Extending private-collective innovation: a case study

Extending private-collectiveinnovation: a case study

Matthias Stuermer, Sebastian Spaeth andGeorg von Krogh

ETH Zurich, Department of Management, Technology and Economics, Zurich, [email protected]

The private-collective innovation model proposes incentives for individuals and firms to

privately invest resources to create public goods innovations. Such innovations are character-

ized by non-rivalry and non-exclusivity in consumption. Examples include open source

software, user-generated media products, drug formulas, and sport equipment designs. There

is still limited empirical research on private-collective innovation. We present a case study to

(1) provide empirical evidence of a case of private-collective innovation, showing specific

benefits, and (2) to extend the private-collective innovation model by analyzing the hidden costs

for the company involved. We examine the development of the Nokia Internet Tablet, which

builds on both proprietary and open source software development, and that involves both

Nokia developers and volunteers who are not employed by the company. Seven benefits for

Nokia are identified, as are five hidden costs: difficulty to differentiate, guarding business

secrets, reducing community entry barriers, giving up control, and organizational inertia. We

examine the actions taken by the management to mitigate these costs throughout the

development period.

1. Introduction

In a private investment model of innovation,firms use internal processes to create ideas,

knowledge, and technologies and commercializethese in the market place. Firms appropriatereturns from private investment in innovationthrough intellectual property rights (Granstrand,1999). This model is contrasted with the private-collective model of innovation, where firms andindividuals expend private resources to createpublic goods innovations. Such innovations arecharacterized by non-rivalry and non-exclusivityin consumption (von Hippel and von Krogh,2003, 2006). It is similar to ‘open innovation,’pertaining to models of innovation where firmsfrequently exchange ideas, knowledge, and tech-nology with outside firms and individuals (Ches-brough, 2003). However, open innovation doesnot assume that intellectual property rights are

forfeited and the resulting innovation is offered tothe public for free. Examples of the private-collective innovation model include collaborativecomposing of music on the Internet by manymusicians, the open and collective developmentof a drug formula for treating malaria, or thecreation and sharing of new designs for sportingequipment among sports enthusiasts. An often-cited example of the model is open source soft-ware development resulting in products such asLinux, MySQL, or Apache. Open source softwarecomes with licenses that make it non-exclusive:the software is free for all to download, use,modify, and redistribute. Open source softwareis also characterized by non-rivalry as one per-son’s use of the product does not diminish anyoneelse’s benefits from using it.

Although researchers have examined indivi-duals’ motivations to participate in open sourcesoftware development, to date, there has been

R&D Management 39, 2, 2009. r 2009 The Authors. Journal compilation r 2009 Blackwell Publishing Ltd,1709600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

enes

Highlight

enes

Highlight

enes

Highlight

limited empirical examination of a firm’s incen-tives for private-collective innovation. Moreover,the literature has emphasized the benefits themodel brings to the innovator rather than thecosts and has not discussed how the latter couldbe mitigated. Research has shown that the im-plementation of new models of innovation oftenhas unintended consequences, including ‘hiddencosts’ (e.g., Crawford, 1992), and there is a needfor more empirical work on the benefits and costsof implementing private-collective innovation.

In this paper, we advance empirical research onthe incentives and costs of the private-collectivemodel of innovation. A case study design allowsan in-depth investigation of benefits, costs, andactions only partly discussed in prior work (vonHippel and von Krogh, 2006). Our case draws onquantitative and qualitative data from the crea-tion of the Nokia Internet Tablet. Nokia basedthe development of this product mainly on opensource software and made a large part of theresearch and product development transparentand accessible as a ‘public goods innovation.’Outside contributors involving firms and indivi-duals, unpaid by Nokia, expended a significantamount of private resources on its development.

The paper is structured as follows: in the nextsection, we discuss the private-collective modelfor innovation incentives. The third section de-scribes the research design, and the fourth sectioncontains the case description. The fifth sectionpresents the findings organized along the topics ofbenefits and costs incurred in the implementationof the model and strategic actions to mitigatethese costs. Finally, we conclude the paper andoutline implications for management practice andfuture research.

2. The private-collective innovation model

There are two predominant models of innovationincentives in the technology and innovation man-agement literature. The private investment modelassumes that innovators step forward and investin innovation if and when they can appropriatereturns from these investments. Intellectual prop-erty rights is a necessary condition for the modelbecause it safeguards returns appropriation (Ar-row, 1962; Dam, 1995). In contrast, the collectiveaction model assumes that innovators, providedwith the right public subsidy, contribute to publicgoods innovations (David, 1992, 1998; Stiglitz,2006). Public goods are characterized by non-rivalry and non-exclusivity in consumption. In-

novations are made freely available to all aspublic goods. Science is often cited as an exampleof this model. However, companies have theoption to free-ride on public goods innovations,such as, for example, a biotechnology companycommercializing scientific knowledge on geneticswithout contributing research back to thescientific community. Therefore, society electsto subsidize the activity of innovators, e.g.,university-based research on the human genomefunded by the government.

Recently, a third model, called the private-collective model of innovation incentives, hasbeen suggested where public subsidy is absentand where the innovator expends private re-sources for public goods innovation (von Hippeland von Krogh, 2003). The model is counter-intuitive: why should I make my innovationsavailable to all and why pay for something thatanyone else can use for free? Generally speaking,in the case of private-collective innovations, theinnovator receives higher benefits when contribut-ing to the public goods creation than by only free-riding on its production by others. One aspect isthe privately retained tacit knowledge that inno-vators receive through the production of freelyavailable knowledge that distinguishes them frompure users of the explicit knowledge. This impliesthat firms receive certain benefits during theprocess of creating publicly available innovations,while the mere application of such knowledgebears less incentives. (see also Grand et al.,2004; Gachter et al., 2006)

2.1. Benefits when applyingprivate-collective innovation

A closer examination of the model outlined in vonHippel and von Krogh (2006) and other literaturereveals six complementary benefits for firms toinnovate in this manner: the cost of controllingknowledge, learning, reputation gains, and fastand widespread diffusion of innovations, as wellas lower cost of innovation and manufacturing.First, in the long run, the cost of protectingknowledge needed to innovate (Liebeskind,1996) might outweigh the benefits of doing so.Often, extensive investments in knowledge man-agement systems are needed to protect informa-tion that ultimately and inevitably spills over tothe public (Alavi and Leidner, 1999; Foray, 2004).For example, although the source code for Sony’srobot dog Aibo was protected, it was ultimatelyhacked and published by Sony’s customers.

Development of the Nokia internet tablet

r 2009 The Authors

Journal compilation r 2009 Blackwell Publishing Ltd

R&D Management 39, 2, 2009 171

enes

Highlight

Second, innovators who contribute to collectivegoods innovations benefit from learning fromtheir own and others’ contributions. In additionto benefits garnered from the public goods itself(products and services), innovators also benefitfrom learning in the process of creating it (Allen,1983; Nuvolari, 2004; Baldwin and Clark, 2006).Thus, it should come as no surprise that manycontributors to open source software projects arecomputer science students. By providing opensource software to the public, contributors mayget others to use it, test it, and provide feedbackon how to improve it (Lakhani et al., 2002). Someauthors have even referred to open source soft-ware projects as ‘epistemic communities,’ wherepeople create shared knowledge of software de-velopment (Edwards, 2003).

Third, innovators may gain a positive reputa-tion by privately expending resources for publicgoods innovations (Allen, 1983; Lerner andTirole, 2002; Muller and Penin, 2006). For ex-ample, PricewaterhouseCoopers gains a positivereputation among regulators, peers, and custo-mers when they provide research to the public onpractices of corporate governance. The firm’sreputation is further enhanced when regulatorsactively use and reference the research duringstandard setting in principles of auditing andcorporate governance (PricewaterhouseCoopers,2005).

Fourth, being the first to contribute a publicgoods innovation increases the likelihood of ben-efiting from fast and widespread adoption of theinnovation (Allen, 1983). As a consequence, firmsmay gain an advantage over competitors stem-ming from network effects (von Hippel and vonKrogh, 2003). Establishing a ‘dominant design’ oran ‘open standard’ onto which the firm can fitother technologies and even preempt the intro-duction of competing technology may provide thefirm with additional advantages (Economides,1996; see also Economides and Katsamakas,2006).

Fifth, by contributing to public goods innova-tions, the firm may lower the cost of innovation.Chesbrough (2003) argued that involving outsidefirms, organizations, and individuals in the devel-opment of products reduces the direct labor costin innovation. In addition, when the firm con-tributes to public goods innovations, such as opensource software, it can also effectively reuseexisting technology found in the public sphere.Research has shown that, in software develop-ment, the reuse of open source software is con-siderably higher than the reuse of firm-internal

software, which should have a positive impact onthe cost of innovation (see Haefliger et al., 2008).However, much of this software comes withrestrictive open source licenses, which requiresthe firm to make any combination between thisand other software adhere to this license. Hence,while the reuse of such products may reduce thecosts of innovation, it also ‘forces’ the firm tocontribute to public goods innovations.

Sixth, the supply of public goods innovationsto the market makes it possible for manufacturersto learn about these innovations and reducemanufacturing-incurred fixed costs related to re-search and development. Free access to innova-tions may incentivize manufacturers to ramp upmanufacturing capacity, pursue economies ofscale, and reduce the price of manufacturedproducts. Additional benefits to the customersof the manufactured product may include en-hanced product quality and product warranties(Kotha, 1995; Harhoff et al., 2003).

2.2. Empirical evidence and hidden costs

Empirical research on the private-collective inno-vation model is mainly found in the field of opensource software development, where the focus hasbeen on individual contributors and projects (e.g.,Baldwin and Clark, 2006; Gambardella and Hall,2006; Roberts et al., 2006; Shah, 2006). Researchon the application of the model by firms is rare,with some exceptions: Dahlander (2004) exploredthe network effects available to firms that provideopen source software to the public. Jeppesen andFrederiksen (2006) explored users’ motives tocontribute voluntarily to the development ofmedia products by firms. Henkel (2006) investi-gated firms that revealed open source softwareembedded in their devices to other firms andfound incentives for them to do so. Given thefocus on public goods innovations in these works,the authors have tended to focus on the cost offorfeiting intellectual property rights in the pri-vate-collective innovation model.

Innovation research has pointed to severaltypes of ‘hidden costs’ in the implementation ofnew innovation models that are not inherent to orcaptured in the models themselves. For example,Crawford (1992) and Smith (2004) pointed to thecosts of implementing the accelerated productdevelopment model in US manufacturing indus-tries. Firms often find that the rapid launching ofproducts that have not been properly tested leadsto costly recalls, or that fast product development

Matthias Stuermer, Sebastian Spaeth and Georg von Krogh

172 R&D Management 39, 2, 2009 r 2009 The Authors


leads to significant delays in pilot- and full-scalemanufacturing. With respect to open innovation,Kessler et al. (2000) found some indication thatthe cost of product development increased withincreasing dependency on external sources oftechnology in the innovation process. Empiricalresearch is needed that validates the specificbenefits in the implementation of private-collective innovation, examines the costs incurredin such innovation, and identifies firms’ strategiesto mitigate these costs. We contribute to thisresearch by investigating the implementation ofprivate-collective innovation in the case of theNokia Internet Tablet development.

3. Research design

The research on the development of Nokia’sInternet Tablet focuses on the process of imple-menting private-collective innovation. We investi-gate benefits in implementation and extend theprivate-collective innovation model by identifyingimplementation costs and strategies to mitigatethese. Research on implementation processes typi-cally requires longitudinal observation (Pettigrew,1990), prompting a case study design. In order toobtain insights into the development process, wegathered different types of data and performedquantitative and qualitative analysis. Done prop-erly, such combined analysis offers a valuableinsight, as Shah and Corley (2006) have recentlyargued. In the following sections, we describe thesampling, data sources, and data analysis.

3.1. Sampling

Our research design is a single-case study focuss-ing particular attention on sampling (see Eisen-hardt, 1989). There are three reasons for selectinga particular case: fit, distinctiveness, and its re-velatory nature (see also Yin, 2003; Siggelkow,2007). First, Nokia’s Internet Tablet developmentrepresents a case that both serves to explore andextend the private-collective model. The InternetTablet was created by private investments byNokia, building on existing open source productsas well as releasing a substantial amount ofknowledge (source code).

Second, the distinctiveness of the case is pro-vided through Nokia’s unique approach in pro-ducing a product based on open source software.While Nokia is not the first manufacturer tocreate a mobile device based on the open source

operating system Linux, its strategy of a conspic-uous commitment to open source software and itsdevotion to building a community of volunteersare unique for a multinational corporation in theconsumer electronics industry. In order to deter-mine the distinctiveness of the case, we comparedit with a wider sample including Sharp, Motorola,Hewlett-Packard, and Sony.

In the 1990s, Sharp introduced a PersonalDigital Assistant called Zaurus, which ran on anadapted Linux operating system. Soon after,voluntary developers programmed a variety ofapplications for the device and started a vividcommunity of developers (among them one of theinterviewees). However, in contrast to Nokia,Sharp did not reveal much of its own sourcecode, scarcely supported the community and itsinterests, and eventually stopped selling the devicein Europe and the United States.

Motorola brought Linux-based cell phones tothe market in 2003. The development of theiroperating system was carried out by MontaVista,a vendor of embedded Linux software, and Troll-Tech, the provider of the graphical user interface‘Qt.’ In contrast to Motorola, Nokia did not relyon a few service providers to implement the soft-ware but collaborated with many, mainly small,software firms in an open fashion. By using opensource software, Motorola expected to cut costsand speed up software development, because theydid not pay per-unit royalties and built applica-tion software on the existing open source software(Shankland and Charny, 2003). Similar to Sharp,however, Motorola neither revealed source codebeyond legal requirements nor did the firm pro-vide extensive developer documentation of thesoftware on its devices. Moreover, their Linuxappliances granted no administrator access to theuser, inhibiting the installation or modification ofnative applications. Other mobile devices runningembedded Linux included HP’s iPAQ and theSony Mylo. However, as we discovered in inter-views and in press articles, these companies re-tained the software’s source code and we couldfind no evidence that these firms attempted tobuild up a community of outside developers asproposed by the private-collective innovationmodel (von Hippel and von Krogh, 2003).

Third, given the research gap on the implemen-tation of private-collective innovation, we alsosearched for a revelatory case. The main criterionfor selecting a revelatory case is the researchers’access to a previously inaccessible setting forscientific observation. Establishing ties to Nokiaand the developer community surrounding the


r 2009 The Authors



Internet Tablet, the researchers gained access to avariety of data including documents, interviews,prototypes, and online conversations. Sheddinglight on the reasons for and the effects of thisinnovation project going open and abstractingthese underlying intentions into a model that canbe used in future research, as well as raising theattention of practitioners to this mode of productdevelopment, motivated the selection of the case.

3.2. Data sources

This study relies on several sources of data. Firstand most importantly, we conducted semi-struc-tured interviews, allowing participants the oppor-tunity to narrate stories, provide anecdotes, andstate opinions. Through an initial reading of themailing lists, relevant stakeholder groups anddata sources in the development project wereidentified as ‘Nokia employees,’ ‘Nokia-paid con-tractors,’ and ‘independent individuals.’ Inter-views were conducted with participants from allthe stakeholders identified. The initial partici-pants were selected from the developer and usermailing lists, and subsequent interviewees wereidentified through snowball sampling (Hecka-thorn, 1997). In total, 23 interviews were con-ducted; 10 with Nokia employees, five with

contractors, and eight with unpaid volunteers(see Table 1). In order to protect their privacy,interviewees were anonymized. The interviewslasted on average 75min. The interview guidelinesincluded questions on the firm–community rela-tionship, the strategies Nokia used to revealknowledge and technology to the community,motivation, and other issues (see the AppendixA for examples of two distinct interview guide-lines). The initial interview guidelines were up-dated and enhanced over time, integrating andbuilding upon the results of interviews alreadyanalyzed. All interviews were transcribed verba-tim and, using the software Max.QDA, codifiedusing an open coding technique (for a discussion,see also Strauss and Corbin, 1998). This led to thecreation of 80 codes, which were subsequentlymerged and reduced to 12 categories, includingseven incentive and five cost categories.

The second data source consisted of the pro-ject’s user and developer mailing list. Themonthly archives were downloaded from theirinception in May 2005 until the end of December2006. The archival data were examined using thestatistical software ‘R’ in order to indicate the sizeand activity of the community. The resultingstatistics are presented later in the text.

Third, one co-author followed the developermailing list over the course of several months

Table 1. List of interviews including identifier keys, role, date of the interview, duration, and the person’scontribution or function in the context of this case study

# Key Role Date Duration Contribution, function

1 N1 Nokia November 15, 2006 89min Head Open Source Software Operations at Nokia2 C1 Contractor November 22, 2006 71min Developed Window Manager3 V1 Volunteer December 6, 2006 55min Linux Distribution Release Manager4 V2 Volunteer December 7, 2006 52min Developed Mapping Software5 V3 Volunteer December 13, 2006 54min Developed Music Player6 C2 Contractor December 14, 2006 89min Performance Measurements and more7 V4 Volunteer December 15, 2006 79min Developed Swap Memory Feature8 N2 Nokia December 15, 2006 58min Maemo Product Manager9 N3 Nokia January 12, 2007 92min Software Architecture Team Leader at Nokia10 N4 Nokia February 19, 2007 83min GNOME Desktop Developer at Nokia11 V5 Volunteer February 20, 2007 85min Ported Remote Control Software12 N5 Nokia February 20, 2007 96min GNOME Desktop Developer for Nokia13 N6 Nokia February 28, 2007 80min Multimedia Player Developer for Nokia14 C3 Contractor March 5, 2007 81min GNOME Cþ þ Bindings Developer15 C4 Contractor March 13, 2007 email Software Developer at Contracted Firm16 N7 Nokia March 21, 2007 60min GNOME Desktop Developer for Nokia17 V6 Volunteer April 4, 2007 71min Developed Geolocation Software18 C5 Contractor April 10, 2007 email CEO of Contracted Software Company19 N8 Nokia April 11, 2007 69min X Windows Developer for Nokia20 N9 Nokia April 12, 2007 77min Maemo Community Manager21 N10 Nokia April 12, 2007 91min Testing Team at Nokia, Volunteer in Browser Project22 V7 Volunteer April 15, 2007 email Linux kernel patching for Maemo23 V8 Volunteer April 23, 2007 66min GNOME Foundation Board Member




and observed the project’s chat channel on IRC(Internet Relay Chat), which is not publiclyarchived. Often, informal discussion takes placeon IRC, giving the researchers a feeling for what‘really happens’ in the community. Such onlineparticipant observation is rather uncommon inthe research on open source software develop-ment but is deemed both necessary and helpful inunderstanding the unfolding dynamics of theInternet Tablet development. No formal analysiswas performed with the data but being immersedin the community helped to interpret mails andunderstand the issues raised in the interviews.

Fourth, secondary sources, such as news re-ports, blogs of Nokia members and volunteers,and corporate web sites of Nokia and contractingfirms were included in the case study database.For example, these sources provided additionalinformation on the extent to which Nokia spon-sors other open source projects. Some Web arti-cles were used to get background information onthe Nokia device and potential competitors. Wealso searched an independent web forum fordiscussions and opinions of users of the InternetTablet.

4. Nokia and the development of theInternet Tablet

In this section, we present a short overview of thehistory of Nokia and the Internet Tablet devel-opment, and we provide a descriptive analysis ofMaemo, the community for the Internet Tabletsoftware platform. The purpose of this analysis isto confirm the correctness of the case to examinethe implementation of the private-collectivemodel for innovation incentives.

Nokia was originally set up in 1865, producingpulp and paper. It underwent a series of remark-able transformations in its business model. In1967, it merged with the Finnish Rubber WorksLtd. and the Finnish Cable Works, forming theNokia Corporation with four major businesses:forestry, rubber, cable, and electronics. A diversi-fied company, with a product portfolio rangingfrom tires to television sets, it first started produ-cing mobile phones in 1981, manufacturing carphones for the first international cellular mobilephone network. The first hand-portable phonesets were introduced in 1987. During the 1990s,Nokia focused on telecommunications, especiallyon mobile phones based on the then emergingGSM standard, which had been published by theEuropean Telecommunications Standards Insti-

tute (ETSI) in 1990. During this period, thecompany also divested its other businesses. In1998, Nokia overtook Motorola as the world’slargest mobile phone manufacturer (ICFAI,2005). R&D operations at Nokia have alwaysbeen scattered across the world, working in adispersed, non-hierarchical structure, allegedly toprevent the development of ‘tunnel vision.’ In2003 and 2004, Nokia suffered a decline in themarket share of its mobile phone business. Thecompany had misinterpreted the market demandfor ‘clamshell’ devices and camera phones andhad failed to adapt fast enough to these newdevelopments (ICFAI, 2005). However, in 2007,Nokia posted EUR 51.1 bn of net sales and anoperating profit of EUR 8.0 bn, spending EUR5.6 bn on research and development.

4.1. Internet Tablet History

According to N1, Nokia started to experimentwith incorporating open source products, specifi-cally based on the Linux kernel, into their devicesin 2000. At the same time, the company sought todevelop a device that would take advantage of theincreasing availability of wireless access pointsand provide access to Internet appliances every-where. N1, the head of the software developmentteam, summarized the vision for this new type ofmobile device as follows:

At the same time, totally independently, we hadanother stream of thought which was this kindof category of Internet Tablets. The big ideabehind that was really the same way mobilephones liberated voice. Not only houses oroffices have phone numbers, but people havenumbers. So you can take the phone whereveryou go. We have the same vision that we wantto do the same thing with the Internet andInternet use cases. You don’t need to fire up aPC and you don’t need to go to your desk. Youhave this very light portable device that givesyou access to the Internet. Whether that isbrowsing, email, chat, VoIP. (N1)

Inspired by this vision, Nokia designed an overallsoftware architecture of the operating systembased on open source components and partlyadapted these components themselves and partlycontracted developers for specific implementationtasks. In 2002, individuals who were active inarchitecture-crucial open source projects wereapproached by Nokia and asked to perform


r 2009 The Authors



some tasks as contractors for the company (e.g.,C1, C2, and C3). They had to sign a non-dis-closure agreement (NDA) that prevented theleaking of much information until the InternetTablet went public.

The prototype device was first publicized onMay 25, 2005. At the GNOME User and Devel-oper European Conference on May 31, 2005,Nokia announced that it would give away 500devices for about a third of the regular sales priceto selected software developers and donate thesales to the GNOME foundation, a not-for-profitorganization dedicated to supporting theGNOME graphical desktop environment (Nokia,2005). The sales of the product did not commenceuntil November 2005, when the 770 InternetTablet was officially released with a price ofUSD 439. Customers could order the productvia a Nokia Internet Tablet-dedicated Web site orthrough the official Nokia Web shop. However,little effort went into promoting the InternetTablet to the wider public. The product was notavailable through other distribution channelssuch as local Nokia mobile phone shops1.

While the Internet Tablet 770 was still beingsold throughout 2006, its successor was an-nounced by Nokia’s CEO Olli-Pekka Kallasvuoon January 8, 2007, at the International Consu-mer Electronics Association show in Las Vegas.The successor product was added to the N-seriesof Nokia devices, a popular brand for Nokia’smajor communication products, and marketed tothe ‘mainstream’ public as the N800. On October17, 2007, the third generation of the InternetTablet, the N810, was presented in San Franciscoat the Web 2.0 Summit. Nokia executive VP andhead of the company’s multimedia business unit,Anssi Vanjoki, stated ‘The N810 is the first ofthese devices targeted at a ‘normal’ consumergroup, beyond the geeks’ (Martin, 2007).

The hardware of the Internet Tablet differs inone main aspect from other Nokia products: itdoes not contain mobile phone functionality. Itoffers a 4.1300 display with – given its size – anunusually high resolution of 800 � 480 pixels,which, using a stylus, can be utilized as a touchscreen. It connects wireless through Bluetooth(connecting to a mobile phone) or through acommon WiFi to the Internet. It is also possibleto connect to another PC through the integratedUSB port. The device has no hard disk, but flash-based storage is included that can be extendedwith external flash storage media. Since the N800,a VGA webcam has been integrated and in theN810 a GPS receiver is also included.

4.2. The Maemo Community

The operating system and the software of theNokia Internet Tablet are based on the Maemosoftware platform, an effort led by Nokia andannounced the same day the device was launched.The intention behind the platform was to provideopen source components usually deployed onLinux desktop distributions and to adapt andenhance these for the environment of handhelddevices (Maemo.org, 2006). Nokia had down-loaded the open source GIMP Toolkit (GTK)graphical toolkit and other components such asthe GStreamer framework for multimedia andmodified these to fit the needs of an embeddeddevice with restricted hardware resources. Nokiaalso added their own software developments andparts from independent software vendors as pro-prietary software, protected by commercial soft-ware licenses and released as binary code only.Figure 1 provides an overview of the softwarearchitecture distinguishing between software pub-lished under an open source software license,commercial software components published bythird-party vendors, and Nokia’s proprietarycomponents.

The Maemo platform uses its own infrastruc-ture, such as a revision control system, a softwarebug tracker, and mailing lists that allow commu-nication between developers and users. In No-vember 2006, the Maemo.org site had 54,000unique visitors. While the operating system con-tains proprietary software (some hardware dri-vers and applications such as the Opera Webbrowser), the main software platform, Maemo,was open source and developed by the Maemocommunity.

A source code repository allows developers toadd new software components to the product and

Figure 1. Maemo software stack.




upload improved software that resolves problemsand bugs in previous versions. The Maemo sourcecode has its own repository of source code. Asof January 2007, our descriptive analysis showsthat 33 developers added more than 7.2 millionlines of code to this repository, forming thecore of the operating system (although much ofthis is unmodified code from other open sourceprojects). Figure 2 visualizes growth in thesource code over time. It should be noted thatthe developers were exclusively employees ofNokia or formally affiliated with the Maemoproject, indicating that Nokia retains tight con-trol over the actual changes that occur to the coresystem in the software architecture (see also Kuk,2006).

In order to help identify and remove softwarebugs, Maemo also has a so-called ‘bug tracker’ ofits own. This is used to enter software errors, orbugs, and keep track of the bug-fixing process.Our analysis shows that in June 2008, this trackercontained 3228 bugs, of which 1133 bugs weremarked as ‘open,’ meaning that they were waitingto be fixed.

The Maemo project infrastructure also offers arepository, garage.maemo.org, where people in-dependently of Nokia can register their projectsand use the developer infrastructure for free. Ouranalysis shows that in June 2008, Garage con-sisted of 615 active projects and 12,446 registeredusers. Projects led and contributed to by indepen-dent individuals ranged from GPS navigation

solutions and system administration tools to anElectronic Flight Information System for use withsmall aircrafts.

Most communication occurs through mailinglists. For discussion purposes, Maemo has adeveloper list where actual technical developmentissues are discussed and a user mailing list fordiscussing issues related to the operation and useof the product. We examined the Maemo devel-oper mailing list from its inception in June 2005until December 2006. In total, 832 participantsidentified by their email address contributed tothe list, with 79 participants (9.5%) posting froman official @Nokia.com or @maemo.org emailaddress. The monthly number of postings to thelist ranged from 98 to 548, with a median of 328(mean 339.8). 19.6% of all mails were sent byNokia addresses. Out of the top 15 posters, onlyfour had an official email address from Nokia/Maemo. Communication on mailing lists is orga-nized in so-called ‘threads.’ Often, a thread isstarted by someone asking a technical question,followed by replies by others who attempt todiscuss and answer the question. One hundredand eighty three threads were started by Nokiaaffiliates, while 931 mails were replies to anexisting thread. Non-Nokia participants, on theother hand, started 1670 threads and posted 4010replies. A w2 test of the frequency contingencytable confirms that Nokia participants differ sig-nificantly in the thread start/reply frequenciesfrom non-Nokia participants: Nokia members

Figure 2. Lines of code in the Maemo Subversion code repository.


r 2009 The Authors



were more likely to post a reply to an existingthread rather than start a new one.

Besides discussing development, there is also aMaemo user mailing list that serves to supportgeneral user discussion. This list follows similarpatterns although it is less frequented than thedeveloper mailing list. Its number of postings permonth ranged from three to 253, with a median of133 and a mean of 127 mails. As can be expected,user-related discussions started to pick up afterthe device was introduced to the market in No-vember 2005, whereas development discussionspreceded this point. Our analysis shows that 511participants contributed to the list, 6.46% of themfrom Nokia. Only two out of the top 15 postersidentified themselves as belonging to Nokia.10.9% of all mails were sent from Nokia employ-ees. These started 29 new threads and replied to220 existing threads. Non-Nokia participants, onthe other hand, started 775 new threads and sent1510 mails to existing ones. Also on the usermailing list, a w2 test confirmed that Nokiaaffiliates were significantly more likely than non-Nokia participants to reply to existing threadsrather than start new ones.

Our brief examination of the Maemo commu-nity, including source code development andtechnical and user discussions, supports the state-ment that the development of the Nokia InternetTablet is an implementation of private-collectiveinnovation. Large parts of the product are apublic goods innovation and Nokia employeesas well as external individuals and organizationsexpend considerable private resources (time,knowledge, and technology) to contribute to theinnovation. In the mailing list geared towardsdevelopment, more than 80% of all emails weresent by non-Nokia affiliates. In the next section,we present the findings from the case study.

5. Findings

The following section presents findings on theimplementation of private-collective innovationin the case of the Nokia Internet Tablet develop-ment. The aim of this section is twofold: first, weillustrate a case of private-collective innovationwith empirical data, providing specific benefits forthe company involved. The findings confirm thesix conjectures on the benefits derived above fromthe existing literature and additionally identify abenefit in the case: faster time-to-market. Inaddition, we extend the model of private-collec-tive innovation, highlighting the hidden costs

related to the implementation of private-collectiveinnovation and strategies to mitigate these costs.The benefits and costs of this extended model aresummarized in Tables 2 and 3.

5.1. Benefits in implementation

First, considering the reduced cost of knowledgeprotection, product development managers atNokia were conscious that costs stemming fromprotecting proprietary knowledge and costs ofvoluntarily revealed knowledge imply a trade-off. For the Internet Tablet platform, Nokiadecided to keep some components proprietary,while developing other components as opensource software. According to the interview withN1, about 25% of the software is unmodifiedopen source software. Another 50% consists ofexisting open source code to which Nokia madeimprovements or adaptations and that was againreleased under an open source license. Only 25%comprises closed code that Nokia either imple-mented from scratch or that is closed commercial(proprietary) software from a commercial vendor.N1 stated that one reason for not releasing sourcecode under an open source license was that Nokiawanted to keep control of the look and behaviorof end-user applications, rather than fearing a lossof valuable knowledge. Disclosing the knowledge

Table 2. Benefits in the implementation of Private-Collective Innovation and findings in the Nokia case

Benefits Findings in the Nokia case

Low knowledgeprotection costs

Revealing source code ratherthan protecting it; however,undetermined costs for revealing.

Learning effects Collaboration with external firmsand individuals

Reputation gain Increased attraction of Nokia asan employer and for buildingtheir own developer community

Adoption ofinnovation

Standard setting of the platformconfiguration

Increased innovationat lower costs

Reuse of open source software,outsourcing of software testingand bug fixing and maintenanceto open source communities.Experimentation andcontributions of new applicationsby lead users

Lowermanufacturing costs

No licensing fees for softwareplatform

NEW: Faster time-to-market

Tapping of distributedtechnology expertise and highflexibility of software platform




was not free either; all source codes had to bechecked by Nokia’s department to ensure thatNokia did not reveal intellectual property orpatents they did not have the right to. Altogether,it remains unclear whether Nokia saved costs byrevealing knowledge rather than spending effortto protect it. Nokia seems to see it as beneficial:

We have evidence that some of our competi-tors are now looking at our code and they areinvestigating if they could use our code in theirproducts. You might say that we help themnow to get their products out fast. [. . .] But ifwe had not put it out then we could not haveused the OSS communities who have alreadyhelped us to develop that code. (N1)

Second, Nokia obtained learning effects by ex-perimenting with open source software technolo-gies and by collaborating with contractors, firms,and volunteers. The platform had started as aresearch project within Nokia, seeking to explore

new possibilities with open source components inorder to build new competences. Although thecompany could have chosen a commercial opensource software integrator such as Red Hat orMontavista, delivering ready-made solutions inorder to release a new product quickly, Nokiadecided to learn about the novel technologiesinternally in order to gain the potential for futureinnovations. N3 reported:

[We wanted to make] something that is morethan just putting a product into the marketthrough learning about the possibilities ofleveraging open source in deeper and moresubstantial ways. As a result, we didn’t takethe quick and simple approach, we went a littlebit deeper and we learned more.

Building upon technologies developed mainly byexternal contractors and volunteers implied thatmany outsiders, rather than the internal researchand development staff, had the most intimateknowledge of certain software components. Theinterview with contractor C2 revealed that in thebeginning, Nokia attempted to solve technicalissues internally. However, when the companyfaced severe time pressure on a product launchdate, it would outsource the request to a trustedexternal firm that specialized in the area oftechnical development. For specific parts of theInternet Tablet software, Nokia contracted sev-eral small enterprises with experience in opensource components (see Table 4). Working withthese experts and listening to feedback from theMaemo community helped Nokia to rapidly learnnew open source technologies and how those wereproduced by their communities.

Nokia’s strategy was deliberately focused onsmall companies, rather than on a few, largecontractors. Collaborating with small firms allover Europe enabled Nokia to fill gaps in thecompany-internal technical knowledge needed fordevelopment, while still retaining control of the

Table 3. Costs and possible mitigation strategies in theimplementation of Private-Collective Innovation

Cost Findings in theNokia case

Mitigation strat-egy

Difficulty todifferentiate

Released sourcecode can be reusedby competitors

Partial revealingof source codeto retain controlof look and feel

Guardingbusinesssecrets

Plans for new pro-ducts

Selective reveal-ing of futureplans and pro-tection of infor-mation throughNDAs

Reducingcommunityentrybarriers

Investments forSoftware Develop-ment Kit, previewversion of platform,device program,staff for communitymanagement, andincreased commu-nication effort

Sharing thecosts with otheractors in thecommunity

Giving upcontrol

Development direc-tion such as scopeof functionality ofopen source pro-jects is controlledby external parties

Hiring of keydevelopers andparticipation inupstream com-munities. Nosingle vendorcontrols plat-form

Organizationalinertia

Required internalrestructuring ofprocesses

Adapting andopening up pro-cesses

Table 4. Some contractors to Nokia in the softwaredevelopment process

Company Expertise Country

KernelConcepts GPE and Embedded Linux GermanyOpenedHand Matchbox UKCollabora Telepathy UKImendio GNOME and D-BUS SwedenFluendo GStreamer SpainMovial Scratchbox Finland


r 2009 The Authors



core software architecture by having internalexperts in these areas (C2).

In addition to learning about the technologiesin use and acquiring the skills to work with them,Nokia learned to cooperate with a diverse com-munity of employees, volunteers, and contractorsrather than relying only on contractors who couldbe forced to keep schedules and timelines:

It’s all about the process . . . You develop thisopenly within the communities and you try tosynchronize your own work with the heartbeatof the communities. Some companies nowunderstand this better than others. We cer-tainly have done our learning. We have madesome mistakes too on this front. (N1)

For example, one of Nokia’s major learningsconcerned the early decision to independentlycontinue development of the GTK, a collectionof software constituting the core of the graphicalinterface. Among others, C3 stated ‘the disadvan-tage of doing this of having a forked or largepatch was experienced by Nokia.’ By following aseparate development stream, Nokia became dis-connected from the code maintenance effort bythe community and thus realized they had tomove their changes back into the main opensource project as architecture team leader, N3,reported:

It was also a learning field that we had to gothrough. If we had known then what we knowtoday, we would have been able to do it with-out such a large patch. We would have beenable to do more directly upstream in techni-cally better ways with less effort for changingcode.

Third, the growing commitment to open sourcesoftware development led to a reputation gain forNokia, and interviewees V2, V3, and V4 sug-gested that this led to the increasing attachmentof volunteers and recruiting benefits. Particularly,the openness of the Maemo platform encouragedvolunteer users and developers to buy such adevice, to improve the operating system, and tocreate applications on top of it. Nokia wasaccredited with pioneering development of anopen embedded platform by users who enjoyedrunning applications from third parties as forexample V6 stated:

Yeah, I definitely have a more positive view ofNokia this way. Especially, I think they are

handling the open source interaction quitewell. I think they are quite a good open sourcecitizen.

For Nokia, recruiting active open source contri-butors enabled them to select the best individualsbased on their prior contribution to the variousprojects that constituted the development of theInternet Tablet. In some cases, Nokia directlycontacted skilled open source software developersand invited them to apply for jobs. Thus, most ofthe Nokia employees working on the InternetTablet were previously active participants inopen source projects and known to the Nokiamanagers through prior collaboration. This factreflects that for many, Nokia became an interest-ing company to work for, as C1 testified:

At that time, it was a dream come true. I didnot have a good job. I was spending all myspare time hacking Matchbox [window man-ager of the Internet Tablet]. I was reallyenjoying it. And then you’re given a chanceto get paid to do that full time. It was prettyfantastic and an amazing piece of luck.

According to interviews with Nokia developmentmanagers, Nokia employees selected through thecommunity were highly motivated to continue towork on technologies they already knew. Possiblyintrinsic motivation, such as fun – often a primarycause of contributions to open source develop-ment – played a role in their continued high-levelefforts (see e.g., Torvalds and Diamond, 2001;Luthiger Stoll, 2006).

Fourth, the private-collective innovation modelproposes that being the first to contribute a publicgoods innovation increases the likelihood of fastand widespread adoption of the innovation. Ac-cording to the interviews, because the communitywas already familiar with underlying technolo-gies, adapting existing applications from otherprojects to the Internet Tablet platform provedto be a relatively easy task. Nokia also invitedcompetitors to participate in the creation and useof their platform, citing a ‘the more the merrierapproach.’ By initiating a vendor-independentembedded software platform intended for use inother mobile devices, Nokia made it easier forvolunteers, contractors, and competitors to con-tribute. Spreading the innovation and invitingothers to participate was seen as crucial:

We believe the world is changing and thecompetitive advantage comes from how many




others you can get to participate in this net-work. (N1)

In fact, in July 2007, Intel announced that theywere adopting the Internet Tablet’s user interfaceframework Hildon into their new product cate-gory called Mobile Internet Devices (Paul, 2007),which will eventually lead to a higher developerand user basis of Nokia’s Maemo platform.

Fifth, by contributing to public goods innova-tions, firms can lower the cost of innovation.Building on existing and mature technologiesthat could be integrated into the new hardware,Nokia enabled the development of a solid, yetcheap operating system for embedded platforms.Collaborating within existing open source pro-jects allowed Nokia to benefit from the collectiveprogramming efforts:

So what is happing in the D-BUS, in the GTK,in the GStreamer, in the Linux kernel is that Iput two guys there, IBM puts two guys,Motorola maybe puts one guy, maybe Novellputs a couple of guys. So for the price of twoguys, I get four or six guys working on thesame problem. (N1)

Additionally, the company benefited from volun-tary contributions leading to enhancements of thedevice. The volunteers contributed several inno-vations including applications, user interface im-provements, translations, bug reports and fixes,testing different peripherals, and making featurerequests. Allowing volunteers experiment with thesoftware also created ‘proof of concepts’ whichenabled innovations previously seen as unrealisticby Nokia engineers (V3). One example of this isthe swap memory enhancement (using the flashmemory as extended virtual memory) which wasinitiated by volunteer V4 and included in a sub-sequent official version of the Internet Tablet’soperating system. Thus, the likelihood of findinga ‘killer application’ (see e.g., Downes and Mui,1998) in the process (by evolution or sheer luck)increased. An open source software developerobserved:

I think from my point, if you let people changethings [. . .] and document them and open themup so people can hack their own stuff, younever know what is going to happen, whatkind of things people are going to write foryour device which ultimately could make it sellmillions and millions if someone writes thekiller application for it. (C1)

As such, both the costs of innovation were keptlow and ideas that would not have been devel-oped otherwise could be tested and integrated:

I can develop, say, twenty ideas a day and thiscommunity can develop a hundred ideas a day.So it’s more important to be part of thecommunity with a hundred ideas than byyourself with twenty ideas. (N1)

Investigating the software architecture, we foundthat most contributions by volunteers were sepa-rate applications that could be installed indepen-dently from the core operating system. Theinterviews showed that volunteers who madesuch contributions showed high commitment toand responsibility for their work, listening to userfeedback and, in some cases, when others de-manded it, even enhancing their software againsttheir personal belief of the usefulness of thefeatures (explained by V2). Through their con-tributions and feedback from users, volunteersslowly gravitated towards more developmentwork in the community.

Sixth, supply by anyone of public goods in-novations to the market enables manufacturers tolearn about innovations and thereby reduce costsin manufacturing. This conjecture indicates parti-cular benefits that private-collective innovation insoftware offers to computer hardware manufac-turers (von Hippel and von Krogh, 2003).Through choosing a software platform that isavailable under an open source license as indi-cated, Nokia reduced fixed costs related to re-search and product development. Nokiamanufactured and sold the Internet Tablet, butthe product’s functionality and, thus, ability tofulfill user needs were, to a large extent, shaped bythe users themselves. User-developed applica-tions, such as mapping and navigation software,could easily be installed by the end users them-selves for free, keeping Nokia’s costs down.Interestingly, in addition to fixed cost reduction,Maemo also has a positive impact on variablecosts in manufacturing because Nokia did nothave to pay a per-device license fee to an intellec-tual property owner. For example, in the begin-ning of 2006, a comparable proprietary operatingsystem, Symbian, demanded USD 7.5 per devicefor the first 2 million units.

In addition to confirming theoretical conjec-tures on benefits, one more benefit emerged in thestudy that should be considered crucial for pri-vate-collective innovation: faster time-to-market.Using external, modular technologies not only


r 2009 The Authors



impacted on costs, but it also led to the creationof a new operating system working on a newhardware platform in a short time. According tointerviews with N1 and N3, this fast developmentcreated flexibility, which, combined with userfeedback of pre-releases, allowed for a quicktime-to-market compared with other devices thecompany had launched.

However, integration of the numerous opensource components into an executable environ-ment is very challenging as, for example, C3explained. One way to rapidly tap into thisknowledge of open source communities wasachieved by Nokia’s strategy of contractingopen source developers and small firms – the‘bridges between Nokia and the communities’(N3):

It’ll get done quicker and probably better ifthey pay us to do it. [. . .] Over the years wehave been involved, we have so much experi-ence and knowledge, we know all the tricks.We know how to get things like X and Match-box up and running quickly on hardware. Weare just basically selling that knowledge aswell. Although they could very well likelyfigure it out themselves, we just can get themthere a lot quicker. (C1)

5.2. Costs of implementation andstrategies to mitigate these

While the benefits of private-collective innovationhave been spelled out previously, the hidden costsof implementing the model have been neglected inprevious work or remain unknown. In this sec-tion, we present the findings from the case studyalong five categories of costs, together with No-kia’s strategies (where applicable) to mitigatethese. First, when software is freely availableeven to direct competitors, it is possible forcurrent and future competitors to design clonesthat look like and behave in a manner very similarto the original product. Competitors are in theposition not only to imitate but to replicate theproduct (see Kogut and Zander, 1992). Thispotential lack of differentiation of products repre-sents a cost to the firm as it forfeits an opportu-nity to gain competitive advantage (Granstrand,1999). Such reuse of open source software iscompletely legal and, in fact, took place as ex-plained in this illustrative example:

The OpenMoko actually bases on our stuff.[. . .] the Moko window is a kind of basewindow for their applications I think. Therein the comments of the source file you can findthat this is based on the Hildon window byNokia. They even have this copyright ‘Nokiacorporation.’ So they are based on our stuff.But, on the other hand, our stuff is not thatrevolutionary. I mean that’s how it goes. Webase on somebody else’s stuff too. (N4)

Nokia’s strategy to mitigate this cost was toselectively open up its software development.Figure 1 visualized how Nokia revealed the mid-dle layer of software, the basic infrastructure,under an open source license, while keeping partsof the bottom layer (hardware specific software)and much of the user-visible applications undertheir own proprietary license. N1 argued that itwanted to ensure a unique ‘Nokia look’ byretaining control over applications, for exampleof the device’s email program. In doing so, theykept crucial parts closed (e.g., power managementand other hardware drivers) and prevented repli-cation of these parts by competitors. In all theinterviews conducted with volunteers, the respon-dents demonstrated understanding for Nokia’sdecision to selectively reveal source code,although some stated their strong preference forreleasing all software under an open sourcelicense.

Second, Nokia was concerned with potentialcosts stemming from losing business secrets, suchas plans for future devices. In order to mitigatethese costs, Nokia revealed knowledge in terms ofsoftware but kept future key product innovationsand business figures, such as devices sold, numberof employees, or investments expended, confiden-tial. For the development of new software fea-tures, they contracted several small enterprisesor even motivated individual volunteers withunique knowledge critical to the developmentof the Internet Tablet to sign an NDA. TheNDA protected Nokia against the leakage ofinformation about planned new product develop-ments. The agreements effectively created athree-tiered community of people ‘who knew,’which meant Nokia insiders, ‘those with a clue,’which meant contractors, and the ‘regulars’ (C3).However, this information imbalance led tostrong tensions for Nokia with its external com-munity members. Thus, Nokia hired N9 e.g., inorder to update the technical roadmap of theInternet Tablet with all the information necessaryfor software developers.




Third, in order to facilitate the increasinginvolvement of volunteers, Nokia needed to carrythe costs of reducing community entry barriers.The company invested in the creation of a Soft-ware Development Kit that enabled new volun-teers to easily start development for the InternetTablets. It is a common practice for softwaremanufacturers to provide such an SDK at highcosts (Jacobson et al., 1999). However, in the caseof the Internet Tablet, Nokia offered the devel-opment tools for free. In order to allow volunteersto adapt their software for upcoming platformreleases, Nokia also offered a development snap-shot of their work-in-progress (often includingsoftware for yet unannounced features) whichcould be used to ensure that an application wouldalso run on future releases. Employees were sentto related conferences in order to increase aware-ness of the platform and answer questions fromcurrent and future volunteers and contractors.Additional staff, such as ‘community representa-tive’ N9, were hired in order to communicatebetween Nokia internal developers and the ex-ternal community members. In order to mobilizemore volunteers to join the Maemo community,Nokia sold 1500 heavily subsidized devices toactive open source developers. While such directcosts by Nokia cannot be mitigated easily, invest-ments in community building, knowledge diffu-sion, and marketing may be lower in the futurethrough sharing the effort with other communitymembers. Some interviewees explained that, forexample, with increasing popularity, well-inte-grated Maemo community members started tosupport new volunteers who were becoming in-volved in the development process.

Fourth, by contributing source code to opensource projects that were not managed by thecompany, Nokia gave up control of the futuredevelopment direction of core technologies de-ployed in the company’s hardware. According toN1, the company traded having full control of thetechnology for participation in joint development,thus benefiting from sharing the cost of innova-tion with outsiders. For example, GTK wasoriginally intended for use on desktop PCs, andaccording to interviewees N1, C1, V2, V4, andN3, it was necessary to adapt GTK to the low-resource environment of the Nokia Internet Ta-blets by decreasing memory consumption. Inorder to regain some control of these criticalsoftware components, Nokia hired key developersfrom the GTK community and contracted smallenterprises with in-depth knowledge in this area.According to N3, the contribution of code im-

provements and modifications, combined with ameritocratic organization of the projects in-volved, gave Nokia enough influence on thedirection of software development. Yet, the prac-tice of contracting developers from incumbentcommunities such as GNOME in order to gainreputation and control raised concerns from someMaemo community members as to how Nokiawould influence the future of the projects. Forexample:

Obviously when they are sponsoring a project,then they are going to have some control overthe direction and what gets into it. [. . .] I think,they truly want to work with the communityand want to keep them happy as well, so it’s alla bit of give and take, I suppose. (C1)

The situation was different for projects that wereinitiated and controlled by Nokia itself. Accord-ing to C3, Nokia granted only write-access to itssoftware repository to its employees, thus retain-ing control of the actual published source code. Itis worth noting that this strategy created sometensions in the community, with the non-Nokiamembers complaining that they could not help orcontribute if they did not know about the futuredirection of the software development or aboutthe estimated timeframes of future software re-leases. The Maemo community attempted toinfluence Nokia’s behavior, mostly through pro-viding intense feedback on mailing lists. Forinstance, in 2000, Nokia released a new versionof the Maemo platform as a binary download asV4 remembers. However, Nokia did not publishthe source code at the same time but staved offthe community by arguing that the legal depart-ment needed three more weeks to clear thecode for reasons of intellectual property protec-tion. A strong, negative reaction from the com-munity taught Nokia to proceed differently nexttime, releasing both binary and source codesimultaneously. Sanctions on the part of thecommunity were mainly ‘withdrawal of love’and the ‘threat of forking.2’ When attempting tobalance control and openness, Nokia consideredthe threat of defection of volunteers to other openplatforms used by emerging competing ‘opendevices.’

Fifth, because the private-collective model ofinnovation incentives breaks with the traditionalprivate-investment model that is prevalent inindustry, it is reasonable to expect that theimplementation of the model in an establishedfirm incurs costs of organizational inertia


r 2009 The Authors



(see Sorenson and Stuart, 2000). Because theInternet Tablets also include software writtenby third-party vendors, Nokia needed to ensurethat their software revealed to the Maemo com-munity did not infringe on intellectual propertyrights. As mentioned above, interviewees com-mented that Nokia’s internal approval of opensource software was slow and bureaucratic. Inaddition, the complex internal processes of alarge multinational organization often made itdifficult for Nokia’s developers to collaboratewith external open source projects. One intervie-wee commented:

The biggest problem is that Nokia is a very bigcompany and that Maemo is a very smallgroup. I think it’s like ten or twenty people intotal working on it. A lot of software whichthey use on the 770 is developed in some othergroups in Nokia. For example, the movie plug-ins and MP3 plug-ins. Those are the same Ithink as they use on their phones in Symbian.So that’s a Symbian decision. You can’t tell theSymbian people to use that bug system insteadof the internal bug system. They would say,‘Why?’ (V1)

Nokia employees also commented that the inter-nal Nokia firewall would not allow connection tothe official developer chat room of the Maemocommunity, which was located outside Nokia’snetwork. In order to mitigate the costs of orga-nizational inertia, Nokia employees stated thatthey would work on a case-by-case basis toremove obstacles. Employees were, for example,assigned to make sure that entries in the external,publicly accessible bug tracking system would bepaid attention to.

Altogether, Nokia managers were conscious ofthe trade-offs between revealing knowledge andtechnology and the benefits from participating inprivate-collective innovation. The following state-ment summarizes well the experiences regardingthe trade-offs between cost and benefits in themodel:

Some people might say that one of the pro-blems is that you are leaking and giving outyour secrets and so forth, but it’s more like atrade-off. What is more important to you: togive some of your secrets an internal work-outor how much help in creating these productsyou get for free. I think, if you calculate, youare far more on the positive side when youdecide to share. (N1)

6. Discussion and conclusion

In this paper, we identified a research gap in theliterature on private-collective innovation (vonHippel and von Krogh, 2003, 2006): little isknown about the implementation by firms of theprivate-collective model of innovation incentives.We argued that the implementation of the modelwill be associated with benefits, ‘hidden’ costs,and strategies to mitigate these. In order toexamine and extend the model through empiricalwork, we used a case study design. Using quanti-tative and qualitative data, we demonstrated thatthe development of the Internet Tablet is a case ofprivate-collective innovation. Next, we analyzeddata from several sources in order to identify thebenefits and costs incurred in the implementationof private-collective innovation and the strategiesby Nokia to mitigate these costs.

Nokia launched the Internet Tablet as a pri-vate-collective innovation project and as a low-cost probe (Brown and Eisenhardt, 1997). At thetime of product launch, neither a product cate-gory nor a market for these devices existed.Rather than following existing market demand,Nokia targeted technology pioneers to find outwho would use the Internet Tablet and how itwould be used in real-life applications (similar towhat Zander and Zander, 2005, called ‘exploitingthe inside track’). Nokia opened up the product’ssoftware using externally developed open sourcetechnologies, allowed for and encouraged contri-butions by outsiders, and in the process created anew market for a product it had envisioned.When the product proved successful, Nokiamoved from targeting technology pioneers to-wards the mainstream market with the subse-quent release of the Internet Tablet N800 andN810.

This study confirmed most incentives to inno-vate identified in previous literature (von Hippeland von Krogh, 2003, 2006). It remains incon-clusive whether Nokia saved knowledge protec-tion costs by revealing most of their software.However, the company gained skills and knowl-edge through collaborating with outside volun-teers and contractors and thus also acted as asystem integrator coordinating a loosely couplednetwork of component providers (see also Bru-soni et al., 2001). As the main contributor in theproject, Nokia enjoyed reputation benefits bothas an attractive employer and as an ‘open source-friendly company’ amongst open source softwaredevelopers who contributed software to theplatform. By creating the vendor-independent




GNOME embedded platform and inviting com-petitors to contribute and use the software plat-form, Nokia facilitated the adoption of thesoftware as a common platform for embeddeddevices. Moreover, Nokia built upon existingtechnology and took advantage of users’ previouscontributions to open source software projects.By taking advantage of existing open sourcecomponents, Nokia managed to create a completeoperating system with only a handful of devel-opers and was able to integrate ideas and im-provements from other Maemo communitymembers. In terms of costs, Nokia’s manufactur-ing could benefit from low-cost software devel-opment and avoided paying the commonper-device license fees. Finally, our study foundthat increased flexibility and a faster time-to-market is a benefit in implementing private-collective innovation.

Thus, although the six benefits at first glancemake it rational for the firm to choose private-collective innovation among alternative models,previous work also raised the awareness of unin-tended consequences or ‘hidden costs’ resultingfrom implementing private-collective innovation.For example, in order to obtain outside contribu-tions, a firm may need substantial investments indocumenting the released software, training po-tential contributors, and developing online tutor-ials. These costs of implementation may offset thebenefits to private-collective innovation.3

We briefly reviewed literature that indicated‘hidden costs’ associated with the implementationof novel innovation models (Crawford, 1992;Kessler et al., 2000; Smith, 2004). The studyfound that the implementation of the private-collective model of innovation incentives in No-kia’s development of the Internet Tablet incurredcosts and that the company found strategies tomitigate these. In particular, the potential lack ofproduct differentiation as well as revealed busi-ness secrets incurred costs to the company. Nokiamitigated these costs by selectively revealingknowledge and technology. Another cost con-cerns the lowering of entry barriers to the Maemocommunity. The company invested in severalmeasures to reduce such barriers, including dis-counted devices and a free SDK, as well asallocating employees responsible for communitycommunication in order to attract further volun-teers. Moreover, using technologies that arepartly maintained externally has the advantageof shared innovation costs but implies giving upfull control of the future development of thattechnology. Through hiring key developers of

software for the product, Nokia regained someinfluence and control. Internal processes some-times proved inadequate to enable a transparentand open development process, incurring somedelays and costs as well as frustration in thecommunity. Nokia acknowledged this challengethat the interviewees described as a ‘learningprocess.’

The extended model of private-collective inno-vation provides additional insights for research-ers. First, implementing private-collectiveinnovation may enhance organizational learningand renewal, in addition to being a form of ‘openproduct development’ (Chesbrough, 2003). Dur-ing the development of the Internet Tablets,Nokia adapted and learned to work with a com-munity of volunteers. An open research questionis to what extent such learning can enable firms towork with outside volunteers across generationsand categories of products. Second, Harhoff et al.(2003), modeling the payoff for innovators tofreely reveal their innovations find that one offour conditions, ‘greater generality of the knowl-edge’ reduces the likelihood of free-revealing bythe innovator (see also Muller and Penin, 2006).However, this case study showed that muchknowledge revealed proved to be generic frame-works: it laid the foundation for a generic em-bedded Linux desktop environment. Morespecific components, such as power managementand some end-user applications, were kept pro-prietary. This contradicts Harhoff’s et al. find-ings, and the issue of what kind of knowledge islikely to be revealed under what conditions re-quires more attention in future research.

Limited by its design, the current study canonly generalize the findings to theory. The ex-tended private-collective innovation model pro-vides a set of benefits, costs, and mitigationstrategies that must be tested on a larger samplein future research using cross-sectional as well aslongitudinal designs. Thus, it will be important togarner insights on technological, industry, andmarket conditions that provide different levels ofbenefits and costs of innovation. For example, inindustries of non-virtual goods or where productdevelopment constitutes a minor share of fixedcosts in production (e.g., cement manufacturing),companies may find it more attractive to pursueprivate-investment innovation. Moreover, if in-novation is largely based on tacit knowledgeacquired through extensive and costly apprentice-ship (e.g., luxury goods), volunteers who joinproduct development products may be rare.Future research will also have to investigate the


r 2009 The Authors



impact of company age and size on the innova-tion incentives in the private-collective model. Aswe found in the case of an established company,the process represented costs of organizationalinertia.

Managers who want to experiment with flexiblesolutions, while keeping their own product devel-opment costs low, should investigate ways toimplement private-collective innovation. Sharingdevelopment costs and enabling contributionsfrom third parties, as well as boosting organiza-tional learning, are powerful reasons why themodel is attractive in practice. However, thereare potential ‘hidden costs’ in implementing themodel. Learning from Nokia’s successful ap-proach, managers should think ahead about pos-sible costs and create strategies to mitigate them.The experience from the development of the Inter-net Tablet provides possible mitigation strategies.

Acknowledgements

The research was supported by the Swiss NationalFoundation (grants 100012-101805 & 105512-106932). The authors wish to thank all interviewpartners for their time and comments, as well asEric von Hippel and Gideon Markman for pro-viding us with insightful comments and reviews.

References

Alavi, M. and Leidner, D.E. (1999) Knowledge man-

agement systems: issues, challenges, and benefits.

Communications of the AIS, 1, 1, 1–37.

Allen, R. (1983) Collective invention. Journal of Eco-

nomic Behavior and Organization, 4, 1, 1–24.

Arrow, K.J. (1962) The economic implications of

learning by doing. The Review of Economic Studies,

29, 3, 155–173.

Baldwin, C.Y. and Clark, K.B. (2006) The architecture

of participation: does code architecture mitigate free

riding in the open source development model? Man-

agement Science, 52, 7, 1116–1127.

Brown, S.L. and Eisenhardt, K.M. (1997) The art of

continuous change: linking complexity theory and

time-paced evolution in relentlessly shifting organi-

zations. Administrative Science Quarterly, 42, 1–34.

Brusoni, S., Prencipe, A. and Pavitt, K. (2001) Knowl-

edge specialization, organizational coupling, and the

boundaries of the firm: why do firms know more than

they make? Administrative Science Quarterly, 46, 4,

597–621.

Chesbrough, H. (2003) Open Innovation. The New

Imperative for Creating and Profiting from Technol-

ogy. Boston, MA: Harvard Business School Press.

Crawford, C. (1992) The hidden costs of accelerated

product development. Journal of Product Innovation

Management, 9, 3, 188–199.

Dahlander, L. (2004), Appropriating the commons:

firms in open source software. Technical report,

Chalmers University of Technology.

Dam, K. (1995) Some economic considerations in the

intellectual property protection of software. Journal

of Legal Studies, 24, 2, 321–377.

David, P. (1992) Knowledge, property, and the system

dynamics of technological change. Proceedings of the

World Bank Annual Conference on Development Eco-

nomics, 215–247.

David, P. (1998), Knowledge spillovers, technology trans-

fers, and the economic rationale for public support of

exploratory research in science. Background Paper for

the European Committee for Future Accelerators.

Downes, L. and Mui, C. (1998) Unleashing the Killer

App: Digital Strategies for Market Dominance.

Boston, MA: Harvard Business School Press.

Economides, N. (1996) The economics of networks.

International Journal of Industrial Organization, 14,

6, 673–699.

Economides, N. and Katsamakas, E. (2006) Two-sided

competition of proprietary vs. open source technol-

ogy platforms and the implications for the software

industry. Management Science, 52, 7, 1057–1071.

Edwards, K. (2003), Epistemic communities, situated

learning and open source software development.

Technical report, Technical University of Denmark.

Eisenhardt, K.M. (1989) Building theories from case

study research. Academy of Management Review, 14,

4, 532–550.

Foray, D. (2004) Economics of Knowledge. Cambridge,

MA: MIT Press.

Gachter, S., von Krogh, G. and Haefliger, S. (2006)

Private-collective innovation and the fragility of

knowledge sharing. Working paper available at:

http://nottinghamnetlearning.com/economics/cedex/

papers12006-24.pdf

Gambardella, A. and Hall, B.H. (2006) Proprietary

versus public domain licensing of software and

research products. Research Policy, 35, 6, 875–892.

Grand, S., Georg von Krogh, D.L. and Swap, W.

(2004) Resource allocation beyond firm boundaries:

a multi-level model for open source innovation. Long

Range Planning, 37, 591–610.

Granstrand, O. (1999) The Economics and Management

of Intellectual Property, Edward Elgar, Cheltham.

Haefliger, S., von Krogh, G. and Spaeth, S. (2008)

Code reuse in open source software. Management

Science, 54, 1, 180–193.

Harhoff, D., Henkel, J. and von Hippel, E. (2003)

Profiting from voluntary information spillovers:

how users benefit by freely revealing their innova-

tions. Research Policy, 32, 1753–1769.

Heckathorn, D.D. (1997) Respondent-driven sampling:

a new approach to the study of hidden populations.

Social Problems, 44, 174–199.




Henkel, J. (2006) Selective revealing in open innovation

processes: the case of embedded linux. Research

Policy, 35, 7, 953–969.

von Hippel, E. and von Krogh, G. (2003) Open source

software and the ‘‘private-collective’’ innovation

model: issues for organization science. Organization

Science, 14, 2, 209–223.

von Hippel, E. and von Krogh, G. (2006) Free reveal-

ing and the private-collective model for innovation

incentives. R&D Management, 36, 3, 295–306.

ICFAI (2005) Nokia and the Global Phone Industry,

ICFAI Center for Management Research, Hydera-

bad, India.

Jacobson, I., Booch, G. and Rumbaugh, J. (1999) The

Unified Software Development Process, Addison-

Wesley, Boston.

Jeppesen, L.B. and Frederiksen, L. (2006) Why do

users contribute to firm-hosted user communities?

The case of computer-controlled music instruments.

Organization Science, 17, 1, 45–63.

Kessler, E., Bierly, P. and Gopalakrishnan, S. (2000)

Internal vs. external learning in new product devel-

opment: effects on speed, cost, and competitive

advantage. R&D Management, 30, 3, 213–224.

Kogut, B. and Zander, U. (1992) Knowledge of the

firm, combinative capabilities, and the replication of

technology. Organization Science, 3, 383–397.

Kotha, S. (1995) Mass customization: implementing

the emerging paradigm for competitive advantage.

Strategic Management Journal, 16, 21–42.

Kuk, G. (2006) Strategic interaction and knowledge

sharing in the KDE developer mailing list. Manage-

ment Science, 52, 7, 1031–1042.

Lakhani, K., Wolf, B., Bates, J. and DiBona, C. (2002),

The Boston Consulting Group hacker survey. available

from: http://www.bcg.com/opensource/BCGHacker-

SurveyOSCON24July02v073.pdf. Accessed 1 July 2004.

Lerner, J. and Tirole, J. (2002) Some simple economics

of open source. Journal of Industrial Economics, 50,

2, 197–234.

Liebeskind, J.P. (1996) Knowledge, strategy, and the

theory of the firm. Strategic Management Journal,

Special Issue: Knowledge and the Firm, 17, 93–107.

Luthiger Stoll, B. (2006), Spass und software-entwick-

lung: zur motivation von open-source-programmier-

ern. PhD thesis, University of Zurich.

Maemo.org (2006), Maemo website. Available from:

http://maemo.org, accessed December 2006.

Martin, R. (2007), Nokia launches ‘Context-Aware’

internet tablet. InformationWeek.

Muller, P. and Penin, J. (2006) Why do firms disclose

knowledge and how does it matter? Journal of

Evolutionary Economics, 16, 1–2, 85–108.

Nokia (2005), Nokia makes donation to GNOME

Foundation. Nokia press release.

Nuvolari, A. (2004) Collective invention during the

British industrial revolution: the case of the Cornish

pumping engine. Cambridge Journal of Economics,

28, 3, 99–119.

Osterloh, M. and Rota, S. (2004) Trust and Commu-

nity in Open Source Software Production. In Lahno,

B. andMatzat, U. (eds.), Trust and Community on the

Internet. Lucius&Lucius, Stuttgart, Germany.

Paul, R. (2007) Intel launches site for open source

mobile Linux development. Ars Technica. Available

at: http://arstechnica.com 19 July 2007. Accessed 15

Jan 2008.

Pettigrew, A. (1990) Longitudinal field research on

change: theory and practice. Organization Science,

1, 3, 267–292.

PricewaterhouseCoopers (2005) Governance, risk, and

compliance series: global best practices, Available

from: http://www.globalbestpractices.com.

Roberts, J., Hann, H. and Slaughter, S. (2006) Under-

standing the motivations, participation, and perfor-

mance of open source software developers: a

longitudinal study of the Apache projects. Manage-

ment Science, 52, 7, 984–999.

Shah, S. (2006) Motivation, governance, and the

viability of hybrid forms in open source software

development.Management Science, 52, 7, 1000–1014.

Shah, S.K. and Corley, K.G. (2006) Building better

theory by bridging the quantitative–qualitative divide.

Journal of Management Studies, 43, 8, 1821–1835.

Shankland, S. and Charny, B. (2003), Linux to

power most Motorola phones. Available from:

http:// http://news.com.com/2100-1001-984424.html

accessed Jan 24th, 2007.

Siggelkow, N. (2007) Persuasion with case studies.

Academy of Management Journal, 50, 1, 20–24.

Smith, P. (2004) Accelerated Product Development:

Techniques and Traps. New Jersey: Wiley.

Sorenson, O. and Stuart, T. (2000) Aging, obsolesence,

and organizational innovation. Administrative Science

Quarterly, 45, 1, 81–112.

Stiglitz, J.E. (2006) Scrooge and intellectual property

rights. British Medical Journal, 333, 7582, 1279–1280.

Strauss, A.L. and Corbin, J. (1998) Basics of Qualitative

Research: Techniques and Procedures for Developing

Grounded Theory. Thousand Oaks, CA: Sage.

Torvalds, L. and Diamond, D. (2001) Just for Fun.

London, UK: Texere.

Yin, R.K. (2003) Case Study Research: Design and

Methods. Thousand Oaks, CA: Sage.

Zander, I. and Zander, U. (2005) The inside track: on

the important (but neglected) role of customers in the

resource-based view of strategy and firm growth.

Journal of Management Studies, 42, 8, 1519–1548.

Notes

1. A brief anecdote: in 2006, when one of the co-

authors of this paper asked for accessories in such

a shop, none of the shopkeepers were even aware of

the product’s existence.

2. A so-called ‘‘fork’’ results when ‘dissatisfied pro-

grammers’ copy the original source code from a

project and continue its development in an alter-

native, competing project.


r 2009 The Authors



3. In addition, Osterloh and Rota (2004) pointed out

that the mere presence of a firm in private-collective

innovation may ‘crowd out’ intrinsic motivation by

voluntary contributors (e.g., fun and enjoyment).

Appendix A

Gaining insight from interview codings

As described in the method section, we conductedinterviews with a continuously updated, semi-structured questionnaire investigating differentaspects of our research question. After transcrip-tion of the interviews, we categorized the quotesin similar statements, eventually leading to 80different codes and 1,026 text codings. Iterativeregrouping and recoding of the statements even-tually led to three main themes with a total of 12patterns. Based on these, we created our extendedmodel of the implementation of private-collectiveinnovation.

Since space restrictions prevent us from quotingtoo many statements within the article, we explainin the following sections how we extracted a generalfinding from analyzing the interview statements.For example, in the case of the subsequentlymerged category ‘Network Collaboration,’ thescreenshot of the text analysis software Max.QDAshows that the pattern consisted of three additionalcodes: ‘Maintenance/Upstream Project,’ ‘Nokiacollaborating with Third-Party Vendors,’ and‘How Nokia Manages OSS Development.’ Insum, 38 codings from the interviews covered state-ments on ‘Network Collaboration.’ Triangulatingthese statements and interpreting them eventuallyled to the case study findings in Section 5.

Excerpt of interview with N1:

[. . .] Suppose that I am a company X and Iwant to build a Linux-based product. Now Igo to the community and I take a snapshot ofthe Linux operating system today. I think acopy of the version 2.6.18. Then I go into mylab. I work two years with that, doing all kindsof fancy things. And then I again come back tothe community and say I took a snapshot twoyears ago, look what I have done. And thecommunity says we don’t care. You shouldhave been participating in the communitywork all the time so we could have seen whatyou do to integrate our great things with whatyou have done. Now we have got two totallyincompatible branches. Your branch may begood, but there is our branch and that is

certainly good. Do you really want to mergethese things? So, corporations didn’t reallyunderstand and still not all of them do, thatOSS is not only about software and what kindof software components you have, but it’s allabout the process that you develop this openlywithin the communities and you try to syn-chronize your own work with the heartbeat ofthe communities. Some companies now under-stand better than others. We certainly havedone our learning. We have made some mis-takes too on this front.

What kind of mistakes?

For example, we did exactly what I describedwith our email client. We took a suite for -mailfrom OSS and took a snapshot of that andcontinued developing that over a year. Wedidn’t synchronize it with the open sourceproject. We are now totally in our own branchand the maintenance costs are increasing. Wehave two options: we either make a huge effortin merging these two branches or then we justforget what we have done and jump back towhere the community is today.

Quotes leading to our abstracted finding,as written in the text:

‘Interviewees N1 and N3 pointed out that Nokialearned over time that this is indeed the case andthat it was inefficient in the long run to maintaintheir own software version while the upstreamproject is continuously improved by a muchlarger group of developers.’

Interview guideline for a volunteer memberof the Maemo community

About the interviewee and his involvement1. What is your education?2. What and for whom do you work at themoment?3. Since when have you been involved in opensource software (OSS) development and whatdid you do?4. In what way is your paid work related toOSS?5. Do you have commited access to OSSprojects? If yes, which ones?6. What mailing lists are you subscribed to andfollow most of the discussion?




7. When and how did you get involved in theNokia Internet Tablet project? Why?8. What are your contributions within theMaemo project?9. How many hours per week do you spend con-tributing to Maemo? Has this changed over time?10. Have you received financial or other mate-rial benefits from Nokia for your participationin the Maemo community?11. What is the main reason for your contribu-tions within the Maemo community?Knowledge Revealing Strategy1. In your opinion: Why did Nokia not justintegrate OSS components but create a com-munity portal and actively participate in estab-lished OSS projects?2. How would external participation haveevolved if Nokia had not published their ownsoftware developments as OSS?3. Would you have participated in the Maemocommunity if Nokia had used a proprietaryoperating system?4. From your perception: What type of soft-ware does Nokia release as OSS and what asproprietary binaries?Knowledge: Knowledge Reuse1. What do you think, why did Nokia not chosea ‘commercial’ Linux distribution such as Red-Hat or SUSE but Debian?2. Expert’s estimation (in %): How many linesof code of the software in the N800 (out of thebox) is1. unmodified, preexisting OSS (e.g., parts of

GTK)2. modified, preexisting OSS by Nokia or

contractors (e.g., parts of GTK)3. newly created OSS by Nokia or contrac-

tors (e.g., Hildon)4. newly created proprietary software by

Nokia or contractors (e.g., Canola)5. unmodified or modified preexisting pro-

prietary software by Nokia or contractors (e.g.,Opera)3. What tasks are usually required in orderto integrate preexisting OSS for the InternetTablet?Knowledge: Distributed Technology Expertise1. Why is it not difficult for external program-mers to develop software for Maemo?2. Why isn’t more software written for theplatform?3. What kind of questions do Nokia developersmostly ask?4. In your opinion, what technical knowledgedo you possess which Nokia employees don’t?5. What are the differences in Maemo commu-nity participation (mailing list, IRC . . .) if

somebody is employed by Nokia, working forone of Nokia’s contractors or if they are avoluntary contributor . . .

1. . . . regarding contributions?2. . . . regarding technical knowledge?3. . . . regarding form of communication?4. . . . regarding helpfulness?5. . . . other particular issues?

Knowledge: Guarding Business Secrets1. Why do external developers demand moreknowledge on plans of future Maemo develop-ments?2. What kind of information could Nokiapublish in a technical roadmap while guardingbusiness secrets?Organization: Network Collaboration1. Do you as a voluntary contributor feel partof the Maemo community? In what way or whynot?2. What are the benefits for Nokia when theirsoftware developments are integrated into es-tablished upstream OSS projects such as LinuxKernel, GTK, GStreamer, D-BUS?3. In which OSS projects is Nokia’s softwarecode successfully ‘integrated upstream’? Why?4. In which OSS projects is Nokia not able tointegrate their developments? Why?5. Are there certain examples how Nokialearned about the importance of this? If yes,which ones?6. How does Nokia gain trust in these projects?Organization: Reducing Network Entry Barriers1. What is your benefit in participating in thedevelopment?2. How does Nokia encourage external pro-grammers to voluntarily contribute to the In-ternet Tablet software?3. What should they improve so you’d spendmore time programming for Maemo?4. How relevant are voluntary contributions sofar for the Maemo platform?5. In which communities does Nokia still needto improve its acceptance?Organization: Recruiting benefits1. What is your image of Nokia?2. Would you accept a job offer by Nokia if youcould work at the Maemo project?Organization: Balancing Control1. How can Nokia maintain control oversource code they integrated in incumbent OSSprojects?2. What type of control is important and whichone can be neglected?3. How does Nokia influence future develop-ments in established OSS projects?4. What made volunteer developers leave theMaemo community?


r 2009 The Authors



5. What are other similar OSS projects volun-teer developers could be attracted to?6. Is it probable that certain componentsof the Maemo software will get forked by thecommunity or by other software or hardwarevendors? Why? Why not?7. How transparent is the development processof the Maemo platform?8. Would you like to have more influence in thedevelopment process of the Maemo platform?Organization: Organizational Inertia1. What organizational aspects make it difficultfor Nokia to participate in OSS communities?2. How is collaboration with volunteers af-fected by Nokia’s organizational structure?Organization: Organizational Learning1. From your perspective, what have the keylearnings been since the release of the Nokia770 Internet Tablet?2. What organizational improvements shouldNokia make in the future?Competitive Situation: Time-to-market1. Did the volunteer community help to speed upthe development process of theMaemo platform?Competitive Situation: Countering Uncertaintywith Flexibility1. How flexible is the hardware/software plat-form to implement completely new use cases forthe Internet Tablet?2. How does uncertainty of the future inhibitinnovation by Nokia resp. by users?Competitive Situation: User Contributions1. What are key contributions from the volun-tary developer community?2. How well are the users aligned with Nokia’splans for the Internet Tablet?3. Could volunteers develop a killer applicationfor the Tablet? Why? Why not?Competitive Situation: Difficulty to Differentiate1. In what way does Nokia lose when it revealsthe source code of their own developments?2. What do you think, for what reasons arecertain parts (hardware drivers, graphical userinterface, applications) not reveale3. Do competing projects (e.g., Openmoko,OLPC etc.) using code fromMaemo?What parts?Conclusions1. In your opinion: What is the greatest benefitand what is the greatest disadvantage of ‘open-ing up’ software and hardware products?

Interview guideline for a Nokia manager

IntroSince when do you work for Nokia?Since when have you been working for theInternet tablet project?

What’s your role within the project?What are the goals of the Nokia 770?Is there a difference for N800?ChoicesWhy did you choose Linux instead of Symbian/Windows CE? (What would have been different?)Why did you choose Debian instead of e.g.,Suse or RedHat?Why did you choose the GNOME/GTK stackand not KDE?Why didn’t you contract an Embedded Linuxcompany such as MontaVista to build theoperating system for you? (such as e.g., Motor-ola)What were the technical and business reasonsto break backwards-compatibility of the OS2007 Edition? What are the benefits and risksof this decision?ContractorsWhy do you contract small firms such asKernelConcepts? (What are their strengths?)Which external companies did you contract forthe Internet Tablet development?When do you do things internally? (For exam-ple . . .)What type of knowledge is transferred to Nokiathrough contractors? (How?)How long would it have taken you to createNokia 770 without the help of contractors?Why didn’t you do it all by yourself?What was the advantage of choosing OSS?What are the challenges contracting small firms?VolunteersHow do you motivate volunteer contributions?(For example . . .)What is the influence of revealing source codein the context of motivation?What kind of contributions are made volunta-rily and for what do you have to pay money?(For example . . .)How business-critical are voluntary contribu-tions really for the functioning of the InternetTablet? (Looking at the interviews, it seems asif the most difficult, central things are done byNokia itself or by contracted firms. ArminWarda said his contribution concerning swapmemory was negligible.)On what occasions do you meet volunteercontributors physically? (How often?)What is the difference between physical vs.virtual collaboration?When do you decide to pay somebody for acertain development effort? (For example . . .)On what criteria of a developer do you decideto hire somebody? (For example . . .)What’s different when recruiting somebodyfrom the community than somebody from




within Nokia? (e.g., in the case of the Maemoproduct manager)What changed in terms of employer attractive-ness when Nokia started OSS development?Do Nokia employees ask technical questions tothe volunteer community? (For example . . .)

CommunitiesWhy do you prefer collaborating with the up-stream project? (instead of forking your ownversion)In what cases is it better to fork an establishedproject? (Why?)How do you make Nokia accepted within theOSS communities? (For example . . .)In what kind of tasks is Nokia strong comparedto the strengths of the volunteer community?(For example . . .)In what kind of tasks is the volunteer commu-nity strong compared to the strengths of No-kia? (for example . . .)What are the challenges being a commercialcompany within OSS communities? (for exam-ple . . .)

CompetitorsAre there competitors in the mobile deviceindustry pursuing the same OSS strategy?Do you know of cases where competitors usedyour revealed source code?How do you collaborate with competitors? (forexample . . .)What is the benefit of setting standards such asthe GNOME Embedded Platform? (Don’tcompetitors catch up more easily if they can

integrate a standard and develop their ownproducts faster?)

Matthias Stuermer studied Business Administrationand Computer Science at the University of Berneand graduated in 2005 with his thesis on open sourcecommunity building. He currently pursues his doc-toral dissertation on firm involvement in open sourcecommunities at the Chair of Strategic Managementand Innovation at ETH Zurich, Switzerland.

Sebastian Spaeth graduated as Master of Scienceat Linkoping Technical University, Sweden. Afterreceiving his Doctor in Business Administrationfrom the University of St.Gallen, Switzerland, in2005, he conducts research in the areas of net-work innovation and open business models as aresearcher at the ETH Zurich, as well as teachingMaster courses on Strategic management. He haspublished in Research Policy and ManagementScience among other journals.

Georg von Krogh is Professor of Strategic Man-agement and Innovation at the ETH Zurich’sDepartment of Management, Technology, andEconomics. He received his MSc and PhD fromthe Norwegian University of Technology andNatural Science. He has published articles andbooks on strategic management, knowledge crea-tion, and innovation, as well as organization andmanagement theory.


r 2009 The Authors



Extending private-collective innovation: a case study

Documents

Transcript of Extending private-collective innovation: a case study