Revisiting the Research Questions

In his recent TED talk, Kenneth Cukier provides a succinct, yet accurate answer to the overarching research question of this study, How is value generated from open data?:

“You have more information. You can do things you couldn´t do before” In an attempt to provide more theoretical depth to this answer, I have suggested that two overarching value generating mechanisms can explain how value is generated from open data, the information sharing mechanism and the market mechanism. New value is generated through new information that contributes to increased transparency, and reduces the adverse effects of information asymmetry. Open data can also drive civic engagement through members of society engaging with and enriching available information, using it to positively impact their decision making capabilities, and to align their behavior to practices that contribute to sustainable value. Alternatively, value is generated through the markets when data are used to support process improvements, resulting in increased efficiency and cost reduction, or as a resource in new products or services, resulting in innovations that can transform markets, creating new companies and job opportunities.

I will now examine the five sub-questions in greater detail. The first sub-question posed was: What are the main enabling factors for value generation through open

data? After synthesizing between literature and qualitative and quantitative data, I have identified four measurable, contextual factors I propose are an important part of a societal level infrastructure that is conducive to value generation through liquid open data. These factors are: 1) robust regulatory data and privacy protection frameworks; 2) digital leadership of government; 3) cost of high-speed networks; and 4) ease of reaching a skilled workforce. Moving down one level of abstraction, other factors will be important for individual organizations. The qualitative data collected for Paper IV revealed one particularly important enabling factor, namely absorptive capacity.

Absorptive capacity is a function of organizational and technical capabilities, which allow organizations to reach outside of their boundary for data, information, or knowledge and to productively utilize these external resources for value generation. It remains a research opportunity to extend the conceptual model presented in this dissertation to include organizational level variables and consequently test the boundary-crossing relationships, using observations at both the societal and organizational level.

The second sub-question asked was: What are the unique features of open data? The answer is in fact two-fold. The first concerns the theoretical economic features, which include open data being (in theory) non-rivalrous and non-excludable. These features have a direct impact on how value is generated from open data, as they influence how market participants view open data as a resource. While the openness factor offers a clear value proposition at the societal level, the value proposition is more complicated for individual companies, which generally desire unique resources (Barney, 1991;

Wade and Hulland, 2004). Lindman and Kuk (2015) suggest that open data should be viewed as a common pool resource, subject to subtractability, as the feature of subtractability acts to encourage commercial use and investment. OECD (2014) utilizes the terminology infrastructural resource to highlight that digital data can be easily reproduced and reused by many stakeholders for many different purposes.

Unfortunately, we still do not have a widely accepted economic concept that can fully describe the unique features of open data.

The second answer lies in the more detailed conceptualization of liquid open data, where I have synthesized different definitions of open data. Diverging from those definitions, I have increased the emphasis on the features, or attributes, that reflect the technical and governance related complexity of making data liquid and open. I would also like to draw attention to the fact that open data can mean many things to many people. Each of the seven dimensions of liquid open data has different relevance for different use cases. I propose, however, that governments should strive to make data as

liquid and open over all of the seven dimensions, as is economically feasible. Doing so will create a bigger opportunity or, alternatively, a more valuable option, potentially resulting in synergistic value generation.

The third sub-question posed was: What are the value generating mechanisms of open data? To which I identified four distinctive value generating mechanisms. As these mechanisms form a hierarchy, I have further categorized two as market mechanisms:

efficiency and innovation mechanisms. These two are already widely recognized as value generating mechanisms. However, as a contribution of this study, I propose two additional mechanisms that do not utilize the markets to generate value. As the overarching mechanism guiding these was not previously recognized, I chose to conceptualize it as an archetypical value generating mechanism and named it the information sharing mechanism. The information sharing mechanism explains how value is generated through positive network externalities that are created when multiple stakeholders share valuable information, effectively reducing information asymmetry and encouraging civic engagement.

The fourth sub-question raised was: How can we identify, conceptualize and measure the value that is generated from open data? This question raises a number of complexities and the answer is manifold. Primarily I intended to highlight the immense value that is both generated and captured from open data, yet never enters formal accounting ledgers. For example, access to superb education, clean air and an efficient and supportive healthcare system are amongst the things valued highly by many, yet this value is not easily quantifiable. Moreover, when I started to work with the secondary data, I saw that there was a high correlation between wealth and the other chosen value indicators such as health, education and sustainable environment. It is, however, not a perfect correlation, and the resulting sustainable value construct rates countries rather differently than the GDP measure. For instance, Qatar ranks number 3 out of 76 countries in GDP dollars per capita but number 29 in sustainable value.

Norway, however, ranks number 1 in GDP dollars per capita and number 2 in sustainable value, due to their more balanced approach to information sharing and market-driven innovation. United States rank number 8 in GDP dollars per capita but 23 in sustainable value. Thus, the model does confirm that the countries that value openness and information sharing tend to rank highly in sustainable value as well.

Regarding measurement, it is extremely difficult to trace how much value is generated from use of open data or how much these data are worth. There are many different methods that can be, and are currently used, each of them with their own strengths and limitations. Firstly, we can use previously created nationwide indicators to measure

sustainable value. There are a number of recently developed indexes that offer a nuanced and multi-dimensional view of societal progress to draw on. For instance OECD Better Life index, United Nations Human Development index and Social Progressive Imperative’s Social Progressive index. The difficult part, however, is to adequately link the act of providing society with open data to the generation of sustainable value. One possible approach is to use correlational, cross sectional comparisons as was done in this dissertation. This method certainly offers a valuable indication, but remains silent in terms of direct causality. Therefore, I have opted to use the CR-based method of retroduction to hypothesize about the possible underlying causal mechanisms that could explain the relationships illustrated in the model.

One method that has been used in previous studies is the cost benefit analysis. This type of analysis evaluates the costs and benefits of opening a single dataset, or type of data (Houghton, 2011). Use of the cost-benefit analysis method has revealed that the socio-economic benefit of making commercially attractive data, such as geographic data, available is typically higher than the costs of making said data available. Cost benefit analysis is however a difficult exercise in the case of open data as the data are used by multiple unknown stakeholders for a variety of purposes. Informed by this research, it is my conclusion that the most comprehensive way to obtain an estimation of the use of data for various purposes is through a bottom-up approach, as is employed for calculating the GDP. One possible way to understand use of open data in the private sector is through self-reporting, a method used by the GovLab´s Open Data 500 index. As a second approach, the Danish BDP has adopted a reporting template where individual organizations use a predefined method to report on both tangible and intangible benefits they observe from the use of data. As a third approach, the Danish Geodata Agency used a survey methodology to create a baseline estimate of use of geographic data, before they were made free-of-charge in January 2013.

However, the cost-benefit analysis approach is by definition an ex-post method. From my identification of the open data value paradox, I conclude that we need to evaluate ex-ante the potential benefits in relation to various scenarios consisting of different configurations of enabling factors or barriers. To accomplish this I have suggested the use of option value methods (further discussion in Paper VIII). Option value thinking can benefit from evaluating opportunity costs: how would the geo-service industry have developed if the USA had not made GPS data available? Alternatively, how would our medical knowledge have progressed if genomics data had not been shared? I propose that this way of thinking will draw our focus towards the importance of

creating the right environment for individuals, businesses and governments that would like to utilize open data, providing them with an option for value creation.

The final sub-question raised was: What are the key implementation strategies and business models that can promote long-term generation of value from open data?

This is in fact two questions in one. I will first address the part that considers key implementation strategies.

I propose that for most ODIs, the seven dimensions of liquid open data can be used to benchmark the desired outcome. I expect that more standardized, more interoperable and more readily reusable data are preferred everywhere. However, as ODIs operate in very different contexts, appropriate strategies for implementing liquid open data can vary considerably. In some countries, the need to digitize data might introduce a first implementation barrier, while in other countries the challenge revolves around making already available data open for external use. These two cases introduce very different implementation barriers. For instance, Estonia, one of the frontrunners in E-Government, has in fact benefitted from having next to no pre-existing systems, thus being able to develop them effectively from the start. This in contrast to a country like Denmark who struggles with outdated legacy systems, legislation and a soft infrastructure that is socially entrenched and difficult to change. Denmark, on the other hand, benefits from high capacity and professionalism in data collection and widespread use of central sources of data across government. In the specific case of Denmark, explored in Paper VI, the main governance challenge was finding a balance between the safeguarding the autonomy of existing data custodians and ensuring an overall coordination between program participants, including use of common standards. Accordingly, I suggest that the approach the BDP adopted could be suitable for other ODIs with a strong focus on government efficiency through open data and interoperability, operating within a system that already supports cross-government collaboration.

The second part of the question concerns business models. The answer to this part remains in observation, however the business models behind two-sided markets or MSPs do provide a hopeful avenue of exploration for creating sustainable value, as discussed in Paper VIII. These business models have solved the complications associated with capitalizing on the intangible value created through information sharing mechanisms. The value inherent in the provided content attracts the non-paying stakeholders to the platform. Moreover, this group of users incentivizes the other group of stakeholders, the other side so to speak, to use the platform. The latter group, or the other “side” of the platform, utilizes market mechanisms create value for

money or new income, and thus contributes to the financial sustainability of the platform.