Appropriability Mechanisms and the Platform Partnership Decision: Evidence from Enterprise Software

The efficacy of intellectual property rights (IPR) as a mechanism to appropriate the returns from innovation has been studied with mixed empirical findings. Following prior literature, this research examines whether ownership of IPR and downstream capabilities such as marketing capability is an effective mechanism to encourage the owner of innovation (or so-called ISV) to join proprietary platform and prevent platform owner to expropriate the innovation. The study is conducted in the environment of software industry, and SAP is the software platform investigated.

The paper proposes that both of the two anti-expropriation mechanisms will increase the probability of ISV partnering with the software platform. It also propose that strong downstream capabilities, therefore, will weaken the value of intellectual property protection mechanism. Finally, it posits the value of IP protection will be enhanced if the markets served by the ISV grow rapidly. In this research, partnership with software platform is indicated using a binary data based on ISVs’ decision to become certified by a platform or not. Downstream capabilities is measured by trademarks and consulting services, and market growth is measured by sales growth. The primary model is built upon a survival model. However, for the purpose of robustness check, it also use generalized estimating equation model and linear probability model with fixed effects, as well as alternative measurement for key variables. Three hypotheses are all verified.

Information Technology and Intangible Output: The Impact of IT Investment on Innovation Productivity

Since the prior research focused on the contribution of IT to firm-level productivity and less research concern about the relationship between IT, knowledge creation, and innovation output, Kleis et al. (2012) try to identify the contribution of IT to innovate production across multiple contexts using a quality-based measure of innovation output.

Kleis et al. (2012) address the conceptual model of knowledge production that based on the knowledge production function (KPF) shown as the following (Figure 1. P47):

Kleis et al. (2012) collected a panel dataset comprised of annual information on innovation spending research and development (R&D), IT spending, and citation-weighted patents for large U.S. firms between 1987 and 1997. For empirical analysis, Kleis et al. (2012) build two sets of analyses to analyze more than 1800 observations: First, involving a thorough examination of the core research questions. Second, involving further analysis of the impact of IT on innovation output.

The main contribution of this study: (1). The estimation results indicate that IT and R&D both play positive and significant roles in innovation production. (2). The results highlight that R&D has a positive impact on citation-weighted patent output, meanwhile, the firm size is inversely related to patent output. (3). The final results demonstrate an important aspect of IT and the innovation creation process (i.e. networks, e-mail, telecommunication, etc. also contribute to the innovation process). (4). The estimates of the return to investment for IT-driven innovation appear to be economically meaningful and have market value impact.

Captive centers are wholly-owned by the multinational corporations (MNCs) to conduct the R&D and new product development offshore, which also refers to captive offshoring.  Previous research indicates that the performance of distributed work is adversely impacted by failures of cooperation, i.e., misaligned incentives, as well as failures of coordination. The authors analyzed survey data from 132 R&D CCs established by foreign multinational companies in India to understand how firms execute distributed innovative work.

There are two generic categories of coordinating mechanisms for building and maintaining common ground, namely, information sharing and modularization. The information sharing strategy involves ongoing communication between interdependent agents to dynamically update common ground. A modularization strategy involves limited ongoing interaction between the agents. The author aim to examine the moderation role of three task attribution: task routineness; task analyzability and task familiarity. The hypothesis are listed below:

Hypothesis 1A (H1A). With high interdependence,
higher levels of modularization will increase performance
only when task routineness is high; when task routineness is low, higher levels of modularization will decrease
performance.

Hypothesis 1B (H1B). With high interdependence,
higher levels of modularization will increase performance
only when task analyzability is high; when analyzability is low, higher levels of modularization will decrease
performance.

Hypothesis 1C (H1C). With high interdependence,
higher levels of modularization will increase performance
only when task familiarity is high; when task familiarity is low, higher levels of modularization will decrease
performance.

Hypothesis 2A (H2A). With high interdependence,
information sharing as a coordination mechanism will
increase performance only when routineness is low; when
routineness is high, higher levels of information sharing
will decrease performance.

Hypothesis 2B (H2B). With high interdependence,
information sharing as a coordination mechanism will
increase performance only when analyzability is low; when
analyzability is high, higher levels of information sharing
will decrease performance.

Hypothesis 2C (H2C). With high interdependence,
information sharing as a coordination mechanism will
increase performance only when task familiarity is low;
when task familiarity is high, higher levels of information
sharing will decrease performance.

Regarding the first set of Hypothesis.  Hypothesis 1A and 1b are supported, showing that Modularization improves performance only when routineness and analyzability are high. In terms of the second set of hypothesis, results are consistent with the interaction plots, which show that on average, information sharing improves performance for R&D work. However, a significant difference in slopes between high and low levels of task attributes is observed only in comparing projects of low versus high familiarity. The difference in slopes for the other
task variables is insignificant. Only Hypothesis 2C is supported.

The key contribution of this study lies in the careful explication of the conditions under which choice of investments in modularization versus information sharing yield different performance outcomes.It shows that the success of this strategy has to be evaluated based on the nature of the underlying work.

The authors examined the basic Internet adoption for reducing the coordination costs of geographically dispersed firm teams. They hypothesized that adoption of basic Internet will be associated with an increase in the likelihood of collaboration for multi-inventor, geographically dispersed teams (hypothesis 1); Adoption of basic Internet will be associated with a smaller increase in the likelihood of collaboration for single-location multi-inventor teams than for geographically dispersed teams (hypothesis 2); and Adoption of basic Internet will be associated with a smaller increase in the likelihood of collaboration for single inventors than for geographically dispersed teams (hypothesis 3).

For analysis, they used fixed effects panel data model. Their findings suggest that there is no evidence of a link between Internet adoption and within-location collaborative patents. There is no evidence of a relationship between basic Internet and single-inventor patents. These evidences show that basic Internet adoption lowered the coordination costs of geographically dispersed research teams. This does not have to be interpreted as easier access to electronic knowledge systems or shared resources rendered an increase in research output of those who adopted basic Internet. .

The result suggests that IT can be used to integrate geographically dispersed operations, either obtained through acquisition or deliberately dispersed. The limitation of the study oriented from the sample data is one sample size and one period of time within U.S. When they design their international R&D organization, firms mostly regard the following options: a centralized organization that provides higher control and decentralized structure that enables local knowledge resources but increasing coordination cost. However, the study result supports that IT investments may substantially alter the trade-off and decentralized model more attractive, which encourages a more distant R&D activities within firms.

Technology advances in distributed work lead to a growing phenomenon that increasing number of firms set up offshore captive centers (CCs) to carry out R&D work. In this context, Mani et al. (2014) studies how organizations coordinate distributed knowledge intensive work (e.g., R&D).

Prior literature has suggested two generic coordination mechanism, information sharing and modularization, but do not reach any conclusion about the superiority of one of these mechanisms for any given task. This study address such research gap by raising a question, what combinations of task characteristics and coordination mechanisms yield high performance in the context of interdependent knowledge work?

The authors analyzed survey data from 132 R&D CCs established by foreign multinational companies in India to understand how firms execute distributed innovative work. They found that 1) modularization of work across locations is largely ineffective when the underlying tasks are less routinized, less analyzable and less familiar to the CCs; and 2) information sharing across locations is effective when the CC performs tasks that are less familiar to it.

The primary contribution of this study is to demonstrate high performance work configurations in offshoring of R&D and product development work, given that the tasks analyzed are highly interdependent. Its results expand studies on organization and coordination of distributed work as well as for practitioners who want to improve the performance of their distributed R&D strategies. This study has following weakness: the cross-sectional analysis may limit its ability to establish causality, the measures are perceptual and the unobserved heterogeneities are not controlled.

From Wires to Partners: How the Internet Has Fostered R&D Collaborations Within Firms

Motivation:

Historically, collaborative work has been hampered by the existence of significant coordination costs that increase with team size, geographic dispersion, and heterogeneity of team composition. It is widely believed that the adoption of information technology increase the returns to collaborative work. However, little systematic empirical work on the implications of IT investment has been done for industrial research.

Research Question:

The main hypothesis of this paper is that by reducing the coordination costs of collaborative work, investments in IT will be associated with an increase in the likelihood of geographically dispersed, multi-inventor collaborative research teams relative to other types of research teams (including output from single location and lone inventors).

Main Findings:

They find that basic Internet adoption is associated with an increased likelihood of collaborative patents from geographically dispersed teams, which is in contrast with previous finding that IT adoption leads to a disproportionately greater increase in collaborations among researchers who are geographically close to one another. On the contrary, they find no evidence of such a link between Internet adoption and within-location collaborative patents nor single-inventor patents.

Hightlights:

The method they use to address the assumption that Internet adoption is exogenous.

• First, reverse causality. They utilize the timing of Internet adoption as the source of a falsification exercise. They find no evidence that the incidence of cross location research collaborations is correlated with a location pair’s future adoption of Internet technology; that is, location pairs who adopt Internet technology experience no increase in the likelihood of a collaborative patent prior to adoption.
• Instrumental variables. We employ two sets of instruments that capture local variance in the costs to adopting Internet technology. The first addresses cross-sectional differences in local regulatory conditions that will shape the costs of purchasing Internet access. The second captures cross-sectional differences in familiarity and expertise with the Internet in the local regions where the establishments reside.

Setia, Rajagopalan, and Sambamurthy, and Calantone (2012) point out that the contribution of peripheral developers in open-source software projects has not been well understood. The authors in this study attempt to reveal the impact of peripheral developers, in comparison with core developers, on product quality and diffusion and the moderating effect of OSS product life cycle.

The authors find that peripheral developers endow a greater impact on the quality assessment because they supply independent, novel, and unique insights for detecting bugs in the projects. Peripheral developers also contribute to expedite the product awareness and adoption owing to their interpersonal network and word-of-mouth communication; since the personal network of peripheral developers does not suffer from institutional influences or corporate promotions, the information shared through the network tends to receive better credibility, leading better exposure and adoption for potential users.

Interestingly, the positive impact of peripheral developers is not supported for quality enhancement in the study’s empirical analysis. The authors suggest, a larger project may require complex coordination efforts which may prevent active participation of peripheral developers. Also, the lack of efficient coordination mechanism can be another culprit to accommodate geographically dispersed contributions from the developers.

A mature product may represent that its OSS product has been well managed with high quality. This signaling effect can be a good explanation for the authors’ findings: the participation of peripheral developers at a mature stage influences positively on quality assessment, product awareness, and product adoption greater than at earlier stages.

In conclusion, the authors show that the contribution of peripheral developers takes up a significant portion of an OSS outcome, especially combined with the moderating effect of project life cycle. Strategically using this dynamic is a mandate for an OSS organization or a corporate, which aims to leverage the synergy based on the volunteering dedication of peripheral developers.

How Peripheral Developers Contribute to Open-Source Software Development

This paper is motivated by the different contributions of Open-Source Software (OSS) Development made by core developers and peripheral developers. Authors of this article explore the role of peripheral developers in OSS product quality and diffusion and how these effects vary across the product life cycle of OSS projects.

Peripheral developers’ motivations and contributions are different from core developers of an OSS project. First, peripheral developers are motivated to contribute the product quality for their own consumption or demonstrate their adherence to the certain community. Second, software adopters rely on the information from their community, and peripheral developers can act as knowledge sources to help the spread information about OSS product. Thus, the paper argues that compared with core developer, peripheral developers have more contributions to the OSS product quality and diffusion. And these effects are more salient in the product mature stage than in the early stages.

Empirically, the paper collects 1966 monthly observations over 147 products. Because of the nested nature of the research design, the paper uses hierarchical linear modeling (HLM). Model1 assesses the variability in diffusion and quality in each level. Model 2 identifies the direct impact of the predictors. And Model 3 tests the cross-level effects at the periodic and product levels. The empirical results showed that peripheral developers’ contribution to product quality assessment is more salient than core developers. And, the peripheral developers’ participation significantly influences product diffusion. Finally, their influences on product quality and diffusion are various across the product life-cycle stages.

Moreover, the paper examined the paradox of peripheral developers’ contribution and found that project size and coordination may moderate the effects of peripheral developers on product quality enhancement. In all, this paper revealed the different roles of developers and differentiated the contributions made by core developers and peripheral developers for OSS projects.