Introduction
The true value of data products lies in their ability to empower all types of analytics developers, including business professionals, to create, iterate, pivot, and retire end-use-case-focused analytics that respond "at the clock speed of the business" (as Mark Stouse aptly puts it). Surprisingly, many organizations still struggle with allowing business users to create business-level metrics directly. This blog explores the concept of Citizen Data Scientists (CDS) and the importance of enabling a culture of data-driven innovation within organizations.
The Citizen Data Scientist Paradigm
The concept of Citizen Data Scientists often gets conflated with AutoML, but AutoML is just one aspect of enabling a CDS culture. A more holistic view positions CDS as business users and customers for internal IT and Data and Analytics organizations that create specific data products and tools for them. Here's how we can nurture this culture effectively:
- Continuous Education: Emphasize data literacy, AI literacy, decision literacy, and value literacy. Bill Schmarzo highlights the importance of educating business users to ensure they understand the potential and limitations of data and analytics.
- Use-Case Templates: Provide example use-cases and templates for solving problems with AI. This not only helps in reusing AI resources across similar problems but also accelerates the learning curve for business users.
- Expert Mentoring: Frame AI problem-solving as a process that requires expert supervision for good problem framing and data quality. Mentoring ensures that business users receive the guidance they need throughout the use-case lifecycle.
- Verification Before Deployment: Implement strict regulatory supervision to avoid compliance and risk issues. A dedicated 'model hardening' process can ensure that models are robust and ready for deployment.
- External Inspiration: Look for problems solved elsewhere to gain new perspectives and avoid pitfalls like 'blind automation fallacy' or the 'Faster Caterpillar Fallacy.' Learning from others can inspire innovative solutions and prevent common mistakes.
Enabling Analytics Users
One major challenge that data product approaches address is enabling analytics users (both business and data science ends) to focus more on actual analytics and model development rather than laborious tasks like data wrangling. Here's how:
- Controlled Experimentation: Foster a culture of controlled experimentation with the ability to test and quickly pivot using the OODA loop. This approach encourages innovation and rapid problem-solving.
- User-Friendly UX and Tooling: Develop user-friendly interfaces and data product-based tooling to speed up innovation. This allows business users to create simple analytics quickly and iteratively while ensuring transparency and compliance.
For the more complex 20% of analytics, providing agility for data scientists and engineers to develop sophisticated models is crucial. However, by 2022, we should be moving away from time-consuming tasks that occupy 60% or more of the process, such as data wrangling. AutoML can assist here, but it is just one part of the toolkit.
Industrial Revolution of Analytics
Just as the industrial revolution liberated people from grinding grain or generating electricity, we are now at a stage where analytics developers can assemble simple analytics building blocks rather than fabricating everything from scratch. This shift allows for greater efficiency and innovation in analytics development.
To achieve this, organizations need to invest in:
- A Culture of Controlled Experimentation: Encourage a mindset of testing, learning, and quickly pivoting to adapt to changing business needs.
- Robust UX and Data Product-Based Tooling: Create tools and interfaces that enable rapid development and deployment of analytics solutions.
Moving Beyond Citizen Data Scientists
The term "Citizen Data Scientist" often evokes mixed feelings. A more appropriate term might be "Citizen of Data Science," highlighting the role of individuals in guiding and coaching data teams without necessarily turning everyone into a data scientist. This new role would involve continuous education in data literacy, AI literacy, decision literacy, value literacy, and ethics, ensuring relevant, meaningful, transparent, and ethical delivery of business outcomes.
Alternatively, a term like "Data Value Architect" could emphasize the role's focus on value creation and solution architecture based on data and analytics resources.
Conclusion
The value of data products and the concept of Citizen Data Scientists lies in their potential to democratize data and analytics within organizations. By fostering a culture of controlled experimentation, investing in user-friendly tools, and redefining roles to focus on value creation, organizations can unlock the true potential of their data assets.
Call to Action
Are you ready to revolutionize your organization's data strategy and empower your team with the tools and culture needed to innovate at the clock speed of your business? Let's transform your analytics capabilities and create a data-driven culture that drives value and fosters continuous improvement. Contact us today to learn more about how we can help you achieve these goals.