Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles Q&A with Claudio Perrone on PopcornFlow / Evolve and Disrupt

Q&A with Claudio Perrone on PopcornFlow / Evolve and Disrupt

At the Agile Eastern Europe 2015 conference Claudio Perrone gave a keynote titled "Evolve and Disrupt". InfoQ interviewed Perrone about continuous evolution, servant leadership, popcorn flow (an approach to continuous evolution through rapid experimentation), and doing experiments to make change more continuous.

InfoQ: Your talk at AgileEE was called "evolve and disrupt". What made you pick this title!?

Perrone: In our lean & agile community we often talk about learning organizations and the need for continuous improvement. But is the improvement we see really continuous? I'd argue that agile is about rapid evolution, not mindless conformance to branded ceremonies.

The core of my talk was based on this core premise: "what if we could continuously evolve the way we work as fast as some of the most adaptive microorganisms on earth?"

To use Kent Beck's beautiful XP metaphor, "I'm turning all the knobs to ten".

I also realized that we may be happy to be operationally better than our competitors, but to disrupt the marketplace with truly innovative products and services, we also need to evolve our models for understanding customer needs. Which is why I also introduced my flavor of "job stories" as a practical approach to Clay Christensen's Jobs-To-Be-Done theory.

InfoQ: Can you give some examples of how continuous evolution can look?

Perrone: I address problems by exploring options through a continuous stream of small change experiments. It may sound obvious, but most people don’t quite grasp how fast and frequently this can happen. Just a couple of weeks ago I had a conversation with a CEO who saw the outcome of this approach on some of his agile teams. I explained: “For me, continuous evolution is a way of life. Even at personal level, every single week I’m about five-experiments older. At the end of your life, how old do you want to be? Two-experiments old or twenty thousands? Imagine such a continuous stream of experiments throughout your organization, how far would you go?”

Some microorganisms can mutate every fifteen minutes. I often ask myself: “Can I sustain my current pace? Can I go even faster? Why only five experiments a week? Why not five a day? Can I help others do more?”

InfoQ: You stated that, for lean to work, managers should be servant leaders. Can you elaborate why?

Perrone: It's a matter of options:

  1. Autocratic leaders are expected to be in control, take all decisions. At best, they obtain compliance, not commitment. People burn out or take their paychecks and switch their minds off. 
  2. Then you have "leaders" who simply let you do whatever you want. They'll pat you on the shoulder and tell you they are like a friend, sister, or father. I have a family, thank you. They often use banners and slogans, all believing that problems will simply go away.
  3. Then, there are servant leaders, who solve problems and develop the critical thinking skills of people. They look at the underlying systems and help people figure things out. They tap into intrinsic motivators such as autonomy, mastery, and purpose. They have method.

You see? You always have options. I just gave you three. If you were a manager reading this, who would you rather be? It is never too late to take a small step towards change.

InfoQ: What do you think of host leadership, as described in I'm not a servant - I'm a host. To me it looks like something that is taking servant leadership one step further?

Perrone: It’s a rich metaphor. I like it. Is it really one step further than the concept of servant leadership, however? Let’s see. This week I’ve been working with teams who have one huge bottleneck along their value stream due to slow testing environments deployed in another country, company policies, and organizational silos.

In my Lean world, I put product development teams on top. They are creating value for the customers who, ultimately, pay everybody’s wages. Who helps them do the work? In many organizations, you’d have roles such as team leads or scrum masters. At team level, this is all well and fine. But a single team has limited vision and scope. Who helps these teams along the whole value stream? It has to be management. Right now, rather than power accumulators, I need managers who can move mountains across countries, realize that policies are created by humans at some point in time and that “by humans shall be destroyed” if they don’t serve their purpose anymore. That’s what I mean by servant leadership.

InfoQ: Earlier you did an interview with InfoQ about Lean Thinking Applied for Organizational Change where you talked about popcorn flow. Can you summarize it?

Perrone: Popcorn flow is an approach to continuous evolution through rapid experimentation.

As I explained in the previous interview, the word POPCORN stands for Problems & observations, Options, Possible experiments, Committed, Review, Next.

Hopefully the following illustration will make the mechanics a bit more vivid!

In lean, we often talk about value streams. Yet, it's not what we do, but rather what we learn by doing it that matters. When I look at a typical scrum or kanban board, however, all I see is a snapshot of the outcome of the thinking behind it. Perhaps we are missing an opportunity. Popcorn flow accelerates, sustains and brings to the surface the reasoning (how and what we learn), specifically through a continuous stream of small and traceable change experiments.

This is a vivid example of what I call a "learning stream". Value streams and learning streams work together and help us make progress like rails on a ladder. The trick is to make both visible. Most teams use two separate boards. But some teams who adopted this approach now split their single visual board horizontally. The top part is a kanban or scrum board, the bottom part is a popcorn board, with each step represented as a column.

Improvement without change is impossible. But change is hard. So, a popcorn flow guiding principle states: "If change is hard, make it continuous".

Rather than dealing with BIG change, we change all the time, through really small experiments. It's like rapid problem solving on complex domains. Rather than full-blown analysis, we are happy to use imperfect information to generate options, negotiate experiments and eventually converge to something that works.

InfoQ: I fully agree with you that improvements should be visible for team, it's something that I also emphasize for actions coming out of retrospectives. Can you name some benefits of having a popcorn board next to the team's Scrum or Kanban board?

Perrone: Typically, team members capture problems as they occur. Their increased visibility enables more structured reasoning. Experiments are negotiated, designed, and acted upon. You can’t forget them. I often bring managers to visit these boards to observe their team’s evolution. In companies where a firefighting attitude is rewarded and sustained, this simple act can have a profound impact in moving towards learning organizations.

Additionally, all experiments are traced, so we can track their throughput.

I'm not alone in thinking that popcorn flow experiment throughput and cycle-time are a much more solid leading indicators of real agility than, say, a consultant's opinion, cross-team velocity comparisons, conformance to certain agile practices, transformation net promoter scores, and other rather silly stuff we see nowadays.

InfoQ: How do retrospectives fit into the popcorn flow?

Perrone: Like a glove. Ah ah. No, seriously. Forget those deja-vu retrospectives that lead to no action.

Problems and options are typically captured and discussed throughout the week.

At a retrospective we review the committed experiments, discuss or introduce new options, and create new experiments.

The review questions look like this:

  • What experiments did we agree to do?
  • Which ones did we actually do?
  • What did we expect to happen?
  • What actually happened?
  • What did we learn?
  • Based on what we learned, what are we going to do next?

I have direct experience with weekly retrospectives, but other teams are experimenting to do all this work just-in-time, right after their stand-up meeting.

The key is to keep the review cadence really rapid and create a sense of kaizen urgency.

InfoQ: Can you give some examples of the experiments that you have done using popcorn flow?

Perrone: Sure! Here are some examples:

Experiment: Let's add explicit WIP limits on the board

Reason: Towards the end of the sprint, many stories accumulate in Coding Done (partially done work).


  • Work will be smoother and in a much better state around Thu/Fri
  • Reviews/JIT demos happen sooner rather than later

Experiment: "Fix as you go": If found small bugs (less than 20mins), just create a branch and fix them. Do a pull request and mark the id on the card.

Reason: too much bureaucracy for small bugs.


  • developer happy to fix things as needed without lengthy triages.
  • steadily improving quality.
  • low bureaucracy, but still able to track it if things go wrong.
  • at least 3 bugs fixed like this by due date.

Experiment: Do an Analytics meet-up to show how analytics work in <new kanban tool>

Reason: <product owner> needs some form of predictability.


  • PO/Team is aware of what is possible now with the current level of analytics
  • We have better understanding of if, how, when we can improve forecasting with minimum amount of estimation.

InfoQ: How did those experiments work out, what did you and the teams you worked with learn?

Perrone: The "wip" experiment is typical of a team at the beginning of a kanban journey. The followup was to further tweak the limits, but overall it was easy. Everybody shared the pain, understood the reasons and participated in devising options.

The "fix as you go" was a more difficult negotiation, as it actually confronted two schools of thought within a newly formed team working on a relatively old codebase: one more conservative and the second more pragmatic. Trying it for a limited time was a small step towards a more aggressive attitude that ultimately led them, months later, to continuous deployment.

The "analytics" experiment, was actually a stepping stone towards dropping point estimation and moving towards a mature #NoEstimate approach.

I think the more important thing to consider is the cumulative effect of all these changes. Some of the experiments have obviously failed to meet our expectations. Like a mutation, some change is positive, some did not bring the effects we hoped for. But when teams are autonomously able to generate at least 5-10 experiments/week as the ones I'm talking about, I'm pretty damn sure that Terminator's Skynet is self-aware and my job is done!

InfoQ: If people are interested to take steps to make improvement more continuous, can you give them some advice?

Perrone: You can start popcornflow as a little experiment. What would you expect to happen? You may consider structuring your next couple of retrospectives by following the popcornflow steps, without even mentioning popcornflow once.

Are you afraid of experiments and failure? One of our growing popcornflow community members suggested using the term “exploration” instead of “experimentation”. Christopher Columbus discovered the Americas after failing to meet his expectations.

We complain about attitude to change that others have. But often the problem is us.

We are limited by our mental cages, what we read, the people we meet, the brands we buy. Continuous evolution is a mindset.

I often remind to myself and others that “There is a world out there that needs to be changed. But it’s a world that needs you… to change.”

About the Interviewee

Claudio Perrone helps companies evolve fast and disrupt the marketplace. In his career, he has been playing key roles in Lean & Agile transformations for global organizations and fast-growing technology startups. He is currently writing a book about PopcornFlow, a new model for continuous evolution through ultra-rapid experimentation. He is also the author of A3 Thinker, a set of brainstorming cards and mobile solutions based on Toyota’s Lean management approach. You can follow Claudio on Twitter (@agilesensei) or learn more about Claudio, A3 Thinker and PopcornFlow on his website. Join the small but growing Popcornflow community! Finally, don’t forget to join the book notification list!

Rate this Article