Scaling with Experimentation

This week’s Cyb3rSyn Newsletter is another special mid-week edition with a guest post by Venkatesh Krishnamurthy (Venky) and also a summary of the third Cyb3rSyn Labs Podcast episode, my conversation with him.

Venky is a seasoned Agile Coach and Australia's first LeSS (Large-Scale Scrum) Certified Trainer. More importantly, he was present and participated during the early experiments that eventually shaped what we know today as LeSS. The experiments and the underlying principles were the key topics of the podcast.

The podcast explores Venky’s journey into systems thinking, early LeSS experiments and fostering adaptable organizations. Here is a quick summary…

Venky's professional journey began in computer science, with a background in C and C++ programming. A pivotal moment arrived when he met Craig Larman, co-inventor of Large-Scale Scrum (LeSS), who introduced him to agile, extreme programming, and Scrum. Inspired, Venky transitioned from programming to process consulting, marking a significant shift in his career.

"Sometimes in life, you find some good people who will inspire you, who will show you that there is something else, and you get hooked on to it, and then you change your career and life itself," Venky reflected. He expressed gratitude for mentors like Craig Larman and Harish Jose, emphasizing the importance of those who share knowledge without expecting anything in return.

The conversation delved into the early days of Less, around 2002-2003, when Venky worked in Bangalore. Craig Larman, then the chief scientist, aimed to implement agile methodologies across the company's software deliveries. However, Scrum, designed for single teams, posed a challenge for their multiple-team projects.

Driven by the question of how to scale Scrum, Larman and Venky embarked on a series of experiments. "Without experiments, how do you really know what works, what doesn't work?" Venky stated. These experiments, numbering around 500, explored various approaches over a period of ten years with Venky documenting the progress.

Venky emphasized the importance of creating a "proper container" or environment for experiments. This meant ensuring that experiments didn't disrupt customer deliveries, timelines, or quality. The experiments also applied concepts from theory of constraints, systems thinking, and queuing theory.

The experiments and their outcomes — what to try and what to avoid — were meticulously captured, eventually forming the basis of the book "Practices for Scaling Lean & Agile Development".

Systems thinking underpinned all of the experiments. Venky explained that traditional hierarchical organizations often focus on individual parts (departments, teams) rather than looking at the interactions as well. Leaders, focussed on specific KPIs, may not see the impact of their decisions on other parts of the organization.

After drawing an analogy to the old Indian tale of the elephant and the blind men, Venky also emphasized that every solution creates new problems, often separated in space and time, a concept leaders need to understand. I brought up Churchman’s note about how the author had taken a privileged position in being able to see the “whole elephant”.

Lean thinking and queuing theory were also central to the experiments. Venky explained that we can see almost everything is a queue - from the product backlog to the sprint backlog. Understanding queuing theory is crucial because the work entering these queues (epics, stories, features) varies in size and complexity. Without this understanding, teams struggle to commit to deliverables and often resort to outdated methods.

The experiments conducted over a decade culminated in the Less framework, designed to scale Scrum while retaining its core principles. Key aspects of Less:

Venky contrasted LeSS with frameworks like SAFe, which can involve numerous roles and significant upfront investment. LeSS advocates starting with minimal roles, experimenting, and adding roles only when justified by learning.

Here is the guest post from Venky which should serve as a good intro to the podcast.

Recently, I bought a barista coffee machine, and like any coffee lover, I was faced with a dilemma—there are tons of coffee beans available, but how do I know which one suits my palate and budget?

I couldn’t just buy every type of coffee bean— that would be too expensive and impractical. Instead, I took a systematic experimental approach:

Even though this was a simple, low-risk experiment, the same principle applies to organizations experimenting in a complex environment.

Here are a few key principles that apply to both small-scale and large-scale experiments:

As many know, LeSS, or Large-Scale Scrum, is built on more than 500 experiments conducted over ten years by Craig Larman and Bas Vodde. I had the opportunity to be in the midst of many of these experiments.

One of the key enablers was strong leadership support—leaders were not just approving experiments, they were actively learning from them. Every experiment was measured against the optimizing goals of value delivery and adaptability. Based on the outcomes, the experiments were either pivoted or continued.

In my recent podcast with Laksh, we go deeper into real LeSS experiments, including before-and-after snapshots of experiments and their outcomes.

We experimented with various ideas of scaling Scrum while keeping the same simplicity that Scrum brings, ensuring that adaptability was not sacrificed. These experiments incorporated lean thinking, systems thinking, queueing theory, and many more influences.

Interestingly, none of these experiments started with predefined principles in mind. Instead, they evolved and emerged naturally as we observed, learned, and iterated through experimentation.

Imagine that you are stuck with only one type of coffee bean for the rest of your life. There is no fun in it, is there? Experimentation allows us to discover, adapt, and continuously improve—whether it’s in coffee or in scaling Agile.

If you’re curious about how real-world Agile transformations happen through empirical experimentation, check out the full conversation.