Towards retail innovation and ambidexterity: insights from a Swedish retailer

Purpose – Incumbent retail organisations need to develop new capabilities to adapt with the increasingly competitive retail landscape. Despite the growing relevance of innovation in retail practice, the strategic management of innovation in retailing is still vastly under-researched. This explorative study thus aims to investigate how incumbent retail firms can organise for innovation from an organisational ambidexterity perspective. Design/methodology/approach – A single-case study of an established Swedish retail firm was conducted from December 2016 to July 2018 and followed up in June 2021. The authors followed the process of implementation of organisational changes aimed to increase innovation in the company, particularly the introductionofadigitalmarketinginitiativeandacorporateinnovationhub.Datacollectionwasbasedonnine semi-structuredinterviewsandparticipantobservationsandunstructuredinterviewsfrom13meetingsandworkshops.Anabductiveapproachtodataanalysiswasfollowed,iterativelycomparingtheoreticalconcepts andempiricaldatausingopen,axialandselectivecodingtodistilfindingsintoaggregatedthemes. Findings – Given the inherently limited formalisation of innovation processes in most retail organisations, structuralambidexterityappearstobenecessarywhentheaimisradical,strategicretailinnovation.Structuralmechanismsareabletosafeguardthespaceandresourcestofocusonlong-termresearchandprojectswith higherriskanduncertainty;however,integrationofinnovationactivitiestothemainstreamorganisationiscritical.Pursuingcontextualambidexterity,whereininsteadofstructuralsolutions,employeesareempowered todivideemployees ’ timebetweeninnovation-relatedandefficiency-relatedtasks,ismorelikelyrelatedtoretail innovations that are incremental and operational. Originality/value – The paper contributes to the emerging topic of strategic management of innovation in retailing, by explicating how incumbent retailers can organise for innovation depending on the type of innovation that is aimed for, using organisational ambidexterity as a novel perspective.


Introduction
Retailing is a fast-paced industry that reflects the changes in technology, supply chain management and consumer behaviour (Roggeveen and Beitelspacher, 2018).Higher demands are being placed on retailers and there is an imperative to implement strategies that would allow them to adapt, innovate and transform to have sustained competitive advantage (Shankar et al., 2021;Grewal et al., 2021;Kumar et al., 2017).The internal capabilities that retailers have traditionally relied one.g.placing physical stores in convenient and accessible Retail innovation and ambidexterity 1 locations, buying power through expertise or scale and a deep understanding of what the consumer needs (Reynolds, 2004)need to evolve along with the changing retail landscape.New conceptual frameworks and theoretical models are necessary in understanding emerging retail practices, one of which is with dynamic capabilities (Paul and Rosenbaum, 2020).
Organisational ambidexterity is a dynamic capability (O'Reilly and Tushman, 2008) of firms to simultaneously manage the paradoxical demands brought about by explorative and exploitative activities (Duncan, 1976;Tushman and O'Reilly, 1996;Jansen et al., 2009;Raisch and Birkinshaw, 2008;Andriopoulos and Lewis, 2009).March (1991) proposes that exploitation, which is associated with activities such as "refinement, efficiency, selection, and implementation" and exploration, which is related to "search, variation, experimentation and discovery" (p.102) are fundamentally different learning activities that divide the attention and resources within a firm.They may, therefore, require different structures, strategies and contexts (Raisch and Birkinshaw, 2008).The management of ambidexterityand, therefore, the balancing of innovation with routine, non-innovation workis industry specific and studies have shown that ambidexterity has a more positive effect in dynamic environments (Simsek, 2009).In the retail context, organisational ambidexterity can be viewed as the capability to pursuing explorative and radical innovation whilst ensuring efficient execution of core retail business operations.However, innovation in retail is a highly complex process that is underresearched and is not yet well-understood (Hristov and Reynolds, 2015).Given the open, hybrid and transversal nature of innovation in this sector that leads retailers to "innovate differently" (Sundstr€ om and Radon, 2014;Sundstr€ om and Reynolds, 2014;Hristov and Reynolds, 2015), there is an opportunity to increase scholarly investigations of the strategic management of innovation in retail firms and how to organise for innovation in retail practice.
The purpose of this explorative research is, therefore, to investigate how incumbent retail firms can organise for innovation from an organisational ambidexterity perspective.A singlecase study is conducted with one of the leading traditional retail chains in Sweden (hereby referred to as "Swedish Incumbent Retail Firm" or SIRF).Being an established retailer with a long-standing focus on cost efficiency and continuous operational improvements, SIRF exemplifies the organisational challenge of developing ambidexterity, maintaining its dominant characteristics for exploitation (i.e. the core retail business) whilst allowing for a greater degree of exploration (i.e.innovation activities).We followed the process of implementation of organisational changes aimed to increase innovation in the company, specifically the introduction of a digital marketing initiative (hereby referred to as M20) and a corporate innovation hub (hereby referred to as I30).The paper contributes with a contemporary understanding of innovation efforts in retail using organisational ambidexterity as a novel perspective.Based on insights from the in-depth case study and relevant literature, we present a proposition on how incumbent retail organisations can organise for innovation from an organisational ambidexterity perspective depending on the type of innovation that is aimed for.

Literature review
2.1 Innovation management in retaila nascent research topic Innovation management is acquiring importance in the retail industry (Pantano, 2016), given that the inherent ability of retailers in innovating continuously is emerging as a critical element for success (Pantano, 2014;Hristov and Reynolds, 2015).However, it is a relatively nascent and fragmented topic in academic research.One research stream involves the application of technology management theories to retail settings (Pantano et al., 2017(Pantano et al., , 2022;;Pantano and Vannucci, 2019).Another approach which considers not just the technological side of innovation investigates retail business model innovation.Retail business model innovations are system-wide changes that result from changing one or more elements of the retail business model (retailing format, activities and governance), modifying the retailer's organising logic for value creation and appropriation (Sorescu et al., 2011).Retail business model innovation resulting from the increasing digitalisation of the retail sector is receiving significant scholarly attention (Mostaghel et al., 2022;Do Vale et al., 2021Jocevski, 2020).
From a firm perspective, however, the strategic management of retail innovation is vastly under-researched.Nonetheless, there have been some attempts to define unique characteristics of innovative activity happening inside retail firms (Hristov and Reynolds, 2015;Reinartz et al., 2011;Hristov, 2007) as well as retail innovation drivers and barriers (Albors-Garrigos, 2020;Pantano, 2014;Reynolds and Hristov, 2009) (see Table 1 for a summary).Depending on the organisational impact, the level at which they are managed and the amount of resources involved, distinctions can be made between operational and strategic innovations (Hristov and Reynolds, 2015).Operational innovation is predominantly (although not exclusively) continuous and incremental, funded through cyclical and shortterm planning and budgeting processes and can take place in any functional area but is usually related to store operations, marketing or new product development (Hristov and Reynolds, 2015).Strategic innovation, on the other hand, is guided by a strategic planning process, can be radical or incremental in nature, has a longer time horizon and is more likely to be a product of multidisciplinary teams (Reynolds et al., 2007).
The approach to innovation in retail has been less organised and more experimental, haphazard and opportunistic, due of the nature of the industry (Hristov, 2007).Innovation in retail firms has been shown to be "embedded in organisational cultural norms, beliefs and values" (Hristov and Reynolds, 2015).The absence of an established innovation function implies that innovation competes with other operational issues at the same location and level, i.e. the delivery of "on-going" business (Olsson and Johansson, 2016).
However, retailers' current ways of adapting to external conditions are being challenged in the face of accelerated technological and market disruption (Grewal et al., 2017).Retail organisations need to evolve to address changing consumer needs, new market opportunities, emerging technological forces and new types of competition.There is an opportunity for retailers to develop organisations that are capable of innovation and exploration, but also to find the right balance with existing capabilities for efficiency in retail operations.In the next section, we present how organisation ambidexterity encapsulates this issue.

Typologies of retail innovation
Retail innovation according to impact (strategic innovation or operational innovation) or according to application areas (offer/customer-related, support-related or organisational-related)

Organisational ambidexterity: simultaneous pursuit of exploration and exploitation
Whilst there is no consensus on its definition, organisational ambidexterity is essentially the capability of an organisation to simultaneously exploit existing competencies and explore new opportunities (Duncan, 1976;March, 1991;O'Reilly and Tushman, 2004;Raisch et al., 2009).Exploitation is coupled to "refinement, choice, production, efficiency, selection, implementation and execution" whilst exploration is about "search, variation, risk-taking, experimentation, play, flexibility, discovery, and innovation" (March, 1991, p. 71).The exploitative business is focussed on exploiting existing capabilities for profit whilst the explorative business is focussed on exploring new opportunities for innovation and growth (O'Reilly and Tushman, 2004).
The nature of exploration and exploitation being contradictory forcesand how organisations could reconcile their demandshas long been the focus of discussion in the innovation management and ambidexterity literature (see, e.g.Birkinshaw and Gupta, 2013;O'Reilly and Tushman, 2013;Andriopoulos and Lewis, 2009;Cao et al., 2009).Research in ambidexterity has identified structural or contextual configurations through which organisations can achieve the balance between exploration and exploitation (see Table 2).
Structural ambidexterity involves the separation of exploration and exploitation into independent units (Tushman and O'Reilly, 1996;Duncan, 1976).The underlying logic is the need to create room for the inherently different natures of exploitation and exploration on the one hand and integration or the balancing of both on the other, referred to as differentiation and integration mechanisms (Andriopoulos and Lewis, 2009;Jansen et al., 2009;Koryak et al., 2018).Contextual ambidexterity, on the other hand, suggests that ambidexterity can be obtained not through dual structural arrangements but by leaders creating a supportive business-unit context (Gibson and Birkinshaw, 2004).A single business unit can thus be ambidextrous if a supportive organisational context allows individuals to engage in both exploitation-oriented and exploration-oriented activities.
Research suggests that the management of innovation and ambidexterity is industry specific and that ambidexterity may have a more positive effect in dynamic environments (Simsek, 2009).There is an opportunity to explore the topic in the retail context, a dynamic industry that is undergoing significant transformation.Whilst there is an increasing heterogeneity of innovations emerging in the current competitive scenario, research still has not empirically identified common characteristics for the development and implementation of successful innovation strategies in the retail industry (Pantano, 2016).

Structural ambidexterity
Ambidexterity is facilitated through structural separation of explorative and exploitative units that require coordination and integration mechanisms including senior management integration.Separate structures are necessary to accommodate the opposing competencies, systems and practices of exploration and exploitation Tushman and O'Reilly (1996), O'Reilly and Tushman (2004), Smith and Tushman (2005), Jansen et al. (2009) Contextual ambidexterity Ambidexterity is facilitated within an entire business unit by creating an organisational context that enables individuals to make their own judgements on how to divide their time between the demands of adaptability (exploration) and alignment (exploitation).This organisational context is characterised by a combination of stretch, discipline, support and trust Gibson and Birkinshaw (2004), Ghoshal and Bartlett (1994) Table 2. Approaches to organisational ambidexterity

Methodology
The paper is based on an in-depth study of a retail company, using a qualitative single-case study methodology.The research design leans on the notion put forth by Yin (2014) in that the study centres around events in a single firm as the empirical object, exploring and investigating the "how" in relation to these contemporary set of events, over which the researcher has little or no control.The case study is a necessary and sufficient method for certain important research tasks and the value of conducting single-case studies in dealing with phenomena holistically has been discussed in literature (Flyvbjerg, 2006;Piekkari and Welch, 2018;Thomas and Myers, 2015).
The selection of this case is based on a previous multiple case study and research project on retail innovation of four major Swedish retail chains addressing how innovation is perceived and how it is managed on a more strategic level.Proper case selection is crucial in single-case study research and Stake (2000) suggests that the researcher should select a case that would offer the greatest potential for learning as opposed to using representativeness as a criterion.For this paper, having a narrower scope, SIRF, being one of the four companies, was purposefully selected as a particularly relevant case, based on how incumbent, leading retailers are being challenged by disruption in the retail industry and have started to face the innovation imperative head-on by implementing activities specifically addressing the need for innovation.SIRF is a leading grocery retail chain that owns and operates over 1,200 hypermarkets, supermarkets, convenience stores and discount stores mostly in Sweden, with over 8,000 employees and a market share of around 36%.
The data collection is based on participant observations and unstructured interviews from thirteen meetings and workshops and nine in-depth, semi-structured interviews (see Table 3 for details).The semi-structured interviews (between 60 and 90 min) as well as the meetings and workshops (between 1 and 4 h) involved individuals in the management team directly linked to the two SIRF initiatives being observed in this study which are digital marketing initiative M20 and corporate innovation hub I30.The main part of the data collection was conducted from December 2016 to July 2018, whilst a follow-up interview with the Head of I30 was conducted in June 2021.In the follow-up interview, information was collected about the developments since 2018.All interviews were recorded and transcribed, whilst field notes were taken during meetings and workshops.Field notes were organised into memos that provided an account of the content of

Retail innovation and ambidexterity
discussions but also our observations and more subjective reflections.Internal and public documents were also used as reference for supplementary information needed to build the case study.
Interview transcripts and memos were analysed using an abductive approach using an iterative process of going back and forth between theoretical concepts and empirical findings (see, e.g.Alvesson and Sk€ oldberg, 1994;Dubois and Gadde, 2002).Both in-vivo coding and constructed theoretical codes were used.Following Strauss and Corbin (1998) general guidelines on open coding followed by axial and selective coding, findings were distilled into aggregated themes.Throughout this process, the emerging patterns in the data were being compared with literature iteratively to see how our empirical findings correspond with earlier premises.
To address the validity and reliability of the research, several techniques were implemented.The data was collected and managed systematically; data and documentation by the different researchers in the research group were always made available and transparent in a shared database, for continuous cross-checking and comparison.Regular documented meetings and discussion sessions involving all members of the research team were also held to coordinate the data analysis process.Aside from doing extensive documentation which promotes reliability of the case study, triangulation of data sources was done to ensure the quality of the research (Yin, 2014).The data collected was extensive enough to answer the purpose of this study and provide a basis for a narrative of the journey that the case company was undergoing, initiating and implementing the two innovation initiatives.Moreover, feedback sessions, which can be a means of respondent validation (Gibbs, 2007), were held with the case respondents in different stages of data collection and analysis.Respondents were given the opportunity to contribute additional perspectives.As suggested by Gioia et al. (2012), qualitative rigour is enhanced when efforts are made to "give voice to the informants in the early stages of data gathering and analysis and also to represent their voices prominently in the reporting of the research, which creates rich opportunities for discovery of new concepts rather than affirmation of existing concepts" (p.17).These implemented techniques for triangulation, transparency and respondent checking are used to avoid obvious errors and omissions, but also give way to new lines of enquiry and novel interpretations (Gibbs, 2007).As a final measure for improved validity and reliability, the preliminary case study was subject to peer-review through a research seminar involving external researchers not part of the research team, using their feedback to further develop the case study.

The case of SIRF: enabling innovation through M20 and I30
This section summarises the case study results.Figure 1 depicts the data structure for the findings, which shows lack of exploration in SIRF and how this was addressed with organisational changes in the form of a digital marketing initiative (M20) and a corporate innovation hub (I30).

Addressing lack of innovation and exploration: a new digital marketing initiative
Innovation is highly relevant, but I wouldn't say that it is an established and well-used word here.From my perspective, we are in the greatest need to develop our capacity for innovation and how we can innovate and what innovation is in our world, in retail.But I would not say we're talking about innovation.At all.We talk about improvements, we talk about changes, we talk about launches.(. ..)Of course you can innovate in retail, but I think my colleagues associate innovation with other industries.I think we talk too little about innovation.(. ..)There are no processes, budgets, or organisation for innovation.No tools.And I don't think there is clarity or control.(. ..)From our department, we think a little self-righteously and had a self-imposed mandate, like no one has given us this mandate but we took it upon ourselves because we find it important for [SIRF], to become the sort of champions for innovation.But we had very little means to do it.

Retail innovation and ambidexterity
In 2016, the concept of "innovation" was not commonly used in the internal vocabulary across the organisation, whilst continuous improvements were more relatable concepts.Given the perceived changes in the business environment, however, senior management expressed a lack of a long-term strategy for innovation and acknowledged the need to explore new forms of business development and innovation.A digital marketing initiative (M20) was launched with the general aim to improve the company's ability to receive, incubate and develop innovative ideas within digital service development, essentially enabling faster adaptability in capturing new opportunities.M20 was primarily focussed on addressing the effects of digitalisation on marketing campaigns.
In this major program [M20] which is about digitalisation, when we talk of innovation, it is in a completely different way from what we have done before, which I think is good.(. ..)We are beginning to realize how we are going to work now, so it's starting to take off.
Part of the framework of M20 was to introduce agile working into the organisation.Agile and design-thinking inspired methods were used such as innovation toolboxes, hackathons, innovation days and design sprints.Thus, M20 was not merely a marketing initiative but an initial attempt to empower individuals in becoming innovative in their existing tasks, by encouraging cultural change and new ways of working.Although the value of entrepreneurship has always been championed in its corporate culture, SIRF being a large and bureaucratic organisation faces its own organisational challenges in adapting with a rapidly evolving consumer base.For many of the employees, M20 was thus seen as a clear direction towards a more active approach to innovation in SIRF.
We are extremely heavy on traditional marketing still and it is partly about history.We have always done it, it works, and in this industry it always worked well.And because we are a big company, we like old systems which means that we haven't actually been able to transform in such a good way as well.And that's where this [M20] program came in.It's all about implementing new technical conditions for modern marketing programs.But also, it's about changing the mindset and culture.So it's a change project as well.
Overall we have been very much about 'gut feel.'It hasn't been like insight-driven where we test and measure.(. ..)I would put it like this, we are a traditional company, it's retail stores.I think that something will of course happen in the next few years, new business models (. ..) that can shake the whole market.We must be prepared for changes.The problem of [SIRF] is that we are quite sluggish today.I think with this [M20], maybe we can be a bit more agile.
Although by SIRF standards M20 had ambitious objectives, due to its location in the organisational structure, innovation was limited within the realm of the marketing and IT functions.Within a year of implementation, the scope of M20's activities was ultimately considered to be insufficient in fully addressing the need for innovation in SIRF.Whilst M20 was continued as a digital marketing initiative, it was acknowledged by senior management that a new unit with a stronger innovation mandate was needed and which would go beyond the digital marketing projects and attempts at agile working that M20 had introduced.
What I think is a shame is that, once again, for innovation, we have defined it to cover only marketing.But there are a lot of elements in it, there is so much more, digitalisation is so much greater than the marketing.

Towards radical innovation: a corporate innovation hub
Because of the acknowledgement from senior management of the lack of radical innovation efforts within the organisation, a corporate innovation hub (I30) was eventually established.I30 was created as a function that was structurally differentiated from the main organisation.It was also an attempt to address the need for a more structured and systematic approach for innovation in the company.Previously there had been no opportunity to develop and test novel, cross-functional ideas because an innovation function of this scale had not yet existed in SIRF.Although M20 existed, as a digital marketing initiative it fundamentally had a limited application area and its innovation process was function specific.
With the corporate innovation hub, the vision was to be able to harness the strengths of both a large corporation and that of a start-up in order to fulfil the purpose of improving SIRF's ability to use digital technologies in enhancing the customer offerings in and between the various companies of the parent organisation, the SIRF Group (including, e.g.financial services, real estate connected to the core retail business).Although the stated mission of I30 was "to identify and analyse new business opportunities and digital innovations at a high pace and turn them into new or updated customer offers," the informal directive was to support the different strategic priorities of the SIRF Group's companies by doing what the parent organisation would otherwise not be able to do, i.e. beyond what is done in on-going business operations.
But we found out that there is no one else, or not so many others, who are looking at 2035 questions.And somebody has to do if [SIRF] is going to be successful for 100 years, I think.(. ..)And I have also realised that I had it wrong before.Because in the past I have said that a dilemma we have is that we have difficulty balancing ideas with high innovativeness and low innovativeness, simple ideas and difficult ideas.But I realise, we have no difficult ideas right now.So that's our problem at the moment.That's something we have to look at because I'm not sure that the existing project processes will lead to these.
As I30 was considered as the Group's de facto incubator, after an innovation project is launched, it was to be integrated into a business unit or department in one of the companies of the SIRF Group.A key issue for the newly established I30 was how to further develop their own innovation development process as well as how to collaborate across the Group.The I30 team was convinced that existing project structures and processes in the organisation were not capable of seizing future opportunities; thus, they developed and iterated their own innovation process.
First we have ideas.When we have chosen some of them, we then test them, prototype them and then we build them and launch, and then scale up.That is the intended model.(. ..)What I was most surprised about was that it's okay with everyone that we have fuzzy frameworks . . .everyone is calm with it and trust us.So I have, to some extent, realized that it's okay that it's a bit fuzzy because it's a bit the name of the game.
I30's activities were formally divided into research, development and partnerships and investments.Research projects involve collecting customer insights but also scanning for developments in digital technologies.This was a way to address the need for more insight-driven ideation and long-term perspective research regarding innovation and digitalisation.Development involves creation and testing of various innovations, including business models and technical services, on a project basis.The aim was to quickly create prototypes that could be validated directly with employees and customers.Lastly, I30 was also responsible for exploring strategic collaborations, potential partnerships and investments to jointly drive development by looking into external sources of knowledge and competence, e.g.technology companies and startups.As an independent unit, I30 has its own budget for research and development projects and it has a steering committee composed of the CEOs of the various operating companies under the SIRF Group, demonstrating support on a senior management level.
It was the top management team, the CEOs of the various companies that together decided put money into [I30].It was not completely clear at the start what it was for, but everyone knew that we need to get faster and we need innovation.
In terms of internal collaboration, drawing from learnings from M20, emphasis was placed on the importance of cross-functional collaboration and transparency with the rest of the organisation.One of the ways with which this was done is through a knowledge sharing

Retail innovation and ambidexterity
"innovation forum" involving managers from various companies of the SIRF Group and the I30 team, wherein potential digitalisation-related synergies across the Group's different business areas were discussed.Furthermore, I30 established a continuous "exchange programme," a form of employee rotation to facilitate knowledge exchange with other parts of the organisation.These employees, usually a project lead or manager, were chosen based on their competence, the focus of the innovation projects being developed and what department or business unit should implement the project after launch.This way, shared responsibility and ownership of the innovation project was promoted.

4.3
Organising for innovation: an on-going, iterative process I30 has continued to grow and reframe its strategy, structure and activities through an ongoing, iterative process.It remains a joint innovation hub of the different business units of the SIRF Group (i.e.including financial services, real estate, etc.) although most of the projects (75%) involve the core retail business.Whilst I30 has completed and handed over innovation projects for launch, expanding the boundaries of exploration in SIRF proved to be a constant challengethe existing needs of short-term improvements for the core business still being more prioritised.Whilst I30's research and development projects included potential new business areas not directly connected to the core business, the scope was still not in the realm of disruptive technologies and new markets.Nonetheless, as the very role and identity of SIRF as a retailer continues to be reflected on by the team in charge of long-term innovation, I30 have maintained its focus on research projects, building in-depth knowledge on selected themes including specific retail technologies and wider societal trends.
It is also about what self-image one has.Are we only 'retailers' who buy and sell?Or are we a tech company?Or are we a platform?Depending on how we should consider ourselves, innovation is different.
One of the main challenges that the I30 team faced relates to innovation performance measurement, choosing the relevant objectives and key performance indicators for their work.In M20, the application areas of the innovations were closer to on-going operations (digital marketing) and thus familiar metrics were used such as measures of brand equity or established project management metrics.As I30's activities involved broader, more novel and complex innovation projects with higher uncertainty that did not follow the traditional development trajectory in SIRF, the established metrics for traditional business development and product development projects were unsuitable.Balancing the need for clear directives and output measures on the one hand and the freedom to act with a stronger innovation mandate on the other is a difficult endeavour and the I30 team has been experimenting with different types of innovation metrics.Formal objectives and metrics were acknowledged as an important way to ensure that the independent innovation hub will still be coupled with the rest of the organisation.
Good innovation metrics, we do not really know yet what this would be for us.We have reached something small, we measure input, throughput, output in this process.How many ideas can we put in?How fast is it going?And what comes out?Number of fails, we have tried to measure.It turned out quite difficult because there have not been so many fails but rather many pivots.Is it failure then, or is it just a change of course?We do not really know yet.These KPIs-we have seesawed a bit on how we should do it.
5. Organising for innovation through structural and contextual mechanisms: insights from SIRF SIRF's initiatives to enable innovation through contextual and structural mechanisms can be viewed as an approach to enable more exploration by an organisation that has been inherently exploitative by nature.As a leading retail chain, incumbent retailer SIRF may have been detached from the concept of innovation a few years ago, but this is not the case today as demonstrated by a more active approach to innovation through the M20 and I30 initiatives.Based on insights from the case study as well as relevant literature, we present the following proposition that describes how incumbent retailers can organise for innovation from an organisational ambidexterity perspective, depending on the type of innovation that is desired: Contextual ambidexterity would more likely lead to incremental, operational retail innovation, while structural ambidexterity would more likely lead to radical, strategic retail innovation.
Contextual ambidexterity is characterised by employees dividing their time between exploration and exploitation on a day-to-day basis in a high-performance organisational context that reconciles forces of discipline, stretch, support and trust (Ghoshal and Bartlett, 1994).It utilises a more behavioural and social means to integrate exploitation and exploration (Andriopoulos and Lewis, 2009).SIRF's implementation of M20an initiative that was the first of its kind in the companyexemplifies a retailer's attempt towards contextual ambidexterity given its ambitious innovation goals and the introduction of experimental and innovative behaviours through agile methods.Exploration-oriented activities were not assigned to a different unit but rather integrated with exploitation-oriented activities under the existing functions of marketing and IT.However, this also meant that M20 was constrained by a limited innovation focus, having defined the boundary of projects within digital marketing.Regardless of application area, i.e. whether the projects were customer-related, support-related, or organisational-related, the innovations in M20 were more operational than strategic in terms of impact.
Given the inherently limited formalisation of innovation processes in most retail organisations, our findings reveal that organising for structural ambidexterity is needed when the aim is radical, strategical retail innovation.Structural differentiation, through the creation of innovation hub I30, was deemed to be crucial because without an independent function with the sole responsibility of long-term innovation work, it was too difficult for the incumbent retailer to escape the inertia of existing business units towards efficiency.The newly established corporate innovation hub thus came with a different strategic intent and a more separated organisational structure.Organising for radical, strategic innovation in retail is thus in line with suggestions in literature emphasising the need for separate structures within an organisation to address the opposing intents and subsequent competences, systems and practices of exploitation and exploration (Duncan, 1976;Benner and Tushman, 2003;O'Reilly and Tushman, 2004).This key finding challenges the existing notion that innovation in retailing often is a product of informal processes that are not institutionalised.The study provides a contemporary understanding of how innovation efforts in retail are increasingly being organised.
With the I30 team having pushed the boundaries for its mandate to cover new markets and technologies, the structural separation from the core business was essential in safeguarding the space for radical innovation.However, finding the appropriate distance and balance was fundamental to its success.The case study, therefore, confirms the critical role of integration of exploration to the mainstream, exploitative side of the organisation.Previous research has suggested that without the subsequent integration with retail operations, a structurally separated innovation function may end up as another functional silo that would not solve the innovation-related aims of any retail company (Olsson and Johansson, 2016).Enabling exploration and exploitation within the same organisation through structural differentiation alone runs the risk of creating an unproductive and inefficient endeavour, unless integrative mechanisms also exist (Koryak et al., 2018;Jansen et al., 2009;Mom et al., 2009;Lubatkin et al., 2006;Smith and Tushman, 2005).In line with suggestions by, e.g.Jansen et al. (2009), our findings demonstrate how informal and formal integration mechanisms were instrumental in ensuring that the work of I30 would continue.Maintaining cross-functional interfaces in middle management, for instance, served this purpose.The knowledge sharing "Innovation Forum" together with the employee exchange programme were formal mechanisms that aimed to foster the linkages between the main organisation and I30.

Retail innovation and ambidexterity
The exchange programme is an example of how liaison personnel can play an important role in resolving differences across the exploratory and exploitative units (Jansen et al., 2009).
Ultimately, integrative mechanisms at strategic and operational levels are made possible when there is senior management integration.Senior executives play an important role in fostering ambidexterity because they are responsible for putting in place systems that allow supportive contexts to emerge, thus encouraging and nurturing adaptability (Gibson and Birkinshaw, 2004).As the case study demonstrates, the establishment of I30 would not have been possible without strong support from senior management.He and Wong (2004) have emphasised the importance of senior managers' awareness of the need to allocate resources between explorative versus exploitative activities.Senior management permitted exploration by allotting resources for a corporate innovation hub even when it was not well-established at the beginning what its deliverables would be, allowing for experimentation in how the I30 team would organise their activities and process.Whilst the vision and strategic intent for innovation proved to be initially limited through M20, I30 later on was equipped with the necessary resources with the approval of a steering committee, composed of the CEOs of the various companies in the SIRF Group, indicating a clear commitment and senior-level buy-in to a differentiated business unit solely for long-term, strategic innovation.
A final insight from the case study was the need for better innovation metrics as a formal mechanism for integration.Whilst there is no lack of performance indicators in retail, quantitative indicators for productivity and profitability are often prioritised rather than measures for milestones and growth.Finding the appropriate metrics for radical innovation in retail proved to be a major issue encountered in setting up I30, as opposed to the M20 initiative where established retail metrics were useful when the projects involved incremental, operational innovation.Whilst the difficulty of finding mechanisms for measuring innovation was acknowledged in practice, the ambidexterity and innovation management literature give limited guidance and is an area of particular interest for future research (Gschwantner and Hiebl, 2016).We suggest that innovation metrics in the context of retailing is a topic that warrants further investigation.

Concluding remarks
Integrating concepts from retail innovation and organisational ambidexterity, this explorative research presents a single-case study on how an incumbent retail firm attempts to actively organise for innovation.When the strategic intent is for innovations that are more strategic and radical in nature, retailers would more likely benefit from pursuing structural ambidexterity.Structural mechanisms can safeguard the space and resources to focus on long-term research and projects with higher risk and uncertainty; however, integration of the innovation activities to the mainstream organisation is critical.On the other hand, pursuing contextual ambidexterity wherein employees are empowered to divide their time between innovation-related and efficiency-related tasks would more likely lead to retail innovations that are incremental and operational.
This study shows that whilst the strategic management of innovation might not have had a clear relevance in retail in the past, this is changing as demonstrated by the establishment of SIRF's innovation initiatives.The paper adds new evidence to the emerging topic of strategic management of innovation in retailing, explicating how incumbent retail organisations can organise for innovation depending on the type of innovation that is needed, using organisational ambidexterity as a new perspective.
Whilst this study offers contributions to research and practice, there are certain limitations that should be considered.This paper is an explorative study that uses a single-case study of an incumbent retail firm in a mature economy.As such, caution should be taken in generalising findings and applying them to other types of retail organisations and contexts.Further research could be carried out in the future to develop the proposition presented in this paper by extending the study into a larger, multiple case study involving different retail actors.Furthermore, innovation metrics as a means for integration proved to be one of the on-going challenges encountered, which is a topic worth exploring in future studies, given the limited understanding on this subject.
Data collection methodDetails (participants, involvement in M20/I20, date)Semi-structured, individual interviews 8 interviews with managers involved in M20 and I30(March 2017   to July 2018) • Digital Experience Manager, Marketing • Project Manager, IT • Manager, Human Resources • Project Manager, Marketing • Manager, IT and Operations • Manager, Customer Relationship Management • Head of I30 (2 interviews) 1 follow-up interview (June 2021) • Head of I30 Participant observations and unstructured interviews 9 meetings with managers involved in M20 and I30 (December 2016 to February 2018) 2 workshops with senior management and I20 team (February and April 2018) 2 workshops with I20 team and external consultants (May and June 2018)

Table 3 .
Details of data collection