’ (episode 8) Broker: ‘OK 23½ online deal?’ The reinsurer says, ‘You're having a laugh aren't you?’ ‘What?’ replies the broker. ‘No, 22.75 is the offer … it's worth everything as before, I'm giving you a gift!’ Broker says, ‘OK, 22.75’ (episode 7) ‘OK, I am going to reduce my line on the three layers we write.’ The broker says, ‘Why are you doing that?’ The reinsurer says, ‘because I am not happy with the price or what's happened with the loss’. They both look at each other for 5 seconds without speaking. The broker says, ‘The loss is not a big problem; it's still paying very good rate on lines’. Reinsurer: ‘But it was a big loss and there is the growth in exposures.’ The broker tries to convince the reinsurer that the client's risk reporting systems are robust, but he thinks they have been under-reporting exposures. The broker says, ‘So what are your thoughts?’ The reinsurer explains, ‘We are going to cut back on that top layer, so I'll do 2’. The broker says, ‘You are halving your line … why do you take such a drastic view on this?’ The reinsurer says, ‘With the growth in exposure we just can't take a chance on it’ (episode 9)	Negotiating work Created the opportunity for a reinsurer and a broker to engage in face-to-face dialogue. Negotiating work enabled the practitioners to reconcile differences, progress the capital allocation decision and establish the grounds for furthering acceptance or decline of the deal. Using this space the practitioners discussed terms and exposures, they put forward capital placement offers or alternative capital solutions, they explained the rationales underpinning their decisions and they tried to work through sticking points. While each practitioner searched for a more attractive position, they were respectful and sensitive to the other's position. The subtle use of silences, facial expressions and gestures allowed the actors to signal areas of (dis)satisfaction, and invite each other to elaborate upon, and alter, positions
	2b. Signalling areas of (dis)agreement	Practitioners used a repertoire of facial expressions (grimacing, smiling, frowning), head movements (nodding, moving head side-to-side, looking up to sky) to signal areas of (dis)agreement and (dis)satisfaction. In episode 8 (see 2a), for example, the reinsurer nods and smiles when he says, ‘This could be the worst business I've seen’. The broker raises his eyebrows and curves his mouth downwards, conveying disbelief when he says, ‘Oh come on, it can't be’. They disagree. Later the broker shrugs his shoulders when he says, ‘If you are not going to do it, you're not going to do it’ signalling resigned agreement (episode 8) In episode 9 there are a lot of facial gestures used in what is quite a difficult negotiation. The broker physically grimaces, as if in pain, and exhales breath when the reinsurer tells her he is going to reduce his line, signalling both disagreement and frustration with the reinsurer's position. As the reinsurer explained his rationale, she maintained an expression of confusion with her forehead scrunched up, again signalling disagreement with the reinsurer's argument that the loss is a problem (episode 9)
	2c. Inviting each other to elaborate upon, and alter, proposals	The reinsurer (Peter) says, ‘let's say it pays 22½. What would we do?’ There is a 3-second pause. Peter says, ‘5% across the board’. He is looking directly at the broker (Mark) who says, ‘OK’ and nods. However, he just continues looking at Peter without speaking for 8 seconds. Peter eventually breaks the silence and starts to elaborate on why he is proposing 5% (episode 4) The broker says, ‘the first layer is where we have an issue’. The reinsurer immediately says, ‘I don't have any movement on that … [3-second silence] … I really don't.’ The reinsurer continues, ‘I know, but that's the mark they [client] have got in the sand. The quote you have yesterday (22½), is there any way you can move down on that? We are not asking you to radically re-underwrite this and think about a huge shift, it's just a perception gap on the first layer.’ He will not alter his position, ‘No I can't. The price reflects the exposure; the underlining business that they are covering is not easy stuff that doesn't have losses. I said to you when we started this process I will do what I can, I will give you where I really can go to. Yeah, and then you come back and say I want a bit more’ (observational note)
3. Restricted space Constructed when actors use constellations of semiotic resources to create moments of disconnection wherein practitioners focus on ‘different’ things	3a. Absorbing information and making notes	As the broker talks the reinsurer writes down the details on his notepad; he does this without looking up. The broker then hands the reinsurer two individual sheets of paper with additional analysis on. The reinsurer pushes his keyboard and notepad forward and quietly reads them for 23 seconds. He then says, ‘You haven't revised the EPI number…’ (episode 7) The reinsurer studies the screen and starts to write notes with his fountain pen. He copies figures from screen and writes them down. While this is happing the interaction is silent (episode 5)	Private work Private work enabled practitioners to temporarily disengage and work in restricted space to which they had exclusive access and were free from interruption. Private work allowed reinsurers to consider and make progress on a decision of whether or not to allocate capital to a particular deal, by allowing them to absorb complex information, wrestle with complex deals, undertake detailed analysis, consider commercial options, generate viable commercial positions and think through possible solutions
	3b. Doing private calculations	Peter (reinsurer) says, ‘So we are now thinking of doing 2 over 3 yeah’. He taps his keyboard four times, takes his calculator out of his drawer and works out some figures. He then returns to the screen and enters figures using the keyboard. He then puts his elbow on the table and rests his chin on his left hand. He is now deep in concentration. He stares intently at his screen – for 41 seconds in silence. He then works on the calculator again, looking back at his screen and entering new figures. He picks up his notebook and thumbs through to a page and starts to write on his notepad. He continues switching between his calculator, writing on the notepad, and looking at the screen. After near 3 minutes of silence Peter says, ‘My head says 25, but I am going to call it 22½’. Throughout this 3-minute period Mark (broker) sat still quietly watching Peter work (episode 4) The reinsurer is now staring directly at the notes on his pad whistling to himself, while the broker just watches. He says, ‘So that's 53.8’. He does not look up or make eye contact with the broker; he just continues looking at the sheet. He then takes the mouse and does some calculations on his computer. He eventually looks up and says, ‘What I am just trying to work out in my own mind is. … I mean they have got to be about 5% up this year?’ (episode 5)
	3c. Focused thinking	The broker is pushing the reinsurer to place more than 5% capital. The reinsurer is thinking about it, but seems to be struggling to justify an increase. He looks straight ahead into space, but does not speak. 54 seconds of silence passes. He suddenly says, ‘It's because it's low, that's the issue’. Silence ensues for another 32 seconds. He then says, ‘OK, 7.5%. It's just because it's low’ (episode 4) The reinsurer sits perfectly still with his right elbow resting on the desk and his chin resting on his right hand (palm up). He stares intently at the screen and does not move for 21 seconds (episode 6)

1. Mutual space

Mutual space was constructed when actors used constellations of semiotic resources to create ‘material connections’ wherein both actors focused on the same material object, such as a figure on a screen, a calculation jotted down on a notepad, a key piece of information in the information pack (see Table 2, section 1, and Images 3 and 5). Mutual space arose as actors used available semiotic resources to ‘share’ materials, such as when they shared the desk, worked through an amended analysis together, or focused on the same notepad. When in mutual space both parties were able to work with and manipulate material objects in a mutual fashion because control of the space and the material had been temporally relinquished, often through some form of invitation (discussed below). To further illustrate the work performed in mutual space, we introduce a third representative strategic episode.1

Adam (broker) and Harry (reinsurer) are discussing a European earthquake and flood deal. Harry has written this deal for several years, but he tells Adam he may have to reduce his capital allocation this year given its poor loss record and increasing exposure. As he is explaining to Adam, Harry turns his computer screen so that Adam can see, pointing at the losses column in the spreadsheet. Adam looks and says, ‘I know, but the pricing is better than last year’. Adam then lays a sheet labelled ‘additional analysis’ on the desk in front of Harry and points to some remodelled data, which he explains as providing more favourable margins. They study the sheet together and discuss the veracity of the data. Harry jots some figures while they talk, does some quick calculations on his calculator, and writes the calculated figures on the margin of Adam's sheet. He then circles and points to the figures with his pen as he talks.

In this, and our other episodes, we noticed particular types of activities being done in mutual space which we labelled ‘Directing each other's attention’, ‘Sketching illustrations pertaining to the deal’ and ‘Mutual calculating’ (see Table 2, section 1). As episode 3 shows, in mutual space actors often turned their screens or placed documents to direct each other's attention to particular files, figures, maps, diagrams, passages etc. Such directing activity raised the actors' sensitivity to particular materials, encouraging them to zoom in on, and jointly engage with, information embedded in the object. This activity often prompted questioning (‘What is the difference?’), surfaced knowledge gaps (‘I don't know’) and motivated future action (‘I will strip that out for you’).

Mutual calculating was another activity performed in mutual space. In Table 2 (1c), for example, the reinsurer and broker worked closely to perform calculations and make sense of the figures. Working together, both actors provided input, vocalized assumptions, exerted influence and worked towards a common understanding of what the figures represented. In such situations, actors often sketched illustrations pertaining to the deal, such as drawing blocks that represented the specific layers of risk. For example, in episode 2, Ben drew circles on the US coastal map to illustrate the risk to Mike.

2. Dialogic space

Dialogic space was constructed when actors created bodily connections wherein they looked directly at each other's faces (see Images 1 and 4). In dialogic space both parties were able to perceive and monitor each other's speech, silence, gaze and expressive gestures. To illustrate the activities performed in dialogic space, we can look at another extract from episode 3.

Adam and Harry lean out and are now looking directly at each other. Harry says, ‘OK, I'll do 8.5% on all the layers’. Adam says ‘right’ and stares back at Harry without speaking, but gently nodding his head. After 7 seconds of silence, Harry starts to justify his offer by referring to the overall firm portfolio: ‘Our European book is really tight now’. Adam says, ‘Look, take it to up to 9% and the client will be happy. You'd be doing me a favour.’

This moment of interaction was typical of the activities performed in dialogic space, which always involved ‘Discussing proposals and terms’, ‘Signalling areas of (dis)agreement’ and ‘Inviting each other to elaborate upon, and alter, proposals’ (see Table 2, section 2). Dialogic space invited a discussion between the parties about issues such as terms, capital allocation and commercial proposals within which each could advance their interests and negotiate outcomes. In this space, actors utilized an array of expressive physical acts, including facial expressions (e.g. grimacing, smiling, laughing, frowning), head movements (e.g. nodding, moving head side-to-side, looking up to sky) and expressive hand gestures (e.g. using hands to fabricate images) to signal areas of agreement and disagreement and build common understanding.

An important semiotic resource frequently employed in this space was conversational silence. Conversational silences were used by actors to invite each other to elaborate upon, and alter, their positions. In episode 3, for instance, we saw how Adam's silence conveyed dissatisfaction with Harry's 8.5% offer, which encouraged Harry to elaborate. The same happened in episode 1 when James's short silence, following Nigel's EuroRe proposal, signalled to Nigel that the proposal was suboptimal, prompting Nigel to further explanation. Similarly in episode 2, Ben's silences appeared to invite Mike to elaborate upon, and potentially alter, his position.

3. Restricted space

Restricted space was constructed when actors used constellations of semiotic resources to create disconnections, wherein actors oriented their bodies, gaze and attention to different objects in the space (see Table 2, section 3, and Images 2 and 6). Restricted space was created when actors used semiotic resources to compartmentalize or bracket a space, which enabled them to conduct analytical work privately without interruption or scrutiny from the other party. To provide an example of the activities performed in restricted space, we provide another extract from episode 3.

Harry turns his body away from Adam and focuses on his computer screen, while Adam sits quietly. Harry taps away at his keyboard, entering figures, and flicking back and forth between two spreadsheets he has open. As he does this he jots down two figures on a small yellow notepad. He drops the pencil and lifts both hands into a prayer-like position, resting them against his lips. He stares intently at the screen (27 seconds) while tapping his two index fingers together. He is deep in thought. At one point Adam asks ‘what do you think?’ but Harry doesn't answer him; he just continues looking at his figures.

This fragment is typical of the activities performed in restricted space. When one actor was in restricted space, the other person's ability to take control or observe the ‘doings’ of the other was significantly restricted because material positioning, bodily orientations and norms of conduct prevented him from doing so without invitation. For instance, Adam could no longer see Harry's screen because Harry had returned it to its original position. Even though Adam used verbal talk to try to interrupt Harry (‘what do you think’) he was restricted from entering Harry's space. In our study, it was primarily reinsurers who generated restricted space because the spatial–material configurations privileged their ownership of the space (e.g. they faced the computer screen and it was their desk). Norms of conduct also privileged the reinsurer withdrawing into restricted space to work on the information brought by the broker, whereas brokers were there to interact and hence had less reason, by custom and work performed, to construct restricted space. Even so, we did sometimes see brokers construct restricted space, for example to read a document, do a quick calculation or consider a piece of information to proffer.

When in restricted space, actors performed private calculations; doing analytical work, calculating on spreadsheets and comparing figures, as well as absorbing information and making notes, including reading documents and writing on private notepads. Focused thinking went hand-in-hand with these activities as actors studied figures, evaluated information, wrestled with alternatives and formulated commercial proposals. While we cannot know for sure what actors were thinking while performing this activity, the visual cues strongly suggest this is what was happening. In episode 1, for example, Nigel took regular ‘time-outs’ to consider information presented, contrast it with his commercial analysis and think through options that he then articulated. Moreover, as shown in the post-interaction reflection following episode 2, our post hoc conversations with reinsurers often confirmed our impressions and interpretations.

Moving between spaces to perform different activities

In sum, our fine-grained analysis revealed that actors created three distinct spaces that enabled them to perform particular activities associated with allocating capital on reinsurance deals (see Table 2). Our categories point to those activities that were germane to each space. For example, mutual calculating could only happen in mutual space, just as we only saw private analytical work in restricted space and actors inviting further elaboration in dialogic space. In other words, these critical activities could only happen because the actors constructed the appropriate spaces.

As our representative examples show, the activities accomplished in these episodes shifted fluidly and easily between the three spaces. These transitions were shaped by the momentary circumstances created in the episode. Sometimes transitions were self-instigated and instantaneous, such as when a reinsurer broke eye contact and switched attention to a document, moving the interaction into restricted space, or turned and focused on the other person, moving the interaction into dialogic space. Other times transitions were invitational, wherein one of the actors would invite the other to enter into mutual space, such as when a reinsurer turned his or her screen towards a broker and invited him or her to look at figures or when a broker placed an information sheet on the desk and invited the reinsurer to look at it. On a few rare occasions, transitions were triggered by disruptive acts, such as when an actor made a comment, tapped a pen on a desk, or performed some other disruptive act that led to a change in the space. Recall how in episode 1 James (broker) used his body (e.g. fidgeting, tapping) and materials (the sweet) to disturb Nigel and move the interaction from restricted space into dialogic space. Finally, transitions were sometimes brought about when an actor closed off space, such as when reinsurers returned their screens back to their original position, out of sight of the broker, or when brokers took back documents that had been the focus of attention.

Discussion

In this paper we set out to examine the question: how is strategic work accomplished through the orchestration of material, bodily and discursive resources? We located our study within the context of strategic episodes, those activities that are both part of everyday work and yet strategically consequential for organizations (Floyd and Lane, 2000; Hendry and Seidl, 2003; Jarzabkowski, 2008; Jarzabkowski and Seidl, 2008). We now draw our findings together into a conceptual framework, illustrated in Figure 2, that revolves around three analytical layers: the creation of spaces; the enactment of strategic work; and the accomplishment of strategically consequential outcomes (hereafter, strategic outcomes). Each of these layers is recursively entwined with the other and each layer is manifested in a particular orchestration of semiotic resources.

As shown in our first-order findings and illustrated in our framework (Figure 2(A)) actors use constellations of semiotic resources – e.g. bodily, speech, material – during interactions, and combine these according to their practical coping with the situation at hand (Chia and Holt, 2009). While there is no predetermined order to employing semiotic resources, their combination creates different connections and spaces for work (Figure 2(i)). We therefore conceptualize these semiotic constellations as important building blocks that shape the conduct of a strategic episode. Specifically, we linked the different semiotic constellations to the construction of different ‘spaces-within-a-space’ during the conduct of an episode. We found that actors construct three types of space – mutual space, restricted space and dialogic space. A strategic episode is accomplished through continuous and fluid transition between these three spaces (Figure 2(ii)), each of which is characterized by distinct activities (see Table 2). As shown in the findings, the co-construction of these spaces is consequential for the types of activities performed within a strategic episode.

The specific clusters of activities in different spaces constitute different types of strategic work: private work performed in restricted space, collaborative work in mutual space and negotiating work in dialogic space (see Figure 2(B)). These three types of strategic work underpin the interactions taking place, shaping the conduct of the episode and its accomplishments. For example, our study showed how fluid transitions (Figure 2(ii)) between the different types of strategic work shaped reinsurers' consideration of capital allocation decisions and their negotiations with critical business partners about these decisions. There is no prescribed pattern in these types of strategic work. Rather, each episode was performed through fluid transitions back and forth between collaborative work, private work and negotiating work according to the unfolding nature of the interaction. Importantly, our study demonstrates that the strategic work performed in each space, and the transitions across spaces, contributed to accomplishing two strategic outcomes (Figure 2(iii)). First, they underpinned enactment of the reinsurer's current strategy within the specific moments of interaction, as each capital allocation decision on a deal is directly tied to enacting the strategic portfolio of the firm. Second, they helped to construct grounds for future strategic work by maintaining critical relationships with brokers, with whom these reinsurers must work daily on pursuing capital allocation opportunities for their firms. While the specific content may vary in different strategy contexts, we suggest that enacting the current strategy whilst also constructing grounds for future strategic work are likely to be relevant outcomes for any strategic episode (see Figure 2(C)). We now turn to the contributions arising from our framework.

Contributions

Our conceptual framework, showing the association between semiotic resources, spaces, strategic work and strategic outcomes, makes several important contributions to the research on strategy-as-practice. The main contribution is our demonstration of how strategic work is accomplished in the orchestration of speech, bodily and material resources (see Figure 2). We show that the construction of spaces is integral to the strategic work performed, and that fluid shifts between these types of work and their associated spaces are critical for accomplishing strategic outcomes. Our study thus elaborates on our understanding of what constitutes strategic work, showing that it constitutes a multimodal accomplishment. Thus far, studies in the strategy-as-practice field have investigated these semiotic resources largely in isolation. While there have been calls to examine a wider range of strategy practices, particularly bringing in multimodal elements such as emotions, bodily positions and material artefacts (e.g. Chia and Holt, 2006; Jarzabkowski and Spee, 2009; Jarzabkowski, Balogun and Seidl, 2007; Vaara and Whittington, 2012), empirical research has largely privileged talk in its various forms over the multifarious affordances of other semiotic resources (LeBaron and Whittington, 2011). Even those studies that have teased apart some of these other semiotic resources (e.g. Jarzabkowski, Spee and Smets, 2013; Kaplan, 2011; Liu and Maitlis, 2014) have tended to focus upon a subset, such as material or bodily resources. Our findings and framework highlight the multimodality of strategic work, and show how such resources cannot be considered in isolation but rather as constellations of semiotic resources within which particular spaces and activities are mutually constituted.

Our framework also advances research on strategic episodes (Hendry and Seidl, 2003; Hoon, 2007; Jarzabkowski and Seidl, 2008). While strategy-as-practice research has emphasized the importance of analysing strategic episodes within the overall flow of strategy making (Hendry and Seidl, 2003), we lack an understanding on the role of multiple semiotic resources in enabling and/or constraining the conduct of episodes. First, our study demonstrates the way strategic episodes are constructed in the orchestration of multiple semiotic resources that create spaces for distinct types of strategic work to be performed. Second, it demonstrates the fluid transitions between episodes as the orchestration of semiotic resources shifts, extending current findings that have examined more on the structuring characteristics of strategy meetings (e.g. Haug, 2013; Jarzabkowski and Seidl, 2008). These prior studies have treated space as a mere background, e.g. referencing the withdrawal to a particular meeting room or off-site location (e.g. Jarzabkowski and Seidl, 2008; Johnson et al., 2010). By contrast, our study makes space an active concept within the analysis of strategic episodes, showing how it is constructed and continuously shifts in accomplishing the episode. Drawing on these insights, future research may explore distinct features of strategic episodes such as the initiation, alternative or competing patterns of conduct and termination and the way these are accomplished within the construction of distinct spaces.

Our research also contributes to the strategy-as-practice agenda to study the actual ‘work that comprises strategy: the flow of activities such as meeting, talking, calculating, form filling, and presenting in which strategy is constituted’ (Jarzabkowski and Whittington, 2008, p. 282; Whittington, 2003) by demonstrating how the construction of spaces and work are entwined. While there have been many calls to study the work of strategy making (e.g. Chia and Holt, 2006; Jarzabkowski, Balogun and Seidl, 2007; Vaara and Whittington, 2012), there are only a few explicit examples of what this work is or how it is performed (e.g. Regnér, 2003; Rouleau, 2005; Samra-Fredericks, 2003), particularly outside boardrooms, and in the work of actors other than top managers (e.g. Regnér, 2003; Rouleau and Balogun, 2011). Our study identified three types of strategic work – private work, collaborative work and negotiating work – and showed how distinct spaces, each created through the constellation of semiotic resources, enable and constrain this work. While transitions between spaces are undetermined and occur instantaneously, each type of strategic work is necessary and performed within a specific space. Yet there is no dominant order or sequence that suggests ‘one best way to strategize’. Rather, strategy ‘unfolds through everyday practical coping actions’ (Chia and Holt, 2006, p. 637); and we show that this practical coping comprises particular configurations of bodily, linguistic and material resources within which actors construct the spaces to perform strategic work. The power of our framework lies in this identification of multiple spaces, the configurations of semiotic resources that underpin them and the fluid transitions between them in performing different and interwoven types of strategic work.

Furthermore, we show that these forms of strategic work are linked to the strategically consequential outcomes of any particular strategic episode. While the consequentiality of micro-strategizing actions has typically been identified in retrospect (e.g. Regnér, 2003; Salvato, 2003), our study provides a more fine-grained understanding of how the outcomes of strategic work are performed in the everyday interactions of strategic episodes, so providing grounds for future research into strategic outcomes (e.g. Jarzabkowski and Kaplan, 2014; Johnson, Melin and Whittington, 2003).

Our study also advances understanding of semiotic resources and the way they shape strategic work in three ways. First, while many studies have pointed to the critical role of discourse in constructing strategic work (see Mantere, 2013; and Balogun et al., 2014, for reviews) silence as a discursive practice has largely been neglected. Yet, linguistic research has shown that conversational silences have ‘perlocutionary’ effects, such as communicating dissatisfaction or indifference, or persuading or convincing an addressee (Jensen, 1973, Kurzon, 2007). Our study demonstrates that silences comprise important strategic resources that, in combination with other semiotic resources, can signal that actors are withdrawing into restricted space, thus shifting the focus of strategic work, or can invite responses such as further clarification or justification during negotiating work. By demonstrating the importance of silence as a discursive practice that shapes the dynamic flow of strategic work, in interplay with other semiotic resources, we provide grounds for future research to extend analysis of strategy talk to include silences. Scholars may, for example, employ more ethno-methodological or conversation analytical approaches to further unpack how these silences function in different strategy contexts.

Second, our study extends insights on the implication of material resources in strategic work. While there are increasing calls to study the materiality of strategic interactions (e.g. LeBaron and Whittington, 2011; Vaara and Whittington, 2012), strategy-as-practice research to date has focused largely on textual artefacts such as PowerPoint presentations (e.g. Kaplan, 2011) or strategic plans (e.g. Spee and Jarzabkowski, 2011). Our study demonstrates the mutability of multiple artefacts in the construction of spaces. For example, a computer screen might be turned towards the other party to construct mutual space in discussing the pricing of a deal (collaborative work), or it may be used as a focal point by a single actor to construct restricted space, so enabling the private work of calculation to take place. The same screen thus contributes to the construction of quite different spaces that afford varying types of strategic work, according to the way it is employed within different configurations of semiotic resources. Our study thus emphasizes the importance of the material turn in strategizing research (Balogun et al., 2014; Jarzabkowski, Spee and Smets, 2013; LeBaron and Whittington, 2011) and suggests a rich set of possibilities for investigating multiple artefacts, not only those that are overtly ‘strategy tools’, for their role in enacting spaces and, concomitantly, strategic work.

Third, our study elaborates on nascent research into the implications of bodily resources in the strategizing process (e.g. LeBaron and Whittington, 2011; Liu and Maitlis, 2014). For instance, Hodgkinson and Wright (2002) illustrated how a CEO sabotaged a scenario planning meeting through her bodily moves and gestures during a strategy workshop. While current studies have shown how gazes and gestures comprise part of the emotional dynamics associated with different types of strategizing (e.g. Liu and Maitlis, 2014), our study illustrates the implication of bodily resources in constructing different spaces for strategic work as well as, in conjunction with other semiotic resources, the fluid shifts between spaces. For instance, averting gaze from face-to-face orientation to quietly looking at the computer screen helps construct the shift from negotiating work to private work. While bodily resources have largely been neglected in empirical research, our study further stresses the need to include bodily orientations in the study of strategic work and its accomplishments. Future research might, for example, address the implication of seating arrangements and the movements of top management team members on the conduct and outcome of a board meeting (see LeBaron and Whittington, 2011) or the embodiment of a firm's strategy in everyday interactions.

Finally, our conceptual framework provides grounds for future research to adopt an explicitly multimodal approach to exploring strategic episodes in more detail, in particular drawing upon the specific traditions of ethno-methodology to further understanding in the strategy-as-practice field (e.g. Goodwin, 2007; Heath, Hindmarsh and Luff, 2010; Kendon, 1990; LeBaron, 2005; Samra-Fredericks, 2003; Streeck, Goodwin and LeBaron, 2011). A multimodal approach could explore the array of semiotic resources and sensory opportunities available to actors in any given moment. In doing so, such research would build upon a growing interest in using video-recordings to examine the multimodal nature of strategizing (LeBaron and Whittington, 2011; Liu and Maitlis, 2014) and address recent calls for research into the materiality of strategy-as-practice (e.g. Dameron, Lê and LeBaron, 2012; Vaara and Whittington, 2012).

Conclusion

In this paper we have developed a conceptual framework for examining the association between, and mutual constitution of, semiotic resources, spaces, strategic work and strategic outcomes. We focused on a particular setting of strategic episodes in the professional services context of reinsurance trading, where reinsurers, as professional actors, have considerable autonomy to enact aspects of their firm's strategy. The conceptual framework may therefore be particularly useful in similar strategy contexts that deal with abstract, knowledge- and relationship-based strategies such as trading, professional advice, branding and reputation (e.g. Fauré and Rouleau, 2011; Løwendahl and Revang, 1998). However, we also suggest that our framework is conceptually valid for other settings of strategic episodes. For example, we expect that different actors will experience different types of material and/or bodily connection as they withdraw from, engage with and influence the flow of strategic planning meetings. Similarly, our findings suggest that actors have an array of semiotic resources with which to advance their agendas, invite other participants to interact with them on strategic work, or bracket themselves out of the strategy meeting in order to pursue their particular interests. These include resources that have had little attention in strategy-as-practice research, such as silence, body and materials.

Our study aimed to contribute to an agenda for research into the way strategic work is shaped by the interplay of multiple semiotic resources; yet, we recognize the limitations of our theorizing. While others may not find the precise strategic work and spaces that we identified such as private work and restricted space – they may find some other space in addition to, or in place of, those we found because the configurations of resources are differently employed – we suggest that such arrangements are liable to be generalizable to multiple contexts in which actors interact in strategic episodes. Board meetings and strategy workshops involve multiple materials and spatial arrangements, including presentations, spreadsheets, tables, seating orders, projector screens and flipcharts, all of which both shape and also provide resources for the bodily, material and conversational interactions between participants. Thus, our findings, while context specific, offer a set of concepts and a framework with which to advance the study of strategic work and spaces.