me

anserg Plad

Management accounting calculations relate innovation to the rm through translationswhere both can change. Based on examples of the management of innovation from threerms the study shows how management accounting calculations rather than describe

see them as inscriptions that produce knowledge (Robson,1992), create visibility (Cooper, 1992), mediate betweencomplementary resources (Miller & OLeary, 2007), andidentify objects and objectives to be managed (Chua,

calculations are not only mobilised by others they alsomobilise others. In this study, this means that accountingcalculations create contexts for something, and in this re-search this something is innovation. The research questionis: how do management accounting calculations mobiliseinnovation activities?

The central nding, which is based on the empiricalstudy of relations between management accounting

0361-3682/$ - see front matter 2009 Elsevier Ltd. All rights reserved.

* Corresponding author.E-mail addresses: jm.om@cbs.dk (J. Mouritsen), ah.om@cbs.dk

(A. Hansen), ch.om@cbs.dk (C.rts Hansen).

Accounting, Organizations and Society 34 (2009) 738754

Contents lists available at ScienceDirect

Accounting, Organiza

w.edoi:10.1016/j.aos.2009.01.006Management accounting calculations relate innovationactivity to the rm through two types of translations; ashort translation which helps extend or reduce innovationactivities in view of an actual or a possible performancevariance; or a long translation which develops competingcontexts for innovation and impacts rms innovationstrategies and sourcing arrangements. This conclusion,which will be developed and justied later, adds weightto theories of management accounting calculations which

per, & Coombs, 1992; Vaivio, 1999). Management account-ing calculations are related to organisational practiceseither in relation to individual managers localised, embed-ded decision making (e.g., Boland & Pondy, 1983; Ahrens &Chapman, 2004,2007), or in relation to change programsthat reach deep into the organisation to manage the labourforce and transform the rm (e.g., Ezzamel, Willmott, &Worthington, 2004; Ezzamel, Willmott, & Worthington,2008; Miller & OLeary, 1994). We follow these ideas butadd one nuance suggesting that management accountingIntroductionthe properties of innovation add perspective to them mediating between innovation con-cerns and rm-wide concerns. This mediation happens through short and long translations.In short translations, management accounting calculations extend or reduce innovationactivities via a single calculation. In long translations innovation activities are problema-tised via multiple calculations. When calculations challenge each other in long translationsthey problematise not only what innovation should be, but also where it should be locatedin time and space. In the three examples, calculations mobilised alternative propositionsabout the relevance of technical artefacts and linked this to innovation strategy and sourc-ing strategy in the rms inter-organisational relations. Tensions between calculationsassociated with technological, organisational and environmental entities framed consider-ations about the value of innovation to the rm strategically differently. All this happensbecause management accounting calculations are partial rather than total calculations ofrms affairs and value.

2009 Elsevier Ltd. All rights reserved.

1995; Hoskin & Macve, 1986; Miller, 2001; Preston, Coo-Short and long translations: Manageinnovation management

Jan Mouritsen *, Allan Hansen, Carsten rts HDepartment for Operations Management, Copenhagen Business School, Solbje

a r t i c l e i n f o a b s t r a c t

journal homepage: wwnt accounting calculations and

ns 3, DK 2000 Frederiksberg, Denmark

tions and Society

lsev ier .com/ locate /aos

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 739calculations and innovation in three rms, is that manage-ment accounting calculations link innovation activities torm-wide concerns rather than describe and representinnovation activities. The visibility, insight and knowledgeproduced by management accounting calculations rarelyconcern the details of innovation practices. It rarely createsdeeper knowledge about the intricacies of innovationactivities; it typically creates insight about links betweeninnovation and wider organisational concerns which aremediated via short or long translations, where length re-ects the number of elements taken into account. In shorttranslations innovation activities are mobilised by a singlecalculation and related to a variance from a standard orbudget which will reduce or increase innovation activitiesdepending on whether the deviation is positive or nega-tive. Short translations mediate between innovation activ-ity and the costs and revenues of the rm.

Long translations have multiple calculations that createtensions about the role of innovation. Here, calculationschallenge each other and develop organisational tensionsand dialogues beyond innovation activities. Long transla-tions develop new possible versions not only of preferredtypes of innovation activities, but also about their locationin time and space. They develop competing propositionsabout the relevance of technical artefacts and link themto innovation strategy and sourcing strategy in the rmsinter-organisational relations. The tensions within longtranslations mobilise technological, organisational andenvironmental entities by framing considerations aboutthe value of innovation to the rm strategically differently.

The remainder of this paper is structured as follows:rst we analyse central discussions about the role ofaccounting calculations in innovation. Here, accountingcalculations are typically not accorded a constructive role,but an emerging literature suggests a positive link betweenmanagement accounting calculations and innovation nd-ing that management accounting calculations are abun-dant in innovative contexts. Yet, the literature is silent onhow the calculation inuences elements of innovation.Then the research strategy and methods are presented;drawing on aspects of actor-network theory we trace rela-tions between proposed management accounting calcula-tions and innovation activities. The empirical sectionpresents three examples of translations between manage-ment accounting calculations and innovation manage-ment. Then the ndings are discussed and nallyconclusions are provided.

Management accounting calculations and innovationmanagement

Often, management accounting calculations and associ-atedmanagement control systems have been understood tohinder the development of innovation. The innovationmanagement literature usually denies a constructive inu-ence of management control systems on product innova-tion (Damanpour, 1991; Dougherty & Hardy, 1996;Gerwin & Kolodny, 1992; Leonard-Barton, 1995; Tidd, Bes-sant, & Pavitt, 1997; Verona, 1999). Formal control systemsconstrain, or at best are irrelevant in, innovation and R&Dsettings (Abernethy & Brownell, 1997; Birnberg, 1988;Brownell, 1985; Hayes, 1977; Rockness & Shields, 1984;Rockness & Shields, 1988). They are obstacles to creativityand incapable of supporting innovation (Abernethy &Stoelwinder, 1991; Amabile, Conti, Coon, Lasenby, &Herron, 1996; Miles & Snow, 1978; Ouchi, 1977; Ouchi,1979; Tushman & OReilly, 1997). Rationalisation is seenas incompatible with the creativity required for innovation(Burns & Stalker, 1961; Hall, 2001; Raelin, 1985).

However, increasingly it is proposed that managementcontrol systems enable innovation (Clark & Fujimoto,1991; Cooper & Kleinschmidt, 1987; Cooper & Slagmulder,2004; Davila, 2000; Davila & Wouters, 2004; Hansen &Jnsson, 2005; Ittner & Kogut, 1995; Ziger & Maidique,1990). Management control systems can be enabling forcorporate activities (Ahrens & Chapman, 2004, 2007), andSimons levers of control framework (1987, 1990, 1991,1994, 1995) suggests that interactive use of managementcontrol systems stimulates innovation (Bisbe & Otley,2004; Widener, 2007). Here, formal management controlsystems can under certain circumstances help rmsfacing rapidly changing product or market conditions. Forexample, Simons (1990, p. 141) suggests that

the prototypical prospector faces strategic uncertain-ties owing to rapidly changing product or market condi-tions; interactive management control systems such asplanning and budgeting are used to set agendas todebate strategy and action plans in these rapidly chang-ing conditions. Defenders, by contrast, use planning andbudgeting less intensively [because they] operate in arelatively stable environment, many aspects of the busi-ness that are important in terms of current competitiveadvantage are highly controllable and managers needonly focus on strategic uncertainties often related toproduct or technological changes that could underminecurrent low cost positions.

When environments are complex and dynamic rmshave management control systems which foster dialogueand interaction about the development of products andmarkets and the innovative pressure may be accommo-dated via interactive use of management control system(Bisbe & Otley, 2004).

Likewise, Davila (2000, p. 402) identies uncertaintyand product strategy as drivers of management controlsystems in new product development and he adds that abroad denition of management control systems is neces-sary to understand their role in relation to product devel-opment (ibid., p. 404):

The study reinforces a broader denition of manage-ment control systems to go beyond nancial measuresand also include non-nancial measures. . . This ndingsuggests that researching management control systemsin new product development cannot be restricted totraditional accounting measures, but needs to encom-pass a broader set of measures. . . As the theory pre-dicted, uncertainty and product strategy are related tothe design and use of management control systems.

Depending on the type of uncertainty facing managersthey will use different combinations of nancial andnon-nancial information. Like Simons, Davila emphasises

decisions about innovation activities can be usefully ex-

also all produced measurement technologies and systems

740 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754tended. How does a calculation make a difference?Robson (1992) argues that accounting calculations de-

velop visibility and create organisational time and space.He analyses how accounting mobilises distant places andmake themparts ofmanagersworld.Management account-ing calculations provide a good deal of the knowledge that isavailable for management (Cooper, 1992, 1997; Law, 1996).This knowledge is an effect of procedures of inscription, i.e.,procedures of how traces such as receipts and statistics areput together and ends in a calculation (e.g., Briers & Chua,2001; Chua, 1995; Miller & Rose, 1990). Focusing more onthe procedure ofmaking a calculation than on its correspon-dence with an underlying reality, Robson makes the man-agement accounting calculation one proposition about thenancial affairs of the rm. So, organisation and marketmay be brought forward and made visible by calculationsof, e.g., of revenues and development in protability (Hines,1988; Quattrone & Hopper, 2005), and the calculations im-pose an agenda requiring a response (Miller, 2001). Theseauthors emphasise that a management accounting calcula-tion is an inscription which develops visibility by

stating what belongs to the past, and of what thefuture consists, by dening what comes before andwhat comes after, by building up balanced sheets, bydrawing up chronologies, it imposes its own spaceand time. It denes space and its organisation, sizesand their measures, values and standards, the stakesand rules of the game (Callon & Latour, 1981, p. 286).

By making things visible, the calculation prioritiseselements to be accounted for. Calculations inuencehow different spaces and different times may be pro-duced inside the networks built to mobilise, cumulateand recombine the world (Latour, 1987). The calculationis an actor. According to Latour any thing that modif[ies]a state of affairs by making a difference is an actor (La-tour, 2005, p. 71). No actor acts alone therefore the calcu-lation is always part of a larger collective that actstogether with it. Actors are made to act by many others(Latour, 2005, p. 46).

Approach and research strategy

The empirical domain is three small and medium sizedcompanies. We interviewed 2025 managers in each rmeach taking between 1.5 and 3 h. We explained managersthat we were interested in their efforts to control and ac-count for innovation. We had a semi-structured question-naire, but often the dialogue would quickly develop itsown momentum. We did not focus on the rms as ethno-characteristics of the situation as drivers of managementaccounting calculations. Simons and Davila forcefully arguethat management accounting calculations do not hinderinnovation. Indeed, they suggest that in innovative contextthere may be many more calculations than in situationswhere innovation is less prevalent. They demonstrate thatmany calculations exist. Yet the analysis of how a controlagenda, such as interactive use of calculations or combina-tion of nancial and non-nancial information, inuencesused in different industries but there were commonalitiesin product technologies (such as a mechanism to receiveand record signals, a computer to manage the signals anda screen to present the signals in a relevant form). Eachhas been given a ctional name to preserve their anonym-ity: SuitTech, HighTech and LeanTech. Through the analysisit was possible to draw out two propositions about innova-tion and two associated management accounting calcula-tions in each example.

The analysis of the empirical material was organised toidentify translations between calculations and innovationactivities. Firstly, we identied propositions about causalrelationships between innovation and value creation med-iated by calculations.We paid attention to how calculationswere accorded power to do things. Secondly, we noted howthe power attributed to calculations translated into pro-posed effects on management of innovation activities(reduction or extension of innovation activities). We tracedhow a presentation of a calculation would propose to inu-ence innovation activities. Thirdly, we then paid attentionto the time and space suggested to be informed by the cal-culation and noted how changes in innovation activitieswould transform into something else such as sourcingstrategies which turned out to be surprisingly important.Last, we used Callons (1986) diagrammatic form to illus-trate the movements around the calculations. His diagramsshow how entities are included in or excluded from anexplanation and they seek to identify the movement ofchanging relations. Figs. 14, which will be presented later,are outcomes of this analytical procedure.

Translations between management accountingcalculations and innovation activities

The empiricalmaterialwas collected in three rms that allinvested in innovation and made this a priority. The concernwasnotwhether innovationwasuseful, butwhich innovationshould be conducted and how it should be organised. In allrms there were manymanagement accounting calculationsbut not all were able to stand for or represent innovation. Ineach of the rms certain calculations were accorded particu-lar signicancewhenmanagers accounted for innovationper-formance. The following sections present how managementaccounting calculations were mobilised to account for andinuence innovation activities.graphic (or cultural) entities, as Yin (1994) would recom-mend, but rather on episodes of translation betweenmanagement accounting calculations and concerns fortechnology. Our interviews were reexive (Alvesson,2003) or analytical (Kreiner & Mouritsen, 2005) whichacknowledges that our theoretical issues, which were pre-sented to mangers explicitly, were the introduction to datacollection. This is not a claim to have researched three rmsin their totalities; the claim is to have researched howman-agement calculations are related to decisions about innova-tion (technology). Management accounting calculations arelikely used for many other purposes as well.

The three rms not only claimed to be innovative andcould all be characterised as HighTech companies. They

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 741Example 1: SuitTech the role of special and customisedcomponents in innovation

SuitTech, a small HighTech rm, produced and soldmeasurement systems to R&D departments and universitylaboratories whose measurement problems varied consid-erably. Some customers measured turbulence in wind tun-nels; others measured water-currents when designing oil-rigs, and yet other customers measured turbulence inames. These different measurement situations confrontedSuitTech with demand for product innovation. Its missionstatement emphasised its ability in providing solutionsand solving problems, and it singled out that customershave depended on the quality and reliability of its prod-ucts and services to solve their problems.

SuitTechs measurement systems were presented as un-ique offerings. Each product was bent tightly around theindividual customer with extreme customisation. In orderto make a unique solution with precisely customised tech-nical functionality, sales engineers could, in cooperationwith the customer, choose from special and customisedcomponents delivered by a broad range of suppliers ordeveloped and produced by SuitTech itself. Finding specialand customised components along with developing andproducing unique components internally was suggestedto be a core competence of the rm.

Mobilisation of sales performance and innovation throughspecialised and customised components

To sales engineers, sales performance was an authori-tative performance measure. The measure calculated theactual gross revenue minus budgeted gross revenue foreach of the major technological areas quarterly. The bud-get was set between the teams of engineers, the salesmanager and the CFO of the rm. Actual gross revenuewas an accumulated measures of all orders signed for atgiven technological area in a given quarter. Thus, salesperformance was recognised in SuitTechs accounting sys-tem when customers signed a contract and an order wasmade. Before signing the contract, customers and salesengineers had a long and intensive dialogue about cus-tomer needs and technical characteristics; they developedmany different propositions about the measurementproblem at hand and about its targeted performance.Therefore, an order symbolised the end of a prolongedprocess of interaction where numerous propositions weredened and considered; the characteristics of an eventualorder could not be predicted at the outset of the processand it was therefore its effect rather than its precondition.The calculation, sales performance, illustrated preciselythat a long process had been ended, which was observedby a sales engineer:

You see the results of what we do in the sales mea-sures. A customer never makes an order before we havehad serious discussions with him or her about the mea-surement problem. And unless we can come up withsomething convincing, we do not get the order.

Sales performance marked the end of a process of inter-action. Together, sales engineers and customers assembledthe measurement system according to detailed require-ments and specications which were developed as partof the process. In principle they could choose any combina-tion of components such as optical receivers, lenses, chass-es, lasers, etc. These could be sourced from a large networkof carefully selected suppliers. The sheer number of possi-ble different components allowed huge exibility in de-sign, and made innovative solutions to the customersmeasurement problems possible:

We can easily be in situations where we need a1.3 mm lens instead of a 1 mm lens. If we let foregothe option to choose from many different items in thedesign (and only use internally produced components)I think SuitTech will create bad customer solutionsand thereby loose competitiveness.

Supplies of external components were used to renethe customers solution and allowed SuitTech to be andstay innovative. In SuitTech, innovation was negotiatedprincipally between sales engineer and customer andwhen needed with the suppliers of special components.Both were professionals and both knew the intricacies ofthe technology. The process of selling, which involvedinventing the product, was time consuming. In principle,it could go on for a long time because both sales engineerand customer would always be able to invent or thinkabout new improved details. Therefore, the process ofdeveloping an order was inspired and would not neces-sarily stop: more timemeant more detail and more quality.

How could such a process be stopped and transformedinto an order? When sales budgets were met and aspira-tions achieved, the sales variance was modest and typicallyunconnected to the process of developing and closingorders. However, in situations where such aspirations werenot met, the sales variance transformed the network ofactivities performed by sales engineers. Unfavourablevariance inuenced sales engineers to redirect their effortsfrom developing orders to closing orders within a shortperiod of time and they were thus persuaded to bracketconcerns about the products. Unfavourable variance ori-ented them to cash ows away from leads; to budget-vari-ances rather than to customisation; and to closing ordersmore than to creating new and elegant combinations ofspecialised components. Unfavourable variance recastsales engineers interests and problematised the dilemmasbetween SuitTechs and customers needs. The sales budgetproblematised the interests of the rm compared withthose of customers and suppliers. Sales performance cre-ated the tension between customisation and closing or-ders. It dened a strategic uncertainty about theinnovation agenda in SuitTech. When sales performancewas favourable it extended technological innovation whilewhen unfavourable it reduced technological innovation.

Extending translations of innovation mobilising direct costsInnovation was in many ways predicated on expansion

of the number of possible components that could be putinto a product. Sales performance framed sales engineersexperimentation with complex designs that prolongedthe sales process as only the best was tolerable. Itprevented much nancial problematisation of the rmsinnovation. A business controller noted the inferiority of

742 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754cost in accounting for the rms sales performance inSuitTech:

A performance measure that is very important for oursales engineers is sales. What I as a managementaccountant miss are indicators for direct cost. We quiteoften debate this. I think this omission to a large extentcomes from the way we innovate. The focus onconstructing unique measurement systems to theindividual customer and producing to order make costindicators less relevant... but I think that we shouldstart considering these things as well. It is possible tobe aware of direct costs even if we are a bunch ofinnovators.

This addition to sales performance of cost items devel-oped a new type of tension in relation to the value of inno-vation. The business controller contended:

It is the contribution margin and not sales that matterswhen it comes to value creation. As a managementaccountant I would say that it is a much more represen-tative calculation of sales engineers value creation.

The contribution margin made revenues less direct costvisible. Such inclusion of cost in performance was pro-posed as a more relevant concern with value creation,but it was also challenged. The sales manager explained:

As soon as we start to use contributionmargin as a per-formance measure some would probably be tempted bythe fact that they could increase performance by reduc-ing direct costs. That is probably good in some situationbut I think that many engineers would probably alsostart to apply cheaper components and new and lessefcient technology in order to reduce the costs whichwould be a disaster for us. We do not compete on costs.We compete on the solution that we are able to come upwith for the customer!We sell a differentiated product a solution that the costumer is willing to pay for. Weshould not be spending our time on reducing costs butinstead on nding the right solution.

Sales performance motivated a strategy of tight cus-tomisation through liberal use of externally sourced specialand customised components but lurking closely in thebackground was the proposition to reduce direct costs;through such behaviour a whole new technology strategythat included a focus more on programmable standardcomponents and software would become desirable. Adapt-able software programming and a narrower range of stan-dard components presented an alternative to the largevariety of special components. Programmable componentdevelopment, which was an appendix to sales and notobligatory to sales engineers, was used to create a bench-mark for technology. The strong form of customer orienta-tion did not favour conventional forms of planning andcontrol. The production manager emphasised that

Actual costs are always different from forecasts; inparticular direct costs depend upon specic measure-ment problems that the customer has and these arehard to forecast and there are no incentives to reducethem for the sales engineers.Thus, the commitment to customisation challenged con-trol of direct cost as well as delivery time since the supplysituation often became complex and impossible to forecastdue to the use of specialised items sourced from externalsuppliers. This concern was, however, only loosely coupledto SuitTechs strategies as delivery time was proposed notto be crucial to the customer.

As calculation, sales performance did not consider directcosts. It did not propose standardisation and it did notstress technological predictability and stability. It framedthe economics of the rm in relation to innovation activitiesbut it did not specify how innovation activities should beorganised because its focuswasmore external than internalto innovation activities. Sales performance motivatedexpansion of activities and propositions in innovation. Asales engineer commented:

We are free to choose any special or customised com-ponent that fulls the customers need. Of course thecustomer has to pay for it but we do not keep recordand set targets for these things. Reducing direct costsis not a performance criterion. Actually, it is a bit of arelief and it makes our job easier. It creates room forinnovation. You may say that it is critical to oursuccess.

Tensions related to the omission of direct cost in theperformance measure was raised by controller whoclaimed that sales engineers should mind costs and reducethe use of the special and customised components:

I do not want to be a pessimist. I think the sales engi-neers do a great job. But is it more the fact that theyshould keep in mind that the special components costsus actually quite a bit in terms of direct costs and time.So why dont we start to incorporate it in our perfor-mance measure.

If they had knowledge of direct cost sales engineerswould perform innovation in new ways and ask questionsabout the appropriateness of special and customised com-ponents. They would reduce the use of such componentsand substitute them with programmable standard compo-nents. The production manager explained:

There are alternatives to special components. I mean,we can go far by programmable standard componentsand by the help of software programming from our soft-ware engineers. Programmable components can neverreplace special components totally but this is anotherpossible technological strategy.

Such a strategy would also affect supplier-relations theproduction manager suggested:

This would also imply that we have to think about oursuppliers in a different way. Currently, we spend a lot ofresources nursing the large network of suppliers deliv-ering customised and special components. However, ifwe used programmable standard components, wewould reduce this network and the resources we con-sume in the purchasing department signicantly. It isa strategic cost, but remember the special and custom-ised components are benecial to us in many ways.

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 743Sales performance privileged heterogeneity in compo-nent selection. The visibility created by calculating costsof special and customised components would encouragea wholly different strategy for innovation. The alternativewould be to focus more on the components programmedby SuitTech itself where variation was created by softwarerather than by hardware:

We might challenge the way that we innovate today.In fact, software is an alternative to the hardware deliv-ered by suppliers.

The tension between the two strategies was to a largeextent created by the demarcation between performanceaccording to sales and direct costs. Direct costs problema-tised the use of special and customised components andproposed to inuence inter-organisational relations.

If we focus more on the components that we can pro-gramme ourselves we might change the way that weare innovative today. This would also affect the waywe see our suppliers. They would rather deliver a rela-tively limited number of standard components. Now weconsider them all as one big supermarket. Lots of oppor-tunities exist out there.

The perspective suggested by direct costs related newelements to the translation of innovation. It required Suit-Tech to upgrade its internal software competences to con-vince sales engineers about the real relevance ofstandardised programmable components for customisa-tion. This challenge was mobilised by associations madeby direct costs and contribution margin which were instark contrast to the ideas of components and inter-organ-isational relationships made by sales performance.

Example 2: HighTech: the concern with technologicalsuperiority

HighTech produced and sold measurement systemstypically to the health sector (e.g., hospitals). Like SuitTech,also HighTechs customers demanded high technology butthey shared industry where the measurement system hadto perform various but specic kinds of medico-technicalanalyses. HighTechs innovation aimed to develop prod-ucts ability to perform all relevant medico-technical anal-yses. Technology development pushed the boundaries ofsupplied technology to the point where HighTech knewmore about possible measurement tasks than customersor users would normally do. HighTech saw itself as a mar-ket-driving rm where customers would buy latest tech-nology when it was made available to them.

Mobilisations of contribution margins and innovation throughtechnological superiority

The product contribution margin was standard vocabu-lary in the new product development department of High-Tech. The contribution margin subtracted expected directcosts from expected sales and the targets set for directcosts as well as sales prices became a measure that coordi-nated and motivated actions taken in each developmentproject. The performance measure, however, paid littleattention to indirect costs which was suggested to havecreated a signicant room for innovation. A developmentengineer explained:

There is not much focus on indirect costs in ourresearch projects and this is fortunate because it givesus freedom to experiment. We are not as accountablefor the resources we spend on each project as wecould be. Before I came to HighTech I worked in adevelopment organisation where this was alwayswas an issue. Here, there are many more possibilities and I think it is benecial for the organisation as awhole.

The development engineer referred to a concern inHighTech whether product development project managerswere to be accountable for the indirect costs of the R&Ddepartment carried out HighTech. The concern waswhether research resources should be reected in productprotability or not; would it be advisable to develop aprot margin after indirect cost or maintain the focus onthe contribution margin accounting primarily for indirectcost? Technological innovation was important to HighTechthat had a history of high quality products. It saw itself as amarket-maker that set the de facto standards of the indus-try. HighTech emphasised application of new technology.The director of research and development suggested thisvery clearly:

We must develop the technology. It makes no sense tous just to copy the products from our competitors. Ourmission is to develop the new products to the marketand we have to be the leading technological rm. Thisis what gives us prot.

HighTech was committed to R&D and prided itself to beable to see customer wants before customers were awareof them. Product developers proposed that they knewmore about relevant uses of the measurement system thancustomers and often customers simply accepted that High-Techs latest product had to have better solutions thanwhat the customer would be able to think of. The individ-ual product was not customised. It was standardised, butas HighTech continuously set new standards for what ameasurement system could do, it created its own demand.It was less a market-driven rm than a market-drivingrm, and HighTech experienced a high degree of technol-ogy elasticity which connected technology developmentwith high growth in prices and revenues. HighTech pro-posed its extensive investment in experimentation andR&D in their development projects as a reason for thiscapability.

HighTechs R&D organisation was separated in two: aR&D department and a development organisation. TheR&D department carried out technology projects aboutchemical uids and electronics and was presented as a ser-vice department for development projects. Technologyprojects initiated to solve technological issues in one newproduct development project could often produce knowl-edge that could be used in a wide range of other develop-ment projects. Individual technology projects produceddeep technological competences in chemical uids as wellas electronics and not merely applications hereof to a

744 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754product line. The costs of R&D were not allocated to newproduct development. The R&D director argued:

Many of the results we get from the technology pro-jects are like public goods. They can be shared byeveryone, as it is a key towards our key competitiveadvantages.

A new product development manager continued:

Often we take detours in the projects. It makes the pro-jects much more expensive in total. But the things thatwe learn provide us with the extra knowledge that is sodecisive to us if we want to keep our position on thetechnological edge. Some may say that we are too care-ful [in research] and spend too many resources in thedevelopment projects. But we learn things that we canuse later in other projects. It is a delicate balance. Butit is a thing that I think that we are good at inHighTech.

HighTechs innovation concerned learning in relation toits technological bases in chemical uids and electronics.Innovation was for purposes beyond the products at hand.The detours in technology projects created extra knowl-edge that could be used in later projects.

New products were considered to produce additionalrevenues which would by far outweigh additional directcosts. Development engineers raised the contribution mar-gin as a justication for complexity in product develop-ment. Even if direct cost was part of the contributingmargin and some concern had to be mustered to managethese costs, the contribution margin justied attention tocomplex organisational development capabilities:

We are allowed to develop our key technological capa-bilities: electronics and uid chemicals. And cost con-trol here is very difcult. But when it comes to directcost we all have a responsibility. Sometime we evenhave to compromise design in order to keep direct costlow. However, this is of less importance in regard to theinnovation lead we get from the development of ourkey technological capabilities.

Sometimes infrequently direct cost could compro-mise design but generally, product innovation was drivenby experimentation with new technologies and large in-house development projects. Concerns with efciency inproduction processes were in large part exported to sub-contractors, as suggested by the purchasing manager:

In our contracts we promise, e.g., to pay for a numberof spools but we will only cover the direct cost and notany prots. If we need less that the number of spoolswe only have to pay for the specic and direct cost ofthe items. So, the subcontractor does not suffer a directloss but neither does he gain any prot. For example,we do not pay for the copper-wire of the spool. It canbe used for other customers. We will only cover thespool.

Product development was concerned with revenues andproduction with cost. Inter-organisational relations mod-elled this difference.Extending translations of innovations mobilising indirectcost

From time to time frustration about the cost conscious-ness of the R&D department was aired. Controllers sug-gested that they start focusing on the resources thatproduct development project consumed in the R&Ddepartment. It seemed that product development projectsinitiated many activities and incurred signicant costs.One controller stated: It is as if you can get technologicaladvice for free.

One way to direct more attention towards the cost-con-sequences of technology development was to allocate thecosts of the technology projects of the R&D departmentto the new development projects of the developmentdepartment. Different types of cost drivers were suggested,e.g., number of requests made to the R&D department, orman hours in the R&D department traceable to individualnew product development projects. The requested labora-tory tests, experiments, etc. were central to solve the tech-nological problems that emerged during the new productdevelopment projects. This would make certain costs ofR&D visible for new product development managers whocould then economise R&D activities. This would haveimportant consequences as a controller argued:

To include a strict focus on indirect cost in our perfor-mance measure would be to introduce an entirely newidea about our business. Nevertheless, I think it is cru-cial that we do this.

Costing would problematise technology projects and newproduct development managers would ask questions aboutHighTechs knowledge banks and look for technological solu-tions elsewhere. A development engineer commented:

Currently, we do not use suppliers much when itcomes to our technology development. But it is de-nitely an option that we should consider in order tobecome more cost efcient in our development pro-cesses. And if we start costing technology requeststhings will change.

In particular in the area of chemical uids possibilitiesfor nding external support, and external partners wereconsidered to be promising while for electronics thiswould be difcult. This was noteworthy, because techno-logical development at HighTech was largely considereda combination of capabilities in electronics and chemicaluids.

We have unique capabilities in HighTech that combineelectronics and chemical uids. We cannot get thatfrom the outside. They are too specialised.

The possible external sourcing of innovation in uidssuggested that relations between the two technologicalareas were to be cultivated in new ways and the R&Ddepartments technological capabilities would change andperhaps even diminish. Costing technology projects wouldfocus too narrowly and hinder corporate-wide value crea-tion the director of R&D argued:

I am sceptical towards the idea of costing our technol-ogy activities. Technology development is something

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 745that emerges gradually and it may involve externalpartners. When we start costing one alternative [i.e.,the internal technology requests], we should also thinkabout the cost of the other alternative [i.e., externaltechnology requests]. But I am not sure what the realcosts for HighTech are if we start sub-contracting tech-nology development.

The suggested real costs were different from account-ing costs. In particular he was concerned whether the con-nections between chemical uids and electronics could beupheld in a situation where, because of costing, the focuswould be on narrow product line effects rather than corpo-rate-wide effects across time and space.

Example 3: LeanTech: the challenge of hardware modules andsoftware programs

Aiming to develop, produce and market high qualityproducts for audio and video transmission, LeanTech haddeveloped a customer base across telecommunicationcompanies and radio- and television stations all over theworld. The past 5 years sharp growth in revenues was ex-plained by the rms innovation activities. All LeanTechsproducts were customised and historically one centralchallenge had been to integrate software and hardwarein a connected offer to the single, individualised customer.Through design and sales work its development- and sales-efforts had focused on expanding markets through cus-tomisation and a exible product program. The resultinggrowth and expansion had made LeanTech outsource alarge part of its production capacity to selected suppliersthat had invested in advanced production technology. Inthis inter-organisational relation an open book arrange-ment had provided time and cost information about theproductions processes of the subcontractors.

Mobilisation of activity-based costing and innovation throughsharing components, modularisation and digitalisation

Design for manufacturability was considered an ele-ment in LeanTechs competitive success and use of com-mon component for modularisation and use digital andsoftware solutions to customisation problems in productinnovation made manufacturing effective. Together theseelements problematised the relationship between hard-ware and software components in innovation activities.

An activity-based costing calculation visualised eco-nomic effects of complexity of engineers design for manu-facturability initiatives. Historically, designers had paidattention primarily to direct cost, but the activity-basedcosting calculation focused differently:

The number of set-up had grown by more than 150%and the machines do not run full time and we had toomuch waste in process time. To meet the market condi-tions we simply have to enable the use of commoncomponents that can be used within and acrossmodules.

This imperative to use common components challengeddesigners because the implication was to reduce numberof components.We were confronted with very high resistance fromthe development engineers when we started to talkabout preferred types. In the development department,they have lots of technical arguments for using manydifferent components but with the open book, we couldshow the time- and cost-consequences of using manydifferent components. As a result, we have been ableto make the development engineers reconsider thedesign and perform some creativity in their design workto reduce the variation of components.

A large number of different components proposed manyset-up operations in the production process, machines hadto be stopped and the labour force had to switch manuallybetween types of components thus increasing time con-sumption and cost. Information about set-up-time andmounting costs in the production process motivated areduction in component selection from 15,000 to 5000components. Focusing on process- and production-aspectsthe role of engineers innovation was to reduce technolog-ical features and components of the products. And in addi-tion to sharing components yet another activity modularisation was proposed as a way to improve themanufacturability of the product. The logistics managerexplained:

By modularisation we pack more potential functional-ities into fewer modules and thereby get a fast reactionto customer orders and eliminate non-value-addedtime. The market condition is that we have to produceas quickly as possible, and by being production innova-tive we can produce everything within 23 weeks.

Modularisation developed a limited number of possibleproduct congurations which would make the productionand assembly process more predictable. In particular, theconcern with modularisation opened a new innovationambition where the distinction between software andhardware gained new signicance.

Historically, LeanTech was concerned with designingand assembling analogue devices but modularisationpushed customisation into digitalisation. Hardware andsoftware could be distinguished and introduce a principleof technology development and production taking intoconsideration predictability in production and creativityin development. Software programming could provideinnovation for customers; various types of software couldbe implemented on largely the same hardware platform.Customisation could be a question of digitalisation (soft-ware) that could quickly be congured according to cus-tomer needs; and the development work and supplies ofsoftware modules could be outsourced more freely to apool of independent software suppliers in LeanTechs sup-ply chain.

Activity-based costing dramatised certain conse-quences of digital rather than analogue technology relatedto design for manufacturability, as explained by the logis-tics manager:

There is simply a potential in software that we have toexploit. If we do this we can increase our productivityand we can deliver very quickly. In principle, we wouldbe able to deliver within just a fewweeks irrespective of

746 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754what the customer wants because our production isgeared towards it. It does create a set of advantages toorient ourselves more to software; we can see that fromour accounting statements.

Activity-based costing expanded the use of digital soft-ware solutions and technology because it presented ana-logue solutions as costly compared with the digitalsolutions.

Translation of innovation mobilising cost of capitalYet the activity-based costing calculation could be

challenged by its disregard for capital costs and depre-ciation that accrued from three types of events: increasein the average cost per unit on inventory, increase inwaiting time for critical components, and increase in R&Dcosts.

Firstly, the value of inventoried components and mod-ules increased since, although reduced in total numbersdue to digitalisation and modularisation, the average costper unit increased. Since all modules and componentshad to be combinable with all other components and mod-ules, they had to have more capabilities and functions. Insoftware, suppliers had to put new resources into pre-pro-gramming the software of modules and in hardware abroader variety of functions required more expensive com-ponents. Secondly, some of the components were criticalcomponents that could be difcult to source and unex-pected waiting time could occur. This risk was partly re-lated to critical components being so special that only asmall number of suppliers would be able to deliver them,or as the logistic manager explained:

Of course the hardware modules we now produceresult in more expensive inventories and if for exampleMotorola designs a new product and use some of thesame components as us it also creates extra costs insourcing and delays but we are not making any calcu-lations on those costs.

Such increase in inventory costs and risk of waitingtime in the supply of these components, due in part tonew surprising competitors, were not taken into accountby activity-based costing, and inventory cost was suddenlya challenge. In some situations, modularisation and digita-lisation could increase cost.

Thirdly, it was cumbersome to make components plug-and-play because they were changed over time and morerecent components had to integrate with older compo-nents. For example, software applications were not onlydesigned by different software-programmers but also atdifferent periods of time by different project teams at dif-ferent suppliers, and therefore a substantial amount of cus-tomisation work was needed in LeanTech. One process ofadditional customisation concerned the challenge ofchanging needs; another was a result of the number ofchanges that were made. Both increased the workload ofchanges to software, as it was explained:

Often there is already a long history of patches andbilateral interfaces resulting in spaghetti of intercon-nected applications, which is time consuming and aexpensive to maintain. But this is a discussion whetherthese costs relate to re-engineering cost of the productportfolio or if they are development costs that alsorelates to future products.

Software tended to grow bigger and become more com-plex because the easiest way to add a new feature or ttwo or more functionalities and packages was to add anew code. At the same time the aging of the software pack-ages fastened and then it became increasingly complicatedto make it work with other packages. As a consequence thetime when new software had to be developed was movedahead. This, together with the fact that the modularisationhad postponed the product differentiation to a point closerto the customer, put pressure on the programmers in Lean-Tech to add new features quickly for connecting differentsoftware packages.

Because of postponed customisation the priority of soft-ware work was often to make customisation work and de-liver to the customer. Making documentation and reviewof changes and new features were not prioritised. The re-sult was that a single delivery could exist in different ver-sions, each with subtly different structures and based onslightly different design concepts and assumptions. Toavoid this and accumulate the specic knowledge that fu-ture deliveries could benet from, changes and new fea-tures had to be studied and documented. A team ofsoftware engineers would review the codes in differentversions and the differences recorded and then agree onthe proper structure that all future changes had to be basedon. This made LeanTech suggest that software changeswere costly and that future changes could only be designedconsistently if the programmers work was based on prop-er documentation of the design and code. Not doing sowould reduce the durability of software. In other words,the frequent changes speeded up the aging of the softwareand the work to review and document became more dif-cult and time consuming as the size of the software in-creased. The logistics manager explained:

Our software packages are growing bigger and thisweight gain is caused by our fragmented supply of soft-ware from internal and external programmers. In mostof our work on software we dont know the originaldesign concept and the changes we make will be incon-sistent with the original concept; in fact they willdegrade the original concept and speed up the agingof the software, and software that has been repeatedlymodied in this way becomes substantially moreexpensive to change and update.

Complexity increased investments in R&D activitieswhich increased depreciation charges by what was sug-gested to be 5060%.

Concerns with cost of capital and depreciation wouldnot only economise R&D but also encourage its substitu-tion towards larger, standardised software packages whichin turn would impact inter-organisational relations. In-stead of several suppliers of software the innovationpotentially could be based onmarket standards frommajorsuppliers instead of own design and programming andexternally delivered software packages. The logistics man-ager explained:

meaningful and powerful by an appeal to their denitional

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 747correctness but only by connections with concerns devel-oped when they participate in mediating multiple actualand potential intra- and inter-organisational spaces andtimes. Table 1 presents and recounts the systems of innova-tion at stake in the three examples.

Table 1 shows that the management accounting calcu-lation speaks for much more than it describes. The surprisearising from the three examples is that the managementaccounting calculation is able to problematise not onlyinnovation activities but also central strategic propertiesof the rm such as its boundaries and capabilities.

The short translationThe primary quality of management accounting calcula-

tions in relation to innovation activities is hardly that theydescribe innovation activities and make them increasinglytransparent. Sales performance is not the same as choicesabout components in SuitTech, but it extends the probabil-ity that sales engineers will use external components. Con-tribution margin is not the same as electronic componentsWe have the option to use software suppliers that offera broad variety of functionalities in one integrated soft-ware package with standard interfaces. Our R&D activi-ties should not be reinventing the wheel. By substitutemany of our current software packages with larger andwell-designed software packages we can slow the agingof our software and minimise the modications anddocumentations work we need to make ourselves.

By using standardised software packages with morefunctionality LeanTechs programmers could meet speciccustomer needs by adding switches and create systemsthat appeared to be different by various functionalitiesbut were only small variations on one basic software pack-age. The software package would last longer before modi-cations were needed and its maintenance costs would bemuch lower. The perspective suggested by capital cost anddepreciation charges required LeanTech to upgrade the useof external software suppliers to fewer, large supplierswith standardised software packages.

Innovation, inter-organisational relations andmanagement accounting calculations

Short and long translations

The main observation from the empirical account is thatmanagement accounting calculations do not calculateinnovation activities per se but theymediate it. They hardlymake the innovation more transparent because they do notmodel it; rather theymediate between innovation activitiesand rm-wide concerns and inuence the intensity anddirection of innovation activities. Management accountingcalculations add a new perspective to innovation activities.This happens in short translations where innovationactivities are related to revenues, contribution marginsand ABCmargins, or in larger translationswhere innovationactivities are linkedwith sourcing strategies and changes inthe competencies of rms through competing calculations.Management accounting calculations rarely becomeand chemical uids, but it extends development engineersexperimentation in HighTech. An ABC margin is not thesame as complex components in LeanTech but it helpssales engineers to be interested in a limited set of preferredcomponents.

The short translation links the innovation to the rm byproblematising when the innovation activity is in excessand has departed from its contribution to making the rmviable. In SuitTech, sales performance only interveneswhenthere is a shortfall which happens when sales engineers in-vest excessive time in assembling a customised product.When sales variance is unfavourable, attention is directedtonalise orders rather than toproduce leads. There is a lim-it to the time a sales engineer can spend combining compo-nents into a product. Sales performance re-frames salesengineers attention towards closing orders when it is injeopardy. Salesperformance thus translatesa complexques-tion of technology into a simple question of time.

A parallel movement can be found in HighTech wherethe contribution margin justies new technology in inno-vation projects and thus encourages developers to experi-ment. The contribution margin helps to explain whetherin fact R&D is able to translate into increasing prices far be-yond the limited direct cost added from innovation. TheR&D activity has to develop a market response in demandand in price increase. There is a constraint to innovation,however, as the technology has to have applicability inan existing product range. While the contribution marginexpands innovation by emphasising R&D innovation as ageneral drive towards increasing prices, it also reducesinnovation by insisting that technology development, overa time period, be integrated with technological capabilitiesof existing product ranges.

In LeanTech the short ABC calculation reduces the num-ber of components that sales people can muster and use ina particular product thus reducing the elements in innova-tion arrangement. The calculation also increases the use ofmore powerful components thus substituting analoguesolutions by digital solutions because it presents costs ofexibility.

These three examples of short translations illustratehow a management accounting calculation can work oninnovation even if it does not directly represent innovationactivities. There is an indirect link between managementaccounting calculations and specic innovation activities,which starts from adding perspective and context to inno-vation. It stipulates a context for innovation that requires itto be protable.

Thus, as has been proposed also by others (e.g., Ahrens &Chapman, 2004, 2007; Boland & Pondy, 1983) managementaccounting calculations do inuence situated decisions andmanagers do use such information in managing R&D pro-jects (Nixon, 1998). Yet, many decisions in innovation areinteresting not only in R&D settings since their effectsspread to manufacturing and sales and therefore, manage-ment accounting calculations help to make the effects ofinnovation economic (Davila & Wouters, 2004; Hansen &Jnsson, 2005; Jnsson & Grnlund, 1988). The usefulnessof management accounting calculations is paradoxical be-cause they are not inherently connected to the activitiesthey help organise. In all examples, the calculation requires

Table 1Translations of innovation management by management accounting calculations in three examples.

SuitTech HighTech LeanTech

Innovators Sales engineers Development engineers Production engineersDominant calculation Sales Contribution margin ABC margin

Short translations Reduction/extensions ofinnovation

Sales performance focuses on ordersclosed and contracts signed. It omitsdirect costs and extends innovationby expanding types of availablecomponents. Yet, when sales varianceis unfavourable it reduces innovationand motivates closing orders quickly

Contribution margin visualises increasing diffe ncebetween steep revenue effects and moderate d ect costeffects of new technology and justies indirec osts ofexperimentation (with electronics and chemic uids).It expands innovation by protecting technologexperimentation but reduces it by insisting thtechnology has to t an existing product progr whenunfavourable contribution margin variance occ rs

ABC margin visualises cost of complexity ofcustomised designs and constrains the numberof technology choices but it increases thepower of each hardware module. It reducesinnovation by stipulating hardware choices butextends innovation by using strongercomponents

Materialisations of theinnovative practice

Combinations of special andcustomised components sourcedfrom anywhere

Electronics and chemical uids mobilised in te nologydevelopment

Hardware modules and software programsdeveloped in a lean supply chain

Long translations Innovation strategy Innovation concerns productvariation vis--vis customerrequirements. Through differentcombinations of special andcustomised components salesengineers search for distinctsolutions fullling individualcustomer needs. Innovation adjuststhe product through combinations ofphysical entities

Innovation concerns development of new prod ctssetting industry standards and create new cus merwants. There is considerable technology elasti ty ascustomers want latest technology. Innovation mbedstechnological capabilities of electronics and ch icaluids and concerns structural adaptation of pr ucts tonew technological possibilities

Innovation concerns process innovationthrough modularisation of hardware andinternal software design, programming anddocumentation

Inter-organisationalrelations

Suppliers play a signicant role indelivering the wide rangecomponents to be drawn in as neededin combinative innovation

Suppliers play no role in regard to developingtechnological capabilities. However, suppliers ay animportant role in optimising direct cost new p ducts

Suppliers play an important role as suppliers ofhardware and specialised software packages

Competing calculation Direct costs/contribution margin(costs of customisation)

Indirect costs of R&D department/gross margin costs ofexperimentation)

Cost of capital (costs of simplication)

Substituting innovationelement

Programmable standard components External technology development Larger software packages from suppliers withsurplus functionality and standard interfacesfor conguring modules

Alternative innovationstrategy

Innovation created by softwareengineers

Innovation created by suppliers with close rela ons Innovation created by increasing use ofsoftware packages with surplus functionalityand standard interfaces

Innovation from the inside Innovation from the outside Innovation from the outside

Alternative inter-organisational relations

Arms length relationship withsuppliers of standard components

Innovation through suppliers unique knowled Innovation through close relationship withsuppliers of standardised software modules

748J.M

ouritsenet

al./Accounting,O

rganizationsand

Society34

(2009)738

754ad

reirt calyatamu

ch

utocieemod

plro

(

ti

ge

cal arte

Calculation 1

nagem

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 749help because its tension is difcult to appreciate withoutmediation: economise time (in SuitTech), develop marketsthough new technology (in HighTech), and make manufac-turable solutions (in LeanTech). The calculation connectsthe innovation activity to other concerns.

A short translation relates the calculation with changesin innovators conduct but it does not question the innova-tion strategy. It is short when it economises innovationthrough inuencing the time, resources and orientation ofinnovators. It bends the innovation to its context by presen-tation of nancial effects in revenues, in contribution mar-gins and gross margin. A relevant management accountingcalculation is specic and therefore partial, and its mobili-sation requires support from others such as the order man-ager (in SuitTech), the new product development manager(in HighTech) and the production engineers (in LeanTech).

The long translationIn addition to the short translation, there are also long

translations generated by competing calculations. Thesetranslations become longer because they develop complexproblematisation of the role of innovation in the rmstrategic consequences beyond the rm by taken manymore entities into account. The tension between calcula-tions is important, and it can be illustrated generally as

Technologi

Innovation strategy

Inter-organisationalrelation

Fig. 1. Elements in the analysis of the role of main Fig. 1.Fig. 1 illustrates that the stake in innovation manage-

ment is a struggle over with technological artefacts. Eachmanagement accounting calculation denes some rulesin this struggle which proposes not only different compo-sitions of technological artefacts but also different innova-tion strategies and sourcing arrangements. Specically, themaps of the translations show connections between man-agement accounting calculations, technological artefacts,innovation strategy, and (inter-) organisational relations.1

Secondly it illustrates that there are competing calculationswhich propose decisions about innovation and (inter-) orga-nisation differently. Thirdly, there are two arrows one boldand one dotted. The bold arrow identies a dominant pro-

1 These elements are clearly the ones identied in our research. Inprinciple, there could have been others.cess of translation while a dotted arrow identies a compet-ing calculation which requires a different settlement ofinnovation and (inter-) organisation. This work on theboundary of the rm may be central in the management ofinnovation in a period of time when rms strategies changemuch faster than they can develop their competencies (Cas-tells, 2000; Parolini, 1999). Figs. 24 show the application ofFig. 1 on the three examples.

Fig. 2 illustrates the production of tensions between thetwo calculations in SuitTech (sales and direct costs) overthe amount of special components that sales engineerscan legitimately take into consideration. The two calcula-tions guide this decision differently. Sales performance ex-pands the number of possible components because itmakes revenue considerations more important than costconsiderations and develops innovation through combina-tion of components arriving from an extended space. Thus,mobilising this calculation, engineers focus on customisa-tion of products through combination of components andthe inter-organisational relation is a large, well-assortedand heterogeneous inventory. Adding the direct cost tothe picture makes problematisation of this relation possi-ble. When direct cost is mobilised, managers identify a ten-sion between resources and efforts invested in designingan order. Innovation through combination of special com-

facts

Calculation 2

AlternativeInter-organisational

relation

AlternativeInnovation strategy

ent accounting calculations in long translations.ponents appears to be costly, and including direct cost inthe performance measures economises innovation activi-ties by shifting attention to programmable componentsthat are more readily available and whose variation canbe guaranteed by software exibility rather than by hard-ware components. Innovation is here to a large extent metby software programming. Inter-organisational relationsare then proposed to be an inventory of a limited rangeof standard components that can be supplied steadilyand predictably. The more standardised the set of possiblecomponents the more amenable innovation is to controlthrough a form of standard cost system.

Fig. 3 illustrates that, in HighTech, the struggle iswhether a large R&D department which takes pride indeveloping general knowledge and not only product spe-cic knowledge is appropriate. The contribution marginapproach sees the costs of the R&D department as a periodcosts and allows it to develop its own distinctions and

750 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754Sales Performance

Broad range of concerns protecting in-house capabilities related to elec-tronics and chemical uids. These activities are discretion-ary investments and only to a lesser extent associated withproduct protability calculations. A positive contributionmargin is proposed to arise from increase in price ratherthan reduction in cost. The cost calculation suggested asa way to convert the period costs of the R&D department

Special compstandard comp

Combination of physical components

innovative suppliers

Fig. 2. Long translation between management accounting calculatio

Electronics/chemical fluids

Contribution margin

Experimentation

Supplier onlymanufacturing

Fig. 3. Long translation between management accounting calculatio

Hardwasoftwa

ABC margin

Software design, program-mingand documentation

Suppliers of special hard-and software

Fig. 4. Long translation between management accounting calculatioDirect costinto a product costs, in contrast, focuses on the efciencyR&D investments and costs and it proposes external com-petences in electronics uids as possible new sources ofknowledge. Thus, the allocation of the period cost toprojects proposes to develop a stronger associationbetween individual R&D projects and individual productdevelopment activities. In addition it also makes directed

onent/onents

Arms length relationship

Software programming

ns, innovation, and inter-organisational relations in SuitTech.

Indirect costs

Suppliers invovled in technology development

Product innovation efficiency

ns, innovation, and inter-organisational relations in HighTech.

re/re

Capital costs

Few suppliers ofstandard software

Configuration of modules

ns, innovation, and inter-organisational relations in LeanTech.

mewhat youmean and then the calculation has to emerge.Mere cognitive interpretation of innovation is not collec-tively actionable; innovation has to be inscribed and madeacalculationbefore it canbeactedon. This is the context thatthe calculation develops and makes possible. Even peoplewho are inside an innovation such as the R&D Director inHighTech have to step out and mobilise the managementaccounting calculation when they want to say somethingto justify innovation. Standing out is a movement, but notamovement fromoneplace toanother. It is amovement intoa calculationwhere some effects can be proposed, surveyedand compared. Mere mental interpretation is not enough. A

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 751outsourcing of R&D initiatives possible and thus develops anew inter-organisational R&D agenda.

Fig. 4 illustrates, in LeanTech, a struggle over the useof exotic components or general standard software. TheABC margin motivates a limited range of complex, some-times exotic, expensive components; cost of capital anddepreciation charges, in contrast, reduce complexity of compo-nents and draw on standard software packages. These prop-ositions reach into the inter-organisational space becauseexotic and specialised components require concerned andintensive interaction with suppliers about the componentsperformance while the use of standard software packagesrequires that the rm interacts with large suppliers whocan develop the technologies of their application largelyby themselves as they dene the industry standard.

The tensions arising in the three examples of proposedtransformation are minimalist. When the three examplesdraw new possible calculations into play they pay atten-tion only to those parts hereof that will make its proposi-tions different from the existing arrangement. It is likely,however, that if the cost strategy would gain power in Suit-Tech and HighTech managers would also quickly concernthemselves with revenues. Rather than seeing the oppos-ing calculations as suggestions of effective managementcontrol systems, they are much more problematising de-vices which challenge dominating arrangements by high-lighting the special features they problematise.

Management accounting calculations in tension

The three examples illustrate that innovation strategycan be an effect of management accounting calculations.The tensions between calculations are important becausethey frame decision making, risk management and strate-gic uncertainty by adding sequences of proposed effects.The short and long translations both create contexts forinnovation activities but there are differences. The shorttranslation develops immediacy between innovation activ-ities and economic effects. In the long translation some ofthe power of a calculation derives from its tensions withother calculations over the appropriate way in which tomake innovation a productive resource for the rm. Thetension is that there is not one but at least two ways inwhich choices over technological components can bemade. The calculations provide these justications whichare inside the process of translation rather than outsideit. The management accounting calculation does not judgethe relative merits of different propositions about innova-tion; the management accounting calculation is part ofthe proposition that it mediates.

If managers do not follow the calculation, they have toproduce another calculation to make their point. In orderto combat one calculation another one is needed. Calcula-tions play a role in the development of new propositions ofthe relevance, power, effects and character innovation inrelation to rm strategies. In LeanTech the ABC calculationis able to rally interest only because it is possible to calculatethecostofhuge inventories. Theproblemofheterogeneityofcomponents is not visible before it has beenmade a calcula-tion. If someonewould claim, say, that innovation should bemore efcient, another voicewould immediately say showcalculation is stronger.The calculation requires a network of practices and

commitments to operate; it will not operate on its own.Any particular economic category performs differentlyacross the three examples. For example, in LeanTech be-cause of cost and time calculations it is possible to proposean integrated, lean supply chain governed from one place.In HighTech, also because of time and cost information it ispossible to contemplate outsourcing of R&D and in Suit-Tech again because of cost information it is possible to con-ceive of in sourcing of many production tasks. Likewise,indirect cost can be proposed to drive value (HighTech)and to destroy value (LeanTech). The calculations do notdetermine their impact; they gain power in interactionwith the development of the entities they engage. Even ifsome parts of the accounting calculation are strengthened,it ows over in new ways; even if, for example, ABC calcu-lations reduced production costs in SuitTech, it opened anew space for increased cost of capital and depreciationcharges. Therefore, calculations gain strength not becausethey are inherently good or reasonable but only by theiroutside found in the activities and strategies it participatesin shaping and developing. This point extends questionsabout the completeness of calculations (e.g., Lawler,1983; Simons, 1995, p. 76-7) which suggests that the con-tribution margin is more complete than sales performance,and ABC margin is more complete than contribution mar-gin. But the three examples show that completeness isnot a property of the calculation. It is useful to substituteconcerns about completeness with the relational qualitiesof the whole network which constitutes the power of thecalculation. Sales performance, contribution margin andABC margin are powerful because they can motivate ac-tions to be performed by innovators. This translation,rather than represent the innovation choices, creates acontext for innovation activities to occur.2

2 The management accountants in the three rms claimed that theirextensions of the calculations were more complete than other calculations.Direct cost was added to a sales gure in SuitTech, indirect cost was addedto contribution margin in HighTech and cost of capital was added toActivity Based Costing in LeanTech. Accountants proposition to addcompleteness in calculation is, however, a stylistic and formalistic concernwith the mathematics of inscription. Inscription is not a copy of the worldbut only a particular ordering of the revenues and costs accumulated in theaccounting system; for the inscription to work, the world of innovationactivity and management has to be added and therefore even if morecomplete stylistically and formally, they can be less complete in the worldof activity and strategy. The power of the calculation derives from itsintertwinement with action.

752 J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754Management accounting calculations as context forinnovation and sourcing arrangements

Research which suggests a constructive role for man-agement accounting calculations in developing innovationobserves that managers develop dialogue about calculationsin the pursuit of innovation (e.g., Davila, 2000; Hansen &Jnsson, 2005; Jnsson & Grnlund, 1988; Nixon, 1998; Si-mons, 1987; Simons, 1990). The addition proposed by thethree examples is that the calculations do not only workby moving closer to innovation and by looking more care-fully at details of innovation practices. It may be that inter-active use, or use of multiple nancial and non-nancialcalculations, focus attention to certain ways of seeing therm through more details and more interactions, and thecorollary probably is that managers know more about thedetails of affairs and develop a unitary interpretation ofthe demands of complex markets. The three examplesshow, in contrast, that the important link is the movementof innovation away from its place into diverging concernsabout the sourcing and strategy.

Like Hkansson and Lind (2004) and Miller and OLeary(2007, 2005) the three examples illustrate that innovationactivities are often inter-organisational and that mediatingtechnologies help rms enrol others in this accomplish-ment. The calculations are involved in coordinating therms inter-organisational eld by extending existing con-gurations of actors and interests into alternative possiblecongurations. As Miller and OLeary point out, markets,knowledge and actors are co-produced in the developmentof innovation activities: markets, science and organisationare co-produced via mediating technologies. In the threeexamples, management accounting technologies mediatethe development of rm boundaries, capabilities and mar-ket requirements.

Management accounting calculations mobilise the envi-ronment and a variety of propositions are added that makeup not only existing environments and but also possibleones. The three examples illustrate how the compositionof the environment is in process. It may be that Simons(1990, p. 142) concerned question How do managersidentify strategic uncertainties? can be addressed by thethree examples. The solution appears simple changethe role of the calculation in the system of explanationand the environment emerges as an effect of the analysis.More particularly, this means that it is possible to contem-plate and prepare for the environment through calcula-tions. Perhaps this is why Simons (1987) prospectors usea lot of information. They are prospectors exactly becausethey have become knowledgeable about many aspects ofthe environment which is then used to design and cultivatethe prospecting abilities. The tensions between calcula-tions produce this opportunity. The three examples illus-trate that management accounting calculation can bemobilised to extend strategy in addition to implementstrategy. In effect management calculations can commandmuch more than they calculate.

Even if the calculation produces visibility, it is not pri-marily about the contours of the objects it proposes tomanage. Rather than making a claim to increase visibilitymore and more into details of organisational spaces (e.g.,Ezzamel et al., 2004; Ezzamel et al., 2008), the manage-ment accounting calculation may also gain by relatingthe economy to other entities such as innovation and envi-ronment. In this optic, sales performance speaks for therm and identies the difference between rm, suppliersand customers in SuitTech. Contribution margin speaksfor the role of technology in developing markets in High-Tech. ABC margins speak to reduce the cost of productioncomplexity developed by innovative arrangements inLeanTech. They all relate concerns about innovation andinter-organisational relations to concerns of other situa-tions and events in the rm and beyond. It transports con-cerns about innovation by relating them to other concernssuch as production within the rm more than to the indi-vidual concern of the innovation situation. The manage-ment accounting calculation is strong because it helps todevelop context (see also Mouritsen, 1999).

Conclusions

A management accounting calculation does not de-scribe or represent innovation and sourcing activities inany detail, but it adds perspective to them and relatesthem to the rm. In effect the management accounting cal-culation is part of a relationship between economy, inno-vation and environment. The management accountingcalculation speaks for the rm and puts pressure on inno-vation to account for its contribution in this respect.

Based on examples from three rms, managementaccounting calculations sales performance, contributionmargin, and ABC margin are mobilised in relation toinnovation and in turn, surprisingly, in relation to sourcingand strategy. The management accounting calculationworks by extending or reducing the number of entities thatinnovation can take into account, less by describing thedimensions of innovation and inter-organisational designand more by adding perspective to them. This mechanismis stronger when a calculation is challenged by anotherone. This is when there is maximum pressure on innova-tion activities to show their strategic signicance. The ten-sions between calculations bend organisational activitiessuch as innovation to considerations such as growth, pro-ductivity, protability, and liquidity.

Management accounting calculations mediate andmobilise innovation through short and long translations.Short translations exist when management accounting cal-culations encourage extension or reduction of innovationactivities when it proposes performance to be adequate orinadequate. Long translations mobilise at least two calcula-tions to problematise the role of innovation for corporatepurposes differently. Management accounting calculationschallenge each other and develop organisational strugglesnot only about the role of innovation, but also about itslocation in time and space technologically, organisationallyand environmentally. The process of developing relationsis, paradoxically, dependent on the management account-ing calculation being partial because then it presentstensions. The calculation can never be total.

Management accounting calculations can motivatevery long sequences of translation as they are associatedwith strategic propositions about technology and theboundaries of the rm. One of the possible effects of such

J. Mouritsen et al. / Accounting, Organizations and Society 34 (2009) 738754 753translations is that the rms strategy for managing inno-vation can undergo drastic reformulation. Another effectof translation is that management accounting calculationsmay create surprising effects very far from their presumedoutcomes. When new calculations come into existencethey reach into new situations that, in turn, inuence therole of the calculation.

Generally, the management accounting calculationholds certain characteristics of innovation in place byshowing their broader justication. Sometimes the man-agement accounting calculation shows this as a shorttranslation where the calculation is tightly coupled to deci-sions regulating the innovation activity. In other situations,however, the management accounting calculation enter-tains a long translation though interaction with other cal-culations where many new elements from whole systemsof innovation are taken into account. Challenging a certaininnovation system, opponents mobilise other managementaccounting calculations that draw other consequences ofinnovation. Innovation is thus not developed merely be-cause of good innovative ideas; innovation has to passthe test of management accounting calculations before itcan be heard, and the challenge is a whole system of inno-vation and sourcing that is given corporate relevancethrough the management accounting calculations. Man-agement accounting calculations problematise the rm,its innovation and technologies, and its boundaries.

Acknowledgements

We wish to thank participants at workshops and semi-nars who have provided valuable comments. The twoanonymous reviewers have been very patient and support-ive, and we extend our thanks to Anthony Hopwood, ArielaCaglio, Angelo Ditillo, Christina Boedker, and HabibMahama.

References

Abernethy, M. A., & Brownell, P. (1997). Management control systems inresearch and development organisations: The role of accounting,behaviour and personnel controls. Accounting, Organisations andSociety, 22(3/4), 233248.

Abernethy, M. A., & Stoelwinder, J. U. (1991). Budget use, task uncertainty,system goal orientation and subunit performance: A test of the thypothesis in not-for-prot hospitals. Accounting, Organisations andSociety, 16(2), 105119.

Ahrens, T., & Chapman, C. (2004). Accounting for exibility and efciency:A eld study of management control systems in a restaurant chain.Contemporary Accounting Research, 21(2), 271302.

Ahrens, T., & Chapman, C. S. (2007). Management accounting as practice.Accounting, Organisations and Society, 32(1/2), 127.

Alvesson, M. (2003). Beyond, neopositivits, romantics and localists. Areexive approach to interviews in organisational research. Academyof Management Review, 28(1), 1333.

Amabile, T. M., Conti, R., Coon, H., Lasenby, J., & Herron, M. (1996).Assessing the work environment for creativity. Academy ofManagement Journal, 39(5), 11541184.

Birnberg, J. G. (1988). Discussion of an empirical analysis of theexpenditure budget in research and development. ContemporaryAccounting Research, 4, 582587.

Bisbe, J., & Otley, D. (2004). The effects of the interactive use ofmanagement control systems on product innovation. Accounting,Organisations and Society, 29(8), 709737.

Boland, R. J., Jr., & Pondy, L. R. (1983). Accounting in organisations: Aunion of natural and rational perspectives. Accounting, Organisationsand Society, 8(2/3), 223234.Briers, M., & Chua, W. F. (2001). The role of actor-networks and boundaryobjects in management accounting change: A eld study of animplementation of activity-based costing. Accounting, Organizationsand Society, 26(3), 237269.

Brownell, P. (1985). Budgetary systems and the control of functionallydifferentiated organisational activities. Journal of Accounting Research,23(2), 502512.

Burns, T. G., & Stalker, M. (1961). The management of innovation. London,UK: Tavistock.

Callon, M. (1986). Some elements of a sociology of translation:Domestication of the scallops and the shermen on St. Brieuc Bay.In J. Law (Ed.), Power, action and belief. A new sociology of knowledge?(pp 196233). London: Routledge and Kegan Paul.

Callon, M., & Latour, B. (1981). Unscrewing the big Leviathan: How actorsmacro-structure reality and sociologists help them to do so. In K.Knorr-Cetina & A. V. Cicourel (Eds.), Advances in social theory andmethodology. Towards an integration of micro- and macro-sociologies(pp. 277303). London: Routledge and Kegan Paul.

Castells, M. (2000) (2nd ed.). The rise of the network society (Vol. 1).Malden, MA, USA: Blackwell Publishers.

Chua, W. F. (1995). Experts, networks and inscriptions in the fabricationof accounting images: A story of the representation of three publichospitals. Accounting, Organisations and Society, 20(2/3), 111145.

Clark, K. B., & Fujimoto, T. (1991). Product development performance.Boston: Harvard Business School Press.

Cooper, R. (1992). Formal organisation as representation: Remote control,displacement and abbreviation. In M. Reed & M. Hughes (Eds.),Rethinking organisation New directions in organisation theory andanalysis (pp. 254272). London: Sage.

Cooper, R. (1997). The visibility of social systems. In K. Hetherington & R.Munro (Eds.), Ideas of difference (pp. 3241). Oxford: BlackwellPublishers/The Sociological Review.

Cooper, R. G., & Kleinschmidt, E. J. (1987). New products: What separateswinners from losers? Journal of Product Innovation Management, 4(3),169184.

Cooper, R., & Slagmulder, R. (2004). Interorg