From Data to Decision Intelligence: The Potential of Decision Accelerator Labs

We live at a moment of perhaps unprecedented global upheaval. From climate change to pandemics, from war to political disharmony, misinformation, and growing social inequality, policy and social change-makers today face not only new challenges but new types of challenges. In our increasingly complex and interconnected world, existing systems and institutions of governance, marked by hierarchical decision-making, are increasingly being replaced by overlapping nodes of multi-sector decision-making.

Data is proving critical to these new forms of decision-making, along with associated (and emerging) phenomena such as advanced analytics, machine learning, and artificial intelligence. Yet while the importance of data intelligence for policymakers is now widely recognized, there remain multiple challenges to operationalizing that insight–i.e., to move from data intelligence to decision intelligence.

DALL-E generated

In what follows, we explain what we mean by decision intelligence, and discuss why it matters. We then present six obstacles to better decision intelligence–challenges that prevent policymakers and others from translating insights into action. Finally, we end by offering one possible solution to these challenges: the concept of decision accelerator labs, operating on a hub and spoke model, and offering an innovative, interdisciplinary platform to facilitate the development of evidence-based, targeted solutions to public problems and dilemmas.

1. Why Decision Intelligence Matters

It is now something of a truism that data can help make better decisions. But while this insight is generally accepted, there is often a lack of clarity about the specific ways in which it is true. In order to better understand why decision intelligence matters–and to overcome obstacles–we need a better understanding of its key value propositions.

Broadly, decision intelligence can:

• Ensure that insights derived from data are more effectively integrated into decision-making processes. This can lead to policymaking that is more responsive and effective, less biased, and characterized by greater situational awareness.
• Allow decision-makers to employ advanced data analytical methods and state-of-the-art simulation technologies. This, in turn, can allow them to better understand and address the complex, interconnected challenges that characterize our era.
• Help incorporate lived experience and thick data into the decision-making process, resulting in more empathetic, realistic, and effective policies and interventions.
• Foster collaboration between diverse stakeholders, breaking down silos, and facilitating the development of comprehensive, cross-disciplinary solutions.

2. Challenges to Decision Intelligence

Despite this list of benefits, several obstacles remain to transform the potential of data intelligence into the streamlined processes and improved outcomes that can result from it. These obstacles limit the possibilities of data and information technology, and more generally constrain the potential positive impact of 21st-century decision-making.

A (partial) list of challenges to decision intelligence would include:

• Lack of awareness of data’s potential: Decision-makers frequently operate within the confines of their respective domains and existing paradigms, leading to a fragmented understanding of the broader context and potential interdependencies between sectors or domains. This can lead to a limited understanding, or even a certain wariness, of the role of data in decision-making. More emphasis on building awareness–e.g., through training and educational outreach–could be helpful.
• Poor Problem Definition: Effective decision-making relies significantly on how problems are framed–i.e., what questions policymakers ask, and how challenges are formulated or understood. These challenges are perhaps more significant when it comes to datafied decision-making. Given the glut of available data, there is a danger that policymakers may address challenges not based on genuine social need or priorities, but simply because the “data is available.” To address such challenges, we have elsewhere advocated for a new science of questions that would help frame challenges and opportunities, and identify the most important priorities.
• Lack of technical capacity: Even when decision-makers would welcome greater use of data, they are often hampered by a lack of technical capacity. This inadequate capacity can manifest in various ways: a lack of familiarity with advanced analytics and data-led decision-making; an absence of necessary tools and frameworks to effectively capture decision requirements; or poor information visualization methods, manifested for instance by static dashboards and a lack of real-time data.
• Lack of advanced methods: Technology moves fast and even data-led decision-making methods often rely on outdated linear and reductionist approaches that fail to capture the complexity and interconnectedness of modern challenges. Such limitations require greater incorporation of advanced methods– such as systems or topic mapping, machine learning, and network analysis–in order to uncover hidden patterns, relationships, and insights that could inform more effective decision-making.
• Limited inclusiveness and stakeholder engagement: Traditional decision-making processes often exclude or inadequately consider the perspectives and experiences of those directly affected by the issues being addressed. Such exclusions can be all the more pronounced–and problematic when it comes to data-led decision-making and can apply as much to the decision-making processes themselves as to the underlying data (which may lack representativeness or contain biases). A lack of inclusiveness leads to less empathetic and effective solutions that may not fully address the needs and concerns of diverse stakeholders–and, therefore, to a lack of trust and less effective policies.
• Fragmented, siloed approaches: Despite the proliferation of data in our economy and society, much of it remains in silos and behind paywalls. Decision makers’ limited ability to combine data and merge insights across sectors and domains is a serious constraint on their ability to maximize the potential of data to create better and more responsive policies. A greater use of structures such as data collaboratives could be helpful in this regard and allow for the development of comprehensive, cross-disciplinary solutions to address complex, interconnected problems.

3. Decision Accelerator Labs

The above challenges represent real obstacles to fulfilling the potential of data in decision-making by policymakers and others. Some possible solutions have been mentioned–e.g., greater use of data collaboratives to overcome data silos, a new science of questions to help establish priorities. In this section, we propose the creation of innovative institutional structures that we call Decision Accelerator Labs (DALs).

DALL-E generated

DALs should function as connectors, bringing together different stakeholders that play a role in the decision-making process. These stakeholders can include parties that need to be consulted or informed (e.g., citizens); parties who play a role as validators (e.g., domain experts or those with lived experience); translators (who translate information or data into meaningful action); and of course decision makers themselves.

DALs can take many forms. They could for instance, take the shape of a real-world think tank; or perhaps virtual immersive and interactive decision theaters (see, for instance, the Decision Theatre efforts at Arizona State University) that provide a conducive environment for decision-makers to visualize data, models, and scenarios. In all cases, DALs should follow a flexible and adaptable spoke-and-hub model, allowing for the creation of tailored decision-making environments that can address diverse challenges across different scales, sectors, and issues (e.g., at the global or regional level, or in a particular domain). By enabling DALs to cater to diverse contexts, the spoke-and-hub model thus ensures that stakeholders have access to the most relevant and effective resources, fostering informed decision-making and enabling more targeted solutions.

DALs would help decision-makers (as well as those affected by their decisions) in several ways. Some of their key enhancements would include:

1. Question science: By bringing together experts and other stakeholders, DALs can play a key role in advancing participatory question science to ensure decision-makers are asking the questions that matter most. In this exercise, DALs can leverage and build upon our 100 Questions methodology.
2. Decision mapping and requirements: DALs can develop new methods to identify decision requirements, as well as where there is the greatest need for decision support and what tools or systems could offer such support. This process involves mapping existing decision systems and needs to identify priorities.
3. Leverage Advanced Data Analytical Methods: DALs can employ new data analytical techniques, such as machine learning, artificial intelligence, advanced simulation, and network analysis, to uncover hidden patterns and relationships within the data. These innovative methods will help identify relevant factors and variables, enabling more accurate predictions and actionable insights; in so doing, they will lead to more effective and responsive policies, and help define the boundaries of 21st-century decision-making.
4. Integration of Lived Experiences and Thick Data: Recognizing the value of firsthand accounts and real-time data, DALs can help incorporate lived experiences and thick data (granular, high-frequency data) into decision-making processes. It will enable such processes through robust thick data-driven methods such as digital ethnography combined with big data-driven insights. This approach will ensure that decisions and policies that emerge from the DALs reflect ground realities and contribute to more inclusive, empathetic, and effective solutions.
5. Rapid Deliberation and Iteration: Relying on real-time data, advanced simulation and analytics, and other methods, DALs can incorporate feedback from experts and policymakers to improve upon its outputs and models iteratively and quickly. In addition, collective intelligence methods and tools can be used to evaluate the impact of policies and programs and test ideas through simulations.

Conclusion: Looking Forward

DALL-E 3 generated

The creation of DALs could offer a promising if incipient, means of advancing toward better decision intelligence, which in turn can lead to more responsive and effective policies and decision-making. In order to fulfill this promise, further research and experimentation are required, both into the specific potential of DALs and the more general opportunities (and challenges) of data-driven decision-making. Join us toward that end.

(Thanks to Alex Fischer with whom I had several conversations on the topic; See and contribute to also our Special Track on Collective Decision Intelligence https://dataforpolicy.org/sp8/)