Can action research improve local journalism?

Norway was considered a good market for the INJECT roll-out. In 2016, there were 227 local newspapers with at least one weekly edition (MedieNorge, 2022). As such, there was a large potential market for the creativity tool in question. According to UNESCO, Norway has the highest rate of newspaper readership worldwide. Norway also has strong social-democratic traditions, strong local communities, and government support of Norwegian cultural and historical values (Halvorsen & Stjernø, 2008). However, each national INJECT ecosystem would have the same basic structure, with companies and institutions at the top level and important tasks for designated individuals.

INJECT aimed to foster relevant journalistic technology development for independently owned local newspapers and to build a sustainable business ecosystem on top of the new technical product. The main presumption was that it is worthwhile for local newspapers to learn to develop tools in their local newsrooms and to commercialise them with local partners instead of buying them ready-made, one-size-fits-all from the shelves of big global suppliers. To fulfil these objectives, it was crucial to foster good collaboration between the partners.

There were three core academic research and development partners in INJECT Norway whose main tasks were to design, develop, manage, and report on the project. A British university and a Greek partner university provided the software technology and conducted all interface design, programming, and development. The same system was used for ecosystem experiments in France and Germany, but the Norwegian system was the core of the project. The British university furthermore had overall leadership over the Innovation Action. A Norwegian university oversaw the user testing and conducted contextual inquiries, focus groups, and other types of qualitative interviewing to learn about how journalists used the tool. Researchers from this university became the most “action research–minded” of the partners: They encouraged collaboration through many visits to the other partners and promoted the project in academic as well as journalistic contexts. Students of journalism and media design were recruited as assistants.

Three local family-owned newspapers contributed as testbeds and initial customers for the INJECT creativity tool. The local newspapers had an ongoing partnership intending to explore possible innovations in their newsrooms, and they already had a line of communication with the Norwegian university involved. A commercial consultancy oversaw public relations, sales, and business development. The company explored ways of securing revenue from licences sold in the Norwegian market, as well as payment for services like technical installations, tutoring of journalists, and courses. It was responsible for growing the ecosystem and finding new revenue streams. The commercial partner also already collaborated with the university for internships in educational contexts.

The EU Commission was an important stakeholder in the project due to its financial investment through the grant. In this sense, the project had a centralised management structure. All partners were reimbursed for their efforts to develop, test, and improve the INJECT tool. The industry partners were compensated with 70 per cent of the costs for work performed during the action, and academic partners 100 per cent. The deliverables and review process in Brussels had to be approved before reimbursements were made.

The main data material for the study is the corpus of official documents produced within the INJECT Innovation Action. During the 18-month project period, 13 deliverables were submitted to the EU Commission, with a volume of approximately 450 pages of text. Deliverables included specifications of the digital services and tools, formative and summative evaluations of the tool, descriptions of the innovation ecosystem in Norway at two stages, dissemination and communication reports, and two versions of a data management plan. In addition, the project proposal, as well as various types of materials in the form of PowerPoint presentations, newsletters, stories published on the INJECT website, and social media posts have informed the analysis.

We stress that we do not evaluate the full suite of technologies and competences stimulated by the INJECT project. The creativity tool lives on as part of a product portfolio called JECT.AI (Maiden, Zachos, Franks et al., 2020). During the active phase of the project, there were also emerging regional ecosystems in the Netherlands and France (see Wagemans & Witschge, 2019). In this article, we evaluate the dynamics of collaboration only in the Norwegian business ecosystem and mainly draw on the four deliverables that deal most actively with the Norwegian innovation ecosystem. Corresponding to requirements set up by the terms of the project, these deliverables dealt with the initial version of the first INJECT ecosystem, the sustainable version of the first INJECT ecosystem, the first INJECT ecosystem “business & exploitation plan”, and finally, the dissemination report for the evolved INJECT ecosystem. We use additional information from the four research articles that have been published from the project (Maiden et al. 2018; Maiden et al. 2019; Maiden, Zachos, Brown et al., 2020; Wagemans & Witschge, 2019).

To reduce exposure, we do not use the names of any of the institutions but give them labels such as “The British university”, “three local newspapers”, and “the commercial partner”. The identity of the participants in INJECT Norway are anonymised in much the same way as the narrative about the town development scheme presented in Greenwood and Levin (2007); however, those who were involved in INJECT Norway will probably understand which companies, leaders, researchers, and journalists we refer to in our analyses. It is not a breach of confidence if someone with previous knowledge about the INJECT Norway project identifies people we refer to in the analysis, since all partners were aware of this possibility when they joined the project.

As described above, INJECT Norway was organised as a business ecosystem consisting of many companies and institutions. It is therefore important to define the concept of ecosystem within the project. Moore (1993, 1996, 2006) established the idea of business ecosystems as a strategy to promote continuous innovation in a disruptive marketplace where conditions regularly change. A business ecosystem is “an economic community supported by a foundation of interacting organizations and individuals”, according to Moore (1996: 26). The strategy of establishing business ecosystems points towards an intended increase in the overall public good emerging from the community. The ecosystems are in themselves intended as “collaborative, innovation-furthering public goods”, Moore (2006: 35) argues. In general, “the fundamental public good offered by business ecosystems is the taking of a challenge that requires coordination and finding a way to take it out of a firm and bring in more participants”, according to Moore (2006: 48). In more specific terms, a business ecosystem for innovation in journalism may be expected to result in public goods such as a more participatory local public, higher quality exposure of corruption or other forms of wrong-doing, and a more enlightened citizenry.

Moore presumes that the success of a business ecosystem fundamentally depends on the collaboration it can achieve:

For every advance there are complementary innovations that must be joined for customers to benefit. These complementary advances often must co-evolve across company lines because no single firm has all of the required specialized knowledge and managerial resources necessary for the whole system.

Collaboration also means that companies must “find ways to align their visions, so that research and development investments are mutually supportive, and capital investments and operating processes are synergistic” (Moore, 2006: 34). Over time, the ecosystem partners presumably co-evolve their capabilities and roles and tend to align themselves with the directions set by one or more central companies (Moore, 1996):

Companies must establish interfaces and protocols for putting together their contributions. Most importantly, they must dialogue closely with customers so that what is created is what the customer wants and is willing to pay for.

Also, motivation is vital to the development of technical solutions, to both testing and implementation and to feelings of personal engagement and community (Moore, 2006). It is important to ensure that the group is open to ideas no matter where or how they originate and that it doesn't discriminate against some potential contributors and favour others. To achieve this, various kinds of talent are needed, and no single organisation can oversee the work, maintains Moore (2006). As the evaluation shows, this became a difficult ideal for INJECT Norway to accomplish.

The INJECT software was intended to support journalists in discovering creative angles on news stories and to discover them more quickly than with existing search tools like the Scandinavian newspaper archive Retriever or Google Search. There would be increased efficiency and profit for the journalism partners if the newsroom spent less time on writing each story and readers spent more time on reading and enjoying them more (for details about the INJECT tool, see Maiden et al., 2018; Maiden et al., 2019).

The design and testing of INJECT was a co-design process in line with the collaborative ideals of action research. According to Järvinen (2007: 15), action research and design science are similar in that both are “primarily oriented at solving the local problem in close collaboration with the local people”. Early in 2017, several British, Dutch, and Norwegian journalists and journalist students were invited into the design process. First, paper-based sketches were discussed, and then digital wireframes were developed and presented. Journalists were interviewed again to discover problems and further requirements.

The first functional interface was only available in English. A contextual inquiry in the summer of 2017 involved Norwegian journalists trying out this interface by testing the functions for international news stories (e.g., Brexit, Donald Trump). The journalists reported that INJECT seemed quite useful for in-depth journalism, but that it did not make much of a difference for breaking news. However, it could give creative support for reportage, background analyses, or portraits and could help them to discover new interview sources and complement established and over-used sources of information. The contextual inquiry also revealed that journalists wanted INJECT to be integrated into their content management systems for easy access. They also wanted to enable the INJECT to search in the newspapers' own internal archives for more relevant and precise creative prompts. On top of this, journalists wanted INJECT to be able to process municipal websites, national statistics, and posts from social media like Facebook, Instagram, and Twitter.

After the test phase, the INJECT tool was adapted to the identified, Norway-specific needs with translations of the user interface and other modifications to fit the national context. In the autumn of 2017, a new version of INJECT was launched, with improvements based on the contextual inquiry. It now had a Norwegian-language interface and was functional in WordPress, Google Docs, and Adobe InCopy. It could be connected to internal archives and would at a later stage also be connected to Facebook profiles and other social media outputs. The INJECT tool at this time accessed over 3.5 million indexed English-language articles; over 300,000 Norwegian-language articles; and 62,160 Norwegian-language articles from the archives of the three newspapers.

The design team, accustomed to co-designing with primary users, worked directly with the journalists. However, journalists and designers didn't necessarily understand each other that well; the British and Greek designers didn't have in-depth understanding of the work practices of local journalists in Norwegian fjords and valleys. Likewise, journalists don't necessarily understand or acknowledge the processes involved in inclusive co-design. Initially at least, they had an expectation that the INJECT tool would work effectively as soon as it was activated. The design team also learnt something about the hierarchy of newsrooms. Journalists had no power to make decisions about technology uptake. The editors and CEOs make decisions based on the need for productivity and cost-effectiveness, which is not necessarily the primary need of the journalist. However, there was a bottom-up pressure from journalists that influenced what editors decided to prioritise. Overall, the INJECT design team felt that they were co-designing for two different groups: journalists who are using the tool in the newsroom and leaders who are making strategic decisions about purchasing tools for the newsroom.

Next, INJECT was introduced in the newsroom, with the aim of learning something about technology design and increased innovation competence, as well as newsroom leadership, by the ways in which the editors and journalists reacted.

At first, there was noticeable enthusiasm among the editors, CEOs, and journalists from the three newspapers. The entrepreneurial spirit of the British and Greek team was contagious and created an optimistic mood among the editors and researchers in the Norwegian ecosystem. They saw that the collaborative innovation process adopted in INJECT had real potential and that the prototype became more useful as the dialogue progressed. This mobilising power of an EU Innovation Action among local stakeholders hence proved to be substantial; however, it is also a perishable resource that may lose its effect over time.

It was well known that one of the greatest barriers to innovation in journalism is people's unwillingness to adapt to new tools in their workday. There was therefore a two-part strategy for involving the participants: First, representatives from the commercial partner organised tutoring and workshops to help increase the adoption of INJECT and followed up by telephone and e-mail. New users were guided through the features of the system and were encouraged to adapt an open mindset towards INJECT. Second, the journalism partners appointed champions, or “agents of change”, in their newsrooms. Journalists in each newsroom that came across as genuinely interested in INJECT and other innovative technologies were given the role of first movers in their newsrooms. The idea was that these enthusiastic early adopters would train their colleagues to use INJECT.

Innovation champions in the newsrooms are important, and journalists who saw the value in INJECT and were prepared to act as internal influencers for its uptake and support were valuable. Typically, it was the younger journalists more experienced with various technologies who were open to INJECT. People who are relatively new in the journalism world – and who have grown up with digital technologies – are typically more open to exploring how these technologies can be adapted for news. Young journalists wanted to explore novel possibilities, while older journalists were sceptical. This tendency was also noted by Grubenmann (2016: 168), who says that journalists had spent very little time studying international examples of digital journalism and trends before their participation in the research project: “Most of the journalists seemed to lack the resources and/or the motivation to follow these developments as part of their daily work”.

Next, time is an enemy of innovation trials. After the initial enthusiasm in the spring of 2017, the newspaper participants increasingly lost interest. Even though the EU Innovation Action's budget had an earmarked post to reimburse the newspaper partners for making their journalists use INJECT actively, the pressure of other journalistic tasks did not diminish during the project, and journalists engaged far less with INJECT than the editors and researchers expected when the project started.

The increasing lack of engagement with INJECT from some fractions was quite detrimental to the innovation efforts. Some journalists considered it yet another tool on the desktop and didn't feel they had time to use it – and even less so for contributing painstakingly to its development through design iterations with teams in Britain and Greece. In fact, there were fractions of seasoned journalists who seemed to actively undermine the engagement with INJECT in the newsroom. Our impression was that they felt their expertise was threatened by the new tools that the younger people in the newsroom handled better. They made fun of the tool's ineffectiveness at an early stage – when it could not be expected to function optimally. The positive vibes surrounding INJECT were slowly morphing into negative ones. In the spring of 2018, it became clear that most journalists in the local newspapers had no interest in continuing the attempt at growing a bigger Norwegian ecosystem. Indeed, in one newsroom, the mere mention of the word “INJECT” caused consternation and negativity among journalists.

Each newspaper received EU funding to be part of the INJECT project and signed a contract stipulating that signing partners would participate professionally. Given that the funds were there to support involvement, not to underwrite non-INJECT activities and internal losses, their lack of involvement towards the end of the funding period caused tension, but it did not lead to acute conflict.

Technology development and maintenance in local newspapers are often assigned to external companies. This likely increases homogeneity among newspapers and results in a low capacity to innovate. INJECT Norway was an opportunity to imagine new ways of organising local newspaper ecosystems so that the low capacity to innovate could be overcome, but this turned out to be a substantial obstacle despite all good intentions.

Small newspapers rely on large companies for their data infrastructure, and there is a tendency toward homogenous infrastructure in the Norwegian news market. Technical specifications are comparably similar across the sector, and for INJECT, this meant that an integration into one newsroom could potentially facilitate integration into several others. The three newspapers all subscribed to data services from the same company, and at first this homogeneous infrastructure aided the INJECT innovation efforts – when something worked well for one newspaper it worked well for all three.

In January 2018, the last design iteration was about to start, and the result of these trials was supposed to corroborate the use value of INJECT to potential customers and investors. All three newsrooms were supposed to use INJECT through their InCopy text editor for easy access. However, it turned out that the company that designed and maintained the newspaper archives and online publishing system failed to integrate the INJECT tool into the newspapers' content management systems, so the journalists in the three newsrooms had to use the web application version of INJECT in a separate browser window. This was a sub-optimal situation that influenced the usability of the tool. It was very difficult to do the adjustments that allowed INJECT to browse the archives, and it was not possible to search Facebook and other social media. It turned out that the contract between the news organisations and the technical company did not include support for services originating outside the company's own ecosystem.

The lesson learned from this step is that many smaller news organisations outsource the maintenance of some technologies to third-party businesses. INJECT was deploying new technologies but could not rely on the news organisations to be in control of technical support. The wider ecosystem of technologies and businesses around any newspaper's operation must therefore be known by innovators for them to assess whether a technological integration is contractually possible or not.

As the INJECT tool matured, it became more important to expand the ecosystem to include more news media and improve market penetration. The commercial partner was the primary point of contact for potential new clients and users. There were three elements in the commercial infrastructure: the legal entity that will sell the product; the support network that supports those who buy the product; and the pricing model that differentiates between customer needs. A range of pricing models were discussed before a three-level model was decided upon:

Creative newsroom (initial setup: EUR 2000, then EUR 50 per user/year). It includes the INJECT tool integrated in the customer's editorial software, search functions for the internal news archive, professional tutoring, maintenance, and follow-up.

A series of events was held to pitch the INJECT tool to potential customers. These events were well received, especially in the Norwegian academic community, and were noticed in media trade journals. An online user community was considered central for onboarding new customers, offering a place for information, community building, raising awareness, and facilitating engagement with the brand and the tool. An INJECT Norway landing page was created to serve as the basis for this engagement.

Ultimately, INJECT was supposed to be a self-sufficient organisation, by onboarding new paying customers into the ecosystem and avoiding the need for continuous third-party funding. At the same time, the promotion dialogue revealed that the pricing model was too high for the local newspaper market, which put the profitability of the venture at risk.

Moore (2006: 33) describes the ideal function of business ecosystems as being “communities of economic actors whose individual business activities share in some large measure the fate of the whole community”; this was not the case with INJECT Norway, where the partners had different interests and priorities, and none were experts on the local newspaper market structure in Norway. Newspapers are relatively careful with investments, and it was not easy to persuade them to try new tools that cost as much as INJECT did. The Norwegian university partner was supposed to promote the product too but was not accustomed to acting as a sales unit. The three newspapers could potentially promote INJECT in the newspaper community, but they were hesitant to recommend the tool – their main activity is to sell subscriptions and advertisement space, not software. Finally, the British management team were too far away culturally and geographically to keep up with the pressure.

The ecosystem was dependent on external funding, and when the funding ended, the Innovation Action ended too. Temporary income, in this case from the EU Commission, can hence be considered a risk in similar projects. The deliverables required all partners to do predefined tasks. For some partners, this by far exceeded the budget and locked them into a workload that was more costly than beneficial. Despite hard work, careful construction of the business ecosystem, and EU support, the INJECT tool did not generate income at the end of the project period. When the external funding ended, newspapers had to choose between continuing with their private funds or ending the engagement. The commercial partner was unable to go on, and the local newspapers were hesitant. They were unable or unwilling to recommend INJECT to their newspaper partners, and the ecosystem was therefore not able to expand in Norway.

It is possible that INJECT Norway could have had more success if it were organised as a business cluster where the collaborators were geographically close to each other (see D'Este et al., 2013). Such industry clusters have proven to, on average, encourage innovation, but the quality of the results vary considerably. Important factors that influence the outcome are industry type and geographical region, as well as internal factors such as concentration within the cluster. A few economically large companies in a cluster have a less positive effect than many smaller companies, and more specialised clusters have better results (Fang, 2015).

The INJECT Innovation Action project reveals the challenges related to managing the local-global dimension. The project showed the need for intermediaries and translators in the local newspaper ecosystem to make “global” innovations accessible and adaptable to the local context, not only language-wise, but also in terms of local organisations and ways of working. This lesson is particularly important as universities increasingly reinvent themselves as entrepreneurial universities in triple helix collaborations where the state provides innovation funding, the private sector provides business opportunities, and the universities provide highly skilled technical experts (Etzkowitz & Zhou, 2018: 7). The Norwegian university sector must become better able to translate between the global and local actors if future action research projects are to become successful.

The project also showed that a dilemma with EU Innovation Actions is that the leadership of the ecosystem can be too far from the local context, and therefore less able to make stakeholders feel heard and to motivate their local innovation efforts. An ecosystem can be artificial in that it must be upheld from a distance with strong communication abilities instead of by everyday interaction near each other. In the case of INJECT, the ecosystem reached the last stages of planning, and the actual implementation was imminent, there was a marked increase in the volume of project communication via e-mail and Skype, and through physical meetings in London and Brussels. As the frequency of communication increased, the complexity of the relationships between participants increased too. In the end, it became almost impossible for the newspaper partners to communicate their local interests in a way that fit with the requirements of the ecosystem leadership's communication practices. The collaboration felt increasingly artificial to some of the partners. There was a sentiment along the lines of “we could just as well have done it in our own way all by ourselves”.

An innovation project organised as an 18-month EU project can be too rigid to develop organically along the lines of validated insights. Objectives were set out in advance, schedules, partner roles, and activities were predetermined, and these all had to be delivered according to the official plan. The funding relied on satisfactory written reports in the form of project deliverables. While this seems reasonable from an administrative point of view, it makes it difficult for the management to allow partners to pivot, follow new ideas, and in other ways explore what seems the best possibilities. This was a point of contention for all partners, but nobody was outspoken about it.

Despite this bureaucratic rigidity, the leadership of the INJECT project was able to adjust the ecosystem strategy. The less-than-expected involvement of the three newspapers led to a reorientation of the project toward the end of the funding period. The leadership pivoted towards more resourceful partners in the consortium and evolved ecosystems in the Netherlands and Germany, leaving the Norwegian ecosystem behind. In a sense, the newspapers contributed to their own narrative that INJECT was not useful for them. In a competitive game, opportunities will move elsewhere, leaving those who don't engage behind.

Action research requires investigators to possess personal and professional abilities that inspire credibility among the local partners. According to Greenwood and Levin (2007), action researchers should be friendly outsiders radiating a basic optimism about themselves and their collaborators, and they should be able to discuss failings in the local group's practices and perspectives on innovation in a way that is experienced as supportive rather than negatively critical or domineering. “Too much feedback”, according to Greenwood and Levin (2007: 125) “can block a group; too little can prevent the group from moving ahead”. It turned out to be almost impossible for INJECT academics to embody these ideals.

The academics in INJECT organised project meetings in Brussels, London, Amsterdam, Paris, and Bergen. Due to the cost of transport, no meetings were held in the towns where the newspapers were located. This seems to have created a sense of distance for the local partners: They were never on “home turf”, and the dialogue suffered from this. Furthermore, the meetings were organised as academic seminars instead of problem-solving meetings. The agenda consisted of speeches and group activities that were often quite unrelated to the most burning issues at any given time.

The friendly outsider is supposed to be a diplomatic coach, not a director or a boss; the INJECT academic team was likely too hierarchical. Greenwood and Levin (2007: 126) assert that the last thing most local groups who are stuck in difficult situations need is someone else telling them what to do. Action researchers must be “personally secure enough to admit ignorance and uncertainty and yet be able to advocate their own understandings and hopes” (Greenwood & Levin, 2007: 128). The INJECT academics were most likely not sufficiently empathic and involved in the perspectives of the problem-owners, and this seemed to make the latter frustrated. “Complex projects with diverse stakeholders in highly charged situations do not yield to quick fixes or magic bullets”, argue Greenwood and Levin (2007: 128). In hindsight, we acknowledge that the challenges faced by local newspapers regarding innovation in news production and audience engagement were more complex and also more subtle than we expected.

Also, action research requires everybody involved to be risk-takers, according to Greenwood and Levin (2007: 127). While conventional social scientists are “trained to avoid risks and to try to look good, no matter what happens”, action researchers must be willing to risk personal failure by supporting a local group that may or may not succeed. If not, they “will not provide the necessary moral support and confidence to people who are trying to persuade themselves to take risks as well” (Greenwood & Levin, 2007: 127).

While the creativity software developed in INJECT is now available in the international journalism market and can be considered a successful design based on action research, the university partners failed to cultivate long-term innovation competence among the local stakeholders. The academic researchers involved in INJECT Norway were good at designing new technology, but not so good at behaving like true action researchers.

It was difficult for the local partners to keep up the positive engagement as the project proceeded. The INJECT creativity tool did not solve their immediate problems with efficiency but would require a long-term engagement to reach its potential in the local newsrooms. This is understandable, because as Greenwood and Levin (2007: 104) maintain, since “local stakeholders take action in their own environments, the consequences of errors are both significant to them and often rapidly apparent”. The academics were at a safe distance from the everyday stress of the newsroom and did not appreciate the negative implications of tests and trials for the work environment.

In addition, the established hierarchies in the newsrooms worked against appreciating the value of the INJECT novelties. Greenwood and Levin (2007: 125) argue that local modes of thinking might cause groups to “shut down or to cycle unproductively over issues without resolving them”, and they also point out that “local people, because of their history together […] often are unable to tell each other uncomfortable things that they clearly are aware of [… and they] overlook some important resources for change” (2007: 125–126). Editors listened to senior members of the editorial staff, who were very knowledgeable, as well as sceptical, about the sudden cultivation of digital innovation competence. Younger staff members, who showed greater interest in the INJECT tool, were not listened to in the same way by the editors. Nobody brought this dilemma to the attention of local leaders. If young members of the editorial staff had been given a stronger role as problem owners, the outcome might have been different. Greenwood and Levin (2007: 129) point out that if conflicts arise between local stakeholders, or between them and the researchers, the process should be recalibrated. The leadership of INJECT was unable to recalibrate after learning that the local hierarchies worked against our innovation efforts. The commercial partner made several personnel changes during the active phase, and we noticed too late that this inadvertently marginalised them in important decision-making processes. Furthermore, this partner was concerned that they spent more work hours than they were reimbursed for and signalled a lack of interest in taking risks in the attempt to make the product successful. Seen together in hindsight, these dilemmas and conflicts were more substantial than when we were able to recognise during the active phase.

The local newspapers finally lost all interest in the project and focused their attention on other possible innovation partners. One of the three newspapers who participated in the INJECT Norway project has now entered into a publishing agreement with Amedia and no longer has responsibility for its technology or future technology development. From the perspective of the INJECT leadership, the local participants did not make the necessary effort to cultivate innovation competence in their editorial staff. Ideally, they ought to have had more patience and to have been better able to suffer setbacks and keep going regardless.