The Research Startup

Scientific research is a pillar for long-term human prosperity. Finding new ways to solve problems by leveraging science and technology are required to grow a sustainable economy, raise societal quality of life, and create opportunities for future generations. The way we organize and incentivize research over the last few decades have created gaps and magnified inefficiencies in our efforts. Economists have shown evidence that research productivity has been falling since the 1970s. We train many more researchers today than any other time in history, yet their collective output and impact on prosperity has stagnated. Many sensible analyses point out ways to improve government policies, industrial strategies, and existing institutions to close the productivity gaps. We believe that, in addition to these changes, we must also create new institutions with new incentives in how we conduct R&D.

The Model

To revitalize the research ecosystem and effectively promote a free flow of ideas, people, and resources, we need to create systems that operate differently than traditional institutions of research such as academia, large corporate labs, and government agencies. We propose a new community of research initiatives based on research startups - focused and independent research teams organized like startup companies. Anomaly Research will act as a parent organization to a portfolio of research startups and manage resources to invest in new research teams. These teams will operate under a new set of incentives:

• Real capabilities
  Researchers are aligned to achieve an ambitious mission and deliver tangible products and services to the real world.
• Focused and flexible teams
  Teams recruit the best talent to accomplish the mission. Accountability is concentrated. People with unconventional backgrounds and experiences are sought out for new perspectives.
• Adaptive structures
  Research startups are structured as independent organizations with fit-for-purpose funding and governance.
• Intelligent risk-taking
  Researchers are given freedom to take risks and learn from experience. Negative results are not punished and suppressed, but analyzed and shared.
• Systemic cooperation
  Problems encountered by one team is an opportunity for another. Teams solve each other’s challenges, share resources, and advise on lessons learned.

An Eye On Utility

One of the defining features of a research startup is the commitment to a specific mission to deliver new capabilities with clear outcomes and milestones. In contrast to academia, researchers are not organized around bodies of knowledge but rather the pursuit of new products and services. New ideas and novel findings alone are not enough. In short, the motto is “products, not papers.” Discoveries no longer need to be “taken out of the lab” as the team must also take ownership of application development. Research startups adapt and expand their capabilities to include product expertise and business functions by bringing in new talent into the team rather than spinning out commercial ventures. The operational model is iterative and the team must generate knowledge and apply them in a self-reinforcing cycle. This new approach provides a number of benefits:

1. The mission aligns team members on desirable outcomes and acts as a guiding star for assessing impact of research activities. The emphasis on product development reinforces a culture of practicality.
2. Focused mission-driven organizations promote cross-disciplinary research. Teams are free to mix research domains, however unconventional, to pursue their goals and achieve practical results. Major scientific breakthroughs are often unlocked by combining perspectives from different research domains.
3. Teams that produce new knowledge also become the primary consumers of their research outputs. In other words, they “dogfood” their own discoveries. As a result, teams spend resources efficiently to conduct reliable science and to self-correct quickly. Researchers are more exposed to the benefits of risk-taking as well as the harms of sloppy science.
4. By combining research and product development into one organization, drivers of high productivity such as tacit knowledge and team relationships are better retained.

The clarity of purpose and concentration of accountability in research startups will lead to faster product development and commercialization. Utility milestones, such as prototypes, demonstrations, and user adoption are naturally favoured in research startups. An emphasis on real capabilities raises the bar on expectations and researchers have more skin-in-the-game in the outcomes. New governance structures, like specialized research boards of directors, provide guidance and oversight to the research team regarding team performance.

Team Science

Research startups expand the types of people that are involved in research and promote healthy team dynamics between researchers. Each team has the freedom to recruit the necessary talent. A shared and focused mission aligns researchers on long-term outcomes. These incentives create better environments for collaboration, talent retention, and capital efficiency. For example:

1. Researchers with unconventional backgrounds and those lacking formal academic credentials will have more opportunities to prove their merit. The inclusion of “outsiders” is highly beneficial for the development of new approaches and creative framing of difficult problems.
2. Teams have the flexibility to concentrate talent by hiring the best candidates, pay market-rate salaries, and issue other types of compensation tied to long-term outcomes. Small elite teams move faster and work more effectively. A corporate structure creates opportunities for all team members to have a stake in the success of the team.
3. Researchers are continuously exposed to other domains, like engineering and business development, which helps them develop better research intuitions by expanding their awareness of the full range of problems their team faces in the pursuit of their mission.
4. Researchers whose work is found to benefit another research startup are free to collaborate. A healthy research startup ecosystem will have researchers cross-pollinate ideas, experience, and knowledge from one team to another.

The research startup model promotes cooperation within teams and takes advantage of talent from outside of traditional research institutions to create intellectual diversity. Shared reward mechanisms reduce toxic competition between researchers for zero-sum recognition. Value created by new ideas, state-of-the-art expertise, and a productive team culture improve group performance and benefit all team members.

Flexible Ventures

Research startups can build a wide range of products and services. Some teams have a clear product vision and one or more potential paths to a viable business model. Others build ambitious infrastructure projects and other types of scientific public goods that benefit the broader research community. Depending on the scope of capital investments, research startups can attract a combination of venture investors, corporate sponsors, philanthropic donors, and government agencies as financial supporters. Each research startup organization can be structured fit-for-purpose, for example:

1. Research-stage businesses pursue projects with a clear vision for marketable products and services. These teams are committed to work out the core scientific and technical problems as well as the necessary business operations for commercialization.
2. Scientific tooling teams work on scientific software and hardware projects that require sustained and focused engineering efforts to bring to life. These tools are often public goods used broadly by researchers and new organizational designs are needed to bring together the talent and resources required for development and maintenance.
3. Data teams work on collecting, curating, and maintaining high-quality scientific datasets. These assets are the core foundations for the development of data-intensive applications such as machine learning models and other types of artificial intelligence.
4. Infrastructure teams work on large-scale projects that often require investments in real estate and physical facilities. These teams are ideal candidates for international cooperation to sustain the high capital costs.
5. Specialist contractors are teams organized around specific technical expertise to assist other research teams. They are defined by their talent pool and do not typically hold intellectual property nor other types of assets.

Some teams may choose to use a structure that crosses multiple categories. Many will create partnerships with other teams via equity, intellectual property, and facilities-sharing agreements. The flexibility in mission and structure are important factors that allow teams to adapt to needs and the whole ecosystem to evolve.

Bold Science

Conceptually, research startups concentrate accountability for end outcomes and provide increased freedom and power in the means of pursuit. As a result, they are more risk-tolerant and more suitable for ambitious ideas compared to other research institutions. Researchers working in a startup model learn faster by taking bigger steps and correcting their approach more promptly. They worry less about the optics of their efforts to their peers and politics of publications. The team exhibit more intelligent risk-taking behaviours not just in science but throughout the entire organizations:

1. Negative but reliable scientific results are valued and treated as effective ways to guide future efforts. Results are judged holistically in relation to the pursuit of the mission.
2. Research startups innovate with all aspects of their operations. Everything from recruitment processes, internal management structure, and legal strategies can all be areas of experimentation. As with scientific experiments, results from organizational design experiments are also shared with other research startups.
3. Tailored governance for research startups provides another layer of risk-tolerance. Specialized boards of directors can provide irreplaceable human-judgement that is necessary to create the culture and space for risk-taking.

Similar to business startups, research startups do not require broad consensus from peers to form teams, acquire resources, and begin their work. They only need to find a small number of supporters who are willing to bet on the team’s abilities and ideas. The lower barrier of entry injects new dynamism into the research ecosystem, whereby even the strangest ideas are allowed to have a chance to prove themselves. The outliers and anomalies are often where the most impactful research work takes place.

Trust and Cooperation

In a vibrant ecosystem of research startups, teams solve difficult problems faced by other teams and find new ways to collaborate. A community of research startups is more powerful, innovative, and resilient than the sum of its members:

1. Teams can create partnerships via equity, intellectual property, and facilities-sharing agreements.
2. When a team encounters difficult scientific problems or technical barriers, new research startups can be created to address them. This is particularly exciting and impactful when the problems in question are bottlenecks for progress for multiple teams.
3. When new discoveries unlock unexpected uses beyond the research startup’s original mission, teams can be spun out to pursue new promising applications.
4. Teams can merge though acquisitions if their efforts are more effective in an unified organization. Talent migrations drive cross-pollination of ideas.
5. As with any risk-taking ecosystem, negative results and outcomes are a fact of life. Research startups that cannot continue to pursue their mission can be gracefully spun down and reabsorbed into the ecosystem. Lessons learned are shared via post-mortems and assets are transferred to new organizations. Team members can join other research startups to bring their expertise and talent towards another mission that they are passionate about. Turning negative outcomes from genuine effort into successes for others creates a culture of resilience and mutual support.

Research startups form dynamic and evolving communities that collaborate more effectively than traditional institutions. The common ethos of practicality, flexibility, and accountability at the core of the research startup model helps facilitate trust and partnerships between teams. Organizations like Anomaly Research act as a central hub that creates ecosystem-level programs, tools, and resources that research startups can tap into.