The academic literature gives several definitions of the concept of Crowdsourcing. A recent study by Estellés-Arolas and González-Ladrón-de-Guevara (2012) identified 40 various definitions, which include the use of a crowd for problem-solving (Chiu, Liang, & Turban, 2014). The most cited definition is the original definition of the concept, coined by Jeff Howe: “The act of a company or institution taking a function once performed by employees and outsourcing it to an undefined (and large) network of people in the form of an open call.” (Howe, 2009).

The second most cited definition is provided by Afuah and Tucci (2012): Crowdsourcing is the act of outsourcing a task to a “crowd,” rather than to a designated “agent” (an organization, informal or formal team, or individual), such as a contractor, in the form of an open call” (Boons, Stam, & Barkema, 2015; Palacios, Martinez-Corral, Nisar, & Grijalvo, 2016).

The two definitions indicate a consensus; it is however challenged by Zheng, Li, and Hou (2011) who perceive Crowdsourcing as a co-creation process, where the firm outsources its internal tasks by using the internet as a platform. Poetz and Schreier (2012) support this while putting the concept into perspective of open-source software. In this paper, Crowdsourcing is an outsourcing process, where a large network of people gets the opportunity to solve a certain task for a company, communicated in the form of an open call. Chua, Roth, and Lemoine (2015) researched the impact of culture on creativity and how this affected the participation in an online Crowdsourcing work. The authors claim that individuals from tight cultures are less likely than individuals from loose cultures, to engage in and succeed at creative Ideation in foreign cultures. The result being, that: “…the greater the cultural distance, the lower the likelihood of creativity engagement and success at foreign creative tasks.” (Chua et al., 2015, p. 216).

Bauer and Gegenhuber (2015) describe how Crowdsourcing affects the roles of consumers and producers. They claim that Crowdsourcing blurs the boundaries between consumer and producer, by getting consumers to employ capacities like their time, effort and knowledge. Hence, consumers act as suppliers. Kohler (2015) analyzed Crowdsourcing platforms to identify the patterns of successful and effective Crowdsourcing-based business models, with the scope of being able to suggest guidance for managers who wish to create or adapt an open, Crowdsourcing-based business model. Feller, Finnegan, Hayes, and O’Reilly (2012) have examined “Solver Brokerages”, which are the intermediaries that facilitate the process of Crowdsourcing. The “Solver Brokerages” are of scientific and practical interest because they enable organizations to access outside knowledge and innovation capacity and as such increase the innovation capacity (Feller et al., 2012). Liu, Yang, Adamic, and Chen (2014) explore how different types of incentives affect the participation and submission quality. They found that higher rewards induced not only more submissions but also submissions of higher quality.

A study of Dell's IdeaStorm community highlights the difficulties in maintaining a steady supply of quality ideas from a crowd over time. Specifically, the study reveals that people who submitted an idea several times are more likely to generate ideas that are valuable to the organization (Bayus, 2013). Another study on Dell's IdeaStorm found, that: “…individuals tend to significantly underestimate the costs to the firm for implementing their ideas but overestimate the potential of their ideas in the initial stages of the Crowdsourcing process.” (Huang, Singh, & Srinivasan, 2014, p. 2138). The ideation possibilities might be overcrowded with ideas, and some unlikely to be implemented. However, over time the average potential of ideas increases, while at the same time, the number of submitted ideas decreases (Huang et al., 2014).

The crowdsourcing process as described by Zhao and Zhu (2014) is powerful and everyone has the potential to participate. According to the authors, a typical Crowdsourcing process starts when the organization identifies which tasks they wish to crowdsource, and deduces them available to a crowd that is willing and capable of executing the tasks for the organization, for a fee or other incentives. The crowd participates either as individuals, or as teams, and tries to solve the task. After completion, the work is submitted to the Crowdsourcing platform. This is followed by the organization's assessment of the quality of the submitted work and the participants can then receive compensation (Zhao & Zhu, 2014, p. 417). Crowdsourcing has been valuable and widely used in practice. In their contemplation, 55 academic articles are identified, out of which, only 9 provided a theoretical basis. The authors claim that there is a lack of research on Crowdsourcing based on organizational theoretical foundations (Zhao & Zhu, 2014, p. 418ff).

Not surprisingly, crowdsourced problem-solving can, under certain circumstances, be more efficient than solving the problem internally, provided that (1) the characteristics of the problem, (2) knowledge required for the solution, (3) the crowd and (4) the solutions being evaluated. This is the case when Crowdsourcing transforms distant search into local search, which enables the company to benefit from distant search without carrying the associated cost. Conditions for Crowdsourcing are ideal when: “(1) the problem is easy to delineate and broadcast to the crowd, (2) the knowledge required to solve the problem falls outside the focal agent's knowledge neighbourhood (requires distant search), (3) the crowd is large, with some members of the crowd motivated and knowledgeable enough to self-select and solve the problem, (4) the final solution is easy to evaluate and integrate into the focal agent's value chain, and (5) information technologies are inexpensive and pervasive in the environment that includes the focal agent and the crowd” (Afuah & Tucci, 2012, p. 356). Bloodgood (2013) however questions the relevance of Crowdsourcing and its problem-solving nature. The author claims that despite its appearing enhancement of problem-solving, Crowdsourcing does not provide a substantive increase in firms’ competitive performance. The author especially focusses on the focal firm, which might compromise its position by risking revealing new products to competitors during crowdsourced problem-solving. The author suggests that research on Crowdsourcing could be conducted with the perspective of awareness-motivation-capabilities (Bloodgood, 2013).

Another managerial consideration is found in Zheng et al. (2011), who suggest that firms need to improve task design and the way they motivate the people that participate in Crowdsourcing contests if they intend to optimize solutions from the process. The authors propose balancing the usage of extrinsic and intrinsic motivation when seeking to encourage participation in Crowdsourcing and that: “Crowdsourcing contest tasks should preferably be highly autonomous, explicitly specified, and less complex, as well as require a variety of skills.” (Zheng et al., 2011, p. 57).

Recent studies have examined organizational capabilities in processing the suggestions they have solicited through a Crowdsourcing process (Piezunka & Dahlander, 2015). The authors claim, that “…organizations that do not handle filtering well may fail to tap into the full potential of Crowdsourcing.” (Piezunka & Dahlander, 2015, p. 858), when organizations face a large pool of suggestions, they can only attend to a subset of the suggestions due to limited attention. Organizations are thus more likely to pay attention to those suggestions that are familiar to internal knowledge pole, which contradicts the reasoning for pursuing external knowledge (Piezunka & Dahlander, 2015).

In the case of PharmaFX, the Challenge category is a Premium Challenge as classified by the crowdsourcing platform provider InnoCentive, here the problem is solved within 45–90 days. InnoCentive handles confidentiality matters and transfer of Intellectual Property Rights (IPR). When the Challenge has a scientific or technical scope, the company only pays rewards for solutions that meet specific criteria. The Challenge has to be “…a well-formed problem, whose solution has high value to an organization. It is specific, detailed, actionable, and delivers a measurable outcome” (InnoCentive, 2016). The Challenge can be idea generation, obtaining new ideas, or a complex technical or scientific problem solving. InnoCentive provides an expert that is suited for the specific Challenge, which helps PharmaFX formulate and articulate the Challenge, and the specific criteria for solutions suitability. The InnoCentive expert will throughout the period of online posting of the Challenge be of service to ‘solvers’, the external crowd member, who have questions or requests regarding the Challenge. When the Challenge deadline is reached the expert ranks the incoming submissions, to help the company with identifying the solution(s) that best meet the specific criteria and needs. When the winner(s) are found, InnoCentive verifies the solvers identity, obtains a Solver Agreement contract and any necessary employer waiver, transfer the award to the solver and finally handles the transfer of IP (InnoCentive, 2016).

Other sources of data are project description, facts about submissions in the Crowdsourcing Challenges and internal evaluation processes at PharmaFX. This data permits us some deeper insights into submissions, evaluation processes, methodologies for identification of the winners of the Crowdsourcing project and finally the reward allocations. In the following a short overview of the data is presented. In total, 58 submissions were sent to PharmaFX's specific Challenge, out of which, 13 were filtered and 45 were forwarded for the first evaluation round.

These 45 submissions were then grouped into different categories based on similar scientific concepts. The number of ideas in each category is presented in percentage in Fig. 1. Noticeable here is that many submissions are well established or had a nearby approach, whereas a small percentage is outside the direct scope of the Challenge.

Fig. 1.

Roughly grouped ideas.

(0.17MB).

The first initial rough evaluation is conducted by four persons in the evaluation team. They colour-coded each of the 45 submissions: red indicated a “bad” solution, yellow a “maybe good” solution and green “good” solution. Internally, both the problem owner and the scientific specialists did not find any of the submitted solutions as being a ‘good’ solution at this stage.

Further in the process, the information about the final evaluation of the submitted ideas is now grouped into the following four categories: “No goes”, “Interesting but out of scope”, “Potential, require further investigation” and “Relevant”. These categories are presented in absolute numbers in Fig. 2. It is noteworthy that the colours presented in Fig. 2 are representing the “Final conclusion traffic light” developed by the evaluation team during the evaluation phase. Of all submitted solutions, approximately 20% are of value to PharmaFX whereas around 80% cannot be utilized.

Fig. 2.

Final evaluation.

(0.28MB).

The evaluation guidelines describe the process of gathering two different kinds of Challenge winners, without communicating this separation to the community. The first type of winner is the one who submitted a solution for a specific need for PharmaFX. Accordingly, the idea must fulfil distinct success criteria. Further, the idea could have existed internally and if so, there is an extra value. Therefore, the idea will most likely be transferred to the problem owner.

The second type of winner is a creative submission, where the idea is not exactly what PharmaFX identified as a solution, but is creative, interesting or holds otherwise possible value, but not necessarily and exactly for this Challenge. This kind of idea should not exist internally.

Next in the process, the focus is on the rewards; how much it is, which type of category and what type of winner is rewarded. Building on the before introduced colour scheme, only ideas with “potential” (Yellow) and ideas which are “relevant” (Green) are awarded. Among the six Challenge winners, half were creative submissions and the other half had a specific value for PharmaFX. Only one of the “a relevant solution (total of two, as seen in Fig. 3) was rewarded and one submission with “potential” received $5000 instead of $2000 as the rest.

Fig. 3.

Rewards per category.

(0.28MB).

In the case of PharmaFX and the crowdsourced Challenge underlying this study, the management team decided not to reveal the sponsor of this Ideation Challenge. The Director of Open Innovation Department (DoOID) elaborated: “First of all there is an incentive not to say it is us because that will not bias the answers so you get more open answers, and secondly it might not be relevant for them to know it is us”. This means that none of the participants knew the identity of the sponsor, nor the identity of the successful submission rewarded at the end of the challenge.

Several researchers have pointed out that using Crowdsourcing to solve tasks can be cheaper than solving them internally. This is also a key argument in TCT, which proposes that firms need to reduce costs to raise their profitability while maintaining their competitive advantage. Hence, it is interesting to see what role cost reduction plays in PharmaFX.

Two categories of costs emerge when looking at PharmaFX's scientific Ideation Challenge. One category is the costs related to the intermediary. PharmaFX collaborated with InnoCentive as an intermediary for their Crowdsourcing Challenge. Two types of costs are associated with InnoCentive as the intermediary: (1) the cost of service provided by InnoCentive and accessing their online platform and (2) the monetary rewards given to the six people that are rewarded for their submissions.

The other category of costs is the one related to the internal resources deployed at PharmaFX. While the costs stemming from the collaboration with InnoCentive are evident from the contractual agreements, the costs arising from the deployment of internal resources at PharmaFX are less so, and no specific accounting for the internal resources at PharmaFX is evident in the data. Nevertheless, the organizational practices and consequences stemming from organizing and handling the Crowdsourcing must be addressed.

The costs of evaluating the submitted proposals and picking the winners puts a high demand on internal resources. The evaluation process started as soon as the Crowdsourcing Challenge ended on InnoCentive's platform. The open innovation team allocated one month to evaluate all the submitted solutions and give senior management feedback. Several evaluation rounds were conducted because the project members required further investigation in discovering the value of some ideas deemed dubious at the early stages. Another reason for the lengthy process was that evaluations of submitted ideas were done according to sets of different criteria. This resulted in some ideas getting a reward, even though the idea would not be implemented at PharmaFX (Fig. 3). As such, these costs could be a direct loss for PharmaFX.

It is arguably a paradoxical outline for a pharmaceutical company such as PharmaFX to crowdsource, while at the same time trying to maintain a certain level of confidentiality. There is a trade-off between Crowdsourcing and keeping the status quo. Evaluating the submissions provided by solvers in a Crowdsourcing Challenge can be time-consuming (Baldwin & Von Hippel, 2011; Pisano & Verganti, 2008), and as such, costly for the host organization. This is especially the case when there is a high degree of uncertainty related to the evaluation (Ye & Kankanhalli, 2015), and this can, in turn, lead to an increase in the overall Crowdsourcing cost for the firm. The question regarding the degree of uncertainty might lead to a lengthy evaluation process and as seen in the analysis, the evaluation process was a large effort. PharmaFX was extensively engaged in the assessment, which arguably resulted in a high expense. At the intermediary partner, the number of submissions was not extraordinary for similar Crowdsourcing Challenges; in fact, it was within the expected range.

The submissions’ high degree of uncertainty was arguably also challenging. The submissions were theoretical ideas, which put extra demand on the evaluating team members because they investigated the submissions. Requesting ideas in a Crowdsourcing Challenge entails a higher level of uncertainty, relative to Crowdsourcing of a specific product.

The evaluation process might have been more complex and time-consuming because each submission was thoroughly evaluated, in addition to carefully investigate the viability of integrating the submission into the organization. During the evaluation, it became clear that the submissions needed to be evaluated from two different perspectives. First, solvers needed to be rewarded for submissions that fulfilled the officially illustrated success criteria of the Crowdsourcing Challenge. Secondly, the organization required additional competences to exploit the ideas. Interestingly, it was a surprise for the organization that the two perspectives did not necessarily correlate. As such, ideas that already existed internally, or ideas that were regarded unsuitable in the company, could still fulfil the deduced success criteria for the Crowdsourcing Challenge. Surely, each submission needed to be evaluated critically and thoroughly, however, the unawareness about the discrepancy, related to evaluating the submissions from two non-correlating perspectives resulted in lengthy evaluations process. It can be argued, that the high degree of uncertainty represents the root cause. As such, this argues for an increase in the overall transaction costs for the Crowdsourcing Challenge.

There were conflicting explanations on how the evaluation round was executed. The director illustrated that the evaluation round was not formalized, and there were not any fixed guidelines provided to the evaluation team members. This is conflicting with a description of the methods employed by InnoCentive in their collaboration efforts with focal companies in their Crowdsourcing Challenge initiatives. The former CEO of InnoCentive predominantly claims that companies are helped with formulating the evaluation success criteria (Allio, 2004). However, he also asserts that the organization might have employed certain criteria at the end of the evaluation. This indicates that the evaluation round ought to be well planned and organized, to avoid a prolonged evaluation process with high evaluation costs.

This study aimed to explore the benefits of Crowdsourcing gained by PharmaFX and has identified the specific benefits which can be obtained, based on the case. Particularly the study is considering the benefits of reduced costs, increased brand visibility and access to specialized skills and their applicability to the pharmaceutical industry. The general academic literature of Crowdsourcing, specifically in the pharmaceutical context, is found to be not conclusive on more than a few vital questions concerning these benefits.

Synthesizing the findings, PharmaFX could gain access to specialized skills and solution diversity. The broadness of the Ideation Challenge allowed greater overall participation, which resulted in comparable greater access solution diversity. The increased access to specialized skills and solution diversity became evident in the amount and difference of the submissions. Concerning cost reduction, we found that Ideation Challenges are subject to relatively high costs in evaluation and codification, which negatively affects the overall cost reduction. However, the Ideation Challenge did not necessarily aim to reduce costs, but rather to produce new and different knowledge.

Using Crowdsourcing for a specific scientific problem would arguably always entail the need for solvers with a very specific background. Furthermore, given that the crowdsourcing might only be a small piece of the puzzle of a larger pharmaceutical R&D process. One can probably also assume, that it may be necessary to provide solvers with a larger amount of information for them to be able to provide suitable solutions.

Managers can constrain the number of generated ideas by casting a smaller net. This could be achieved by selecting a smaller group of external contributors, where a viable scenario could be identifying contributors most likely to provide distant knowledge. It is also suggested that organizations, before pursuing Crowdsourcing, establish procedures to facilitate the processing of large quantities of suggestions, define a goal ratio for how many ideas to implement, and establish criteria for how to prioritize suggestions.

In the context of Ideation Challenge, it is the sponsor looking for a greater amount of diversity, not only in the crowd but also in the solutions. The broadness of the Challenge is necessary, though it comes with greater difficulty in phrasing, resulting in greater codification costs. Furthermore, while this broadness of Ideation Challenges might increase the access to specialized skills and diverse solutions, it also increases the number of overall submissions, leading to increased evaluation costs. This can potentially result in Crowdsourcing not being as cost efficient as expected, and even have no cost reductions.

Pharma companies can benefit from opening the innovation processes, since this enables leveraging externalized knowledge, while managing the risk of competitors exploiting the same openness. Great potentials can be gained for pharma, if the fear of openness is managed.

As with all research, it is important to recognize limitations of this study. The single case, albeit explorative, of necessity limits the generalizability of the findings. Yet this unique case offers insights about the phenomena and invites to new research territory (Patton, 2002). Further restrictions of this study are regarding the potential long-term effects of the Crowdsourcing initiative.

Further research can consider the potential of cost reduction from scientific Ideation Crowdsourcing Challenge compared to internal scientific Ideation generation. Other external search alternatives, which have the same benefits as scientific Ideation Challenges, to a comparable price. Finally, longitudinally study assessments of the value creation from scientific Crowdsourced problem solving.