Jeremy Adamson: Minding the Machines: Building and Leading Analytics Data Science Teams
- 1. Minding the Machines Building and Leading Data Science & Analytics Teams DSCamp, Dec 2021
- 2. Jeremy Adamson Independent Consultant Jeremy Adamson is a leader in AI and analytics strategy, and has a broad range of experience in aviation, energy, financial services, and public administration. Jeremy has worked with several major organizations to help them establish a leadership position in data science and to unlock real business value using advanced analytics. Jeremy holds a Masters in Transportation Engineering and a Bachelors in Civil Engineering from the University of New Brunswick, as well as a Masters in Business Administration from the University of Calgary. Jeremy is a Professional Engineer in the province of Alberta. Learn more at www.rjeremyadamson.com
- 3. Minding the Machines John Wiley & Sons, July 2021 Traditional Analytics • Descriptive reporting • Collect, analyze, report • Ad-hoc analyses • Back-room function • Order takers and technical resources • Objective: Better understanding of what happened Value-driven Analytics • Process driven and parsimonious • Integrated with the business • Business Analysts • Analytical evangelism • Advisors and facilitators • Objective: Unlock value opportunities using advanced analytics and create a culture of data-driven decision making. The practice has evolved from its roots in decision support and business intelligence into a core component of corporate strategy, and an enabler for all areas of the business.
- 4. Failure to launch Organizations, regardless of size and technical maturity, have been overwhelmingly unsuccessful at realizing value from their investments in this space, despite significant investment. Half of organizations believe data science will fundamentally change their organization in the next 1-3 years (Ammanath, Jarvis & Hupfer, 2020) “85% of analytics projects end in failure” – Gartner, 2017 “8 out of 10 organizations engaged with AI say their projects have stalled” – TechRepublic, 2019 “Only 27% of executives we surveyed described their Big Data initiatives as successful” – Capgemini, 2014 “most organizations report failures in their AI projects, with a quarter of them reporting up to a 50% failure rate” – Fast Company, 2020
- 5. It is not just about the technology Process Portfolio Project Management Project Execution Strategy Value Proposition and Case for Change Organizational Design and Target Operating Model Closing the Gap People Team Conventions Relationship Management True Value Creation The practice has evolved from its roots in decision support and business intelligence into a core component of corporate strategy, and an enabler for all areas of the business.
- 6. Strategy – What do we do? How do we do it? Well-articulated strategies establish the “how” and the “what”, providing a framework for all decision making. Strategy Playbook • Current state assessment • Assess readiness • Capability mapping • Defining the future state • Principles, functions services • Organizational design • Target operating model • Closing the gap
- 7. Process – Cognitive Fluency Properly developed and understood processes give analytics teams the ability to scale and focus on core competencies. Process Playbook • Intake and prioritization • Scoping and requirements • Project planning • Project execution • Change management • Closeout and delivery Illustrative swimlane process diagram - MTM
- 8. People – Data Scientists are Artisans Skilled, passionate, and creative practitioners are the bedrock of an analytics function. Data science teams succeed or fail on the strength of their people. Artisans need… • to apprentice with an experienced technical leader • to have business leaders that they respect • to operate within a guild • to be stimulated in their work, and to have real agency • to feel respected and heard, and to be able to express themselves Data scientist hierarchy of needs - MTM
- 9. Tip #1 – Consider the Last Mile Effective Data Storytelling – Brent Dykes Assets and deliverables that cannot be operationalized are value destructive. Each project should begin with the end in mind.
- 10. Tip #2 – Hear Statistics, Feel Stories John Snow & CF Cheffins, 1854 The numbers do not have value – only the actions that they initiate are important. Building a compelling narrative is essential to success. John Graunt, 1662
- 11. Tip #3 – The Model is Not the Product Each project and activity must align to the organizational goals. Begin with the business outcome and work backwards. Picture from blog.militarytobusiness.com
- 12. Tip #4 – Pursue Parsimony Continually assess each project and process for signs of bloat. Be mindful of the multiplicative impact of well-intentioned decisions.
- 13. Tip #5 – Value Creation Every activity, task, and project, must be directed towards value creation for the organization. Each team convention, KPI, and communication must be designed under this rule. Shareholders Employees Customers Outcomes not targets Number of POCs Number of projects Number of collaborators Time saved Reports generated Quality of models Time to market
- 14. Tip #6 – Be Intentional About Relationships Data science and analytics is a tool for helping humans to make better and faster decisions. Adopt a service mindset to enable greater success. Picture from Health and Social Care Alliance Scotland Individual Lead Goal Terry F. (Finance) Jane P. Encourage Tableau use Judy W. (HR) Peter B. Peer review models Wanda N. (HR) Jane P. Encourage Tableau use Spencer P. (Tax) Peter B. Migrate ad hoc scripts Spencer P. (Tax) Jane P. Encourage Tableau use
- 15. Takeaways Having a successful data science and analytics function will be essential to organizational success in the future Analytics must be aligned with corporate strategy, and all activities in support of overarching organizational goals. Establishing consistent and parsimonious processes encourages adoption. Data science is a creative practice and end-to-end ownership is critical to embedding value into deliverables. Technical capability is secondary to human factors.