Drunk Agile

Dan and Prateek take a humorous dive into the nuances of Little's Law, shedding light on its mathematical foundations and real-world applications. They tackle its limitations, highlighting the pitfalls of relying on averages in unpredictable environments. The discussion shifts to managing unvalued tasks in project management and the notion of flow debt generated by expedites. Finally, they explore essential assumptions behind Little's Law, tying it all together with a light-hearted approach to resource management during the pandemic.

Dan, an expert in flow metrics, and Prateek, a knowledgeable figure in agile practices, dive into the basics of flow metrics like WIP, Cycle Time, and Throughput. They humorously discuss how Work in Progress impacts productivity and the nuances between cycle and lead time. The conversation emphasizes the significance of tracking throughput and work item age for improving project efficiency. Their insights reveal how actionable metrics can enhance workflow predictability and accelerate process improvements.

Seth Stephens-Davidowitz, author of 'Everybody Lies', dives into the nuances of customer behavior and data. He debunks the myths around customer feedback, arguing that surveys can mislead product development. Through examples like Coca-Cola's 'New Coke' and Facebook's newsfeed, Seth highlights the gap between what customers say they want versus their actual usage. He emphasizes the importance of rapid testing and treating features as experiments to discover true value, drawing parallels between Agile practices and poker strategies.

Joining the conversation are Mike Ditka, the iconic Bears coach, and NFL legend Tom Brady, known for his unparalleled Super Bowl success. They discuss the misjudgments in valuing talent, using the 1999 NFL Draft as a cautionary tale. The pitfalls of prioritization methods like Cost of Delay are explored, alongside the power of Monte Carlo simulations in decision-making. The conversation emphasizes adapting project management strategies, urging a flexible, data-driven approach for better outcomes. Don't underestimate the lessons from the gridiron!

The hosts discuss the concept of cost of delay and its role in maximizing value. They explore how cost of delay is used in daily work and the challenges of calculating return on investment. The relationship between weighted shortest job first (WISJF) and cost of delay (CD3) is examined. The limitations and assumptions of cost of delay calculations are discussed, emphasizing the importance of a rigorous definition. They also mention the next episode on modeling for uncertainty and mitigation strategies.

An exploration of Wilt Chamberlain's 100-point game and his decision making. Discussion on irrational behavior despite having accurate data. Parallel drawn to a basketball player's decision to switch shooting style against evidence. The relationship between behavioral economics and agile methodologies explored. Pitfalls of mitigating risk through estimation and planning discussed.

Dan and Prateek discuss the concept of right sizing in Agile development, emphasizing the need to find the optimal size for user stories, features, and initiatives. They explore how cycle time and customer value are related, and suggest using data and probabilistic understanding for correct sizing. The hosts also highlight the importance of breaking down large features and stories to maintain predictability and maximize productivity. Overall, right sizing is crucial in Agile development to prevent problems and enable valuable feedback.

Colleen Johnson, CEO of ProKanban and ScatterSpoke, joins the hosts to discuss retrospectives, data, ProKanban, and creating an inclusive community. They delve into the journey of creating ScatterSpoke, the importance of user feedback, and the role of feelings in decision-making. They also explore starting small and getting feedback in product development, balancing guardrails and freedom in agile transformations, and the importance of data and simplified metrics in retrospectives.

Hosts discuss the use of daily throughput and feature level metrics for forecasting. They analyze Monte Carlo simulations and debate the disadvantages of certain methods. They explore the caveats of random sampling and aggregating zero-dominated data. They also analyze the impact of granularity on team performance and highlight the challenges of forecasting and prioritizing features in agile development.

Dan and Prateek take on using Monte Carlo simulations to predict probability of features getting done. They talk about the overwhelming effect that Feature WIP (number of simultaneous features in progress) has on your chances to finish your highest priority features. Note - Some parts of this podcast use visual aids. For the full experience, please refer to the Drunk Agile youtube channel - www.youtube.com/channel/UC758reHaPAeEixmCjWIbsOA