Jeffrey Fredrick and Douglas Squirrel

Douglas Squirrel opens the recording as if it were the Troubleshooting Agile podcast, then notes that the audience can see them today and that this is not the normal podcast audience. Squirrel introduces himself as a CTO in London, a consulting CTO who works with many organizations, and co-author of Agile Conversations.

Jeffrey Fredrick introduces himself as a longtime agile practitioner from before the word agile was coined, also based in London, currently managing director at a fintech company four days a week and consulting on the fifth day. He says that, together with Squirrel, he hosts the Troubleshooting Agile podcast.

Squirrel frames the setting as DevOps Enterprise London and says they are talking about dumping ceremonies, one of his favorite topics. Fredrick says readers of Agile Conversations will know they focus on conversations more than ceremonies, but this talk is not a repeat of the book. It is compatible with the book and aligned with the core message, but different.

Fredrick says the common audience is people trying to make a transformation: they are changing how they do things, making a good-hearted attempt, and yet frustrated because things are not working. The obvious answer seems to be buying the IT Revolution catalog, reading the books, and doing what they say. That is good, he says, especially Agile Conversations, but knowing what to do is not enough to get the results. They want to talk to people who are doing the work, buying the books, bringing in the practices, and still not getting the results.

Squirrel adds that even people who read their book could fall into the same traps. They will explain those traps, why ceremonies are not going to work, and why the audience should do something else.

Fredrick says the idea that culture matters is not new. People in DevOps, Agile, Lean, and digital transformations all talk about the importance and difficulty of culture. Squirrel jokes that the easy advice is to change the culture and then everything will go fine, but people keep failing. Fredrick cites a statistic that half of UK transformations are delayed, and says something is going wrong: people either do not know how to create a good culture or do not know what they are aiming for.

Fredrick says many people have read Accelerate and encountered Ron Westrum's three cultures, which describe information flow and what people in power care about. Those concerns produce recognizable cultural experiences. Squirrel jokes that, because the first culture is listed first, the audience must want to be pathological. Fredrick says probably not: generative sounds better by name.

Squirrel says pathological cultures can still be attractive because they are keenly interested in results and directed from the top, and many successful companies run that way. Bureaucratic organizations dominate, and people are used to following rules. He does not think people set out to be pathological or bureaucratic, but they wind up there because that is the easy thing to do. They also wind up with ceremonies that do not work in a bureaucratic culture because that is what the books say to do.

Fredrick says the tricky part is how to get to generative. The way people approach transformation reflects the culture they start with. A focus on ceremonies can easily become a bureaucratic exercise: people say they are doing it because it is in the book and that is where their focus is.

Squirrel says it seems like doing what the books say should work, because the books contain knowledge. Fredrick agrees that advice such as small teams and daily stand-ups has real advantages and value. But ceremonies only have value if people approach them and live the right way. Ceremonies are effective only when the people in the culture buy into them; ceremonies that work are authentic expressions of the culture.

Squirrel says wanting to be generative is not a case where one plus two equals three. A top-down transformation push plus middle managers saying they will do it by the book does not get a generative transformation. It gets empty ceremonies, which is what they are telling the audience to dump.

Squirrel describes the empty ceremony symptom: in stand-up, planning, or retrospective, everyone stares at their shoes, or at Zoom while looking like they are staring at their shoes, and says, "This is what I did. This is what I'm doing. No blockers." If that is the feeling, they are the audience for this talk, because there is something they can do about it. Most books, training, consultants, and others say to find a way to improve the culture and get to generative, but do not explain how. They will give step-by-step things the audience can do and want to hear whether they work.

Fredrick says effective collaboration is something people already understand. He introduces a two-by-two diagram about how collaboration goes wrong. Squirrel says he knows they want the green upper right. Fredrick says the surprising thing is that the red bad zone is not the bottom left. The diagram describes moving from a traditional top-down culture with low disagreement and low trust/respect toward true collaboration.

Fredrick explains two major failure modes. If people are willing to disagree but do not have trust and respect, they get unproductive conflict. If they have trust and respect but are unwilling to disagree because they are too worried about being nice, they get groupthink. Real collaboration requires people who trust each other enough to care what others think and trust each other enough to say what they really think. Conflict between ideas is required.

Fredrick says trust and respect are the number-one enablement on both sides. Without trust to share what you are thinking and trust to hear what others are saying, collaboration is not possible. Squirrel challenges this as sounding like platitudes, saying he does not believe trust and respect can be created except by already having people who trust and respect each other.

Fredrick answers that there is something you can do: start with conversational transformation. They did not invent it; they came to it over eight years as they learned it themselves. It goes back to Chris Argyris, the Harvard professor, and a body of work going back decades. The consistent message is that if you change your conversations, you can change the way people relate to one another, and that way of relating is your culture.

Squirrel says everyone knows what the end results should look like: not merely being in the upper-right green quadrant, but what good collaboration looks like. They ask the audience to get hands on keyboards in Slack, joking that Squirrel will use the power of his mind to predict a week in advance what people will type.

Fredrick asks the audience to imagine making a group decision about where the next DevOps Enterprise Summit should be. The question is not which city to choose, but how the group should decide. Each audience member should imagine being responsible for deciding the process and type recommendations into Slack.

Squirrel predicts the answers: be diverse, get lots of different ideas, share those ideas including your own, let those ideas conflict, discuss pros and cons, and reach a decision everyone is happy with. Fredrick says this prediction is not hard because they have asked many groups in many settings, though never quite by predicting the future in a recorded talk. People always say some variation of asking everyone for their views and sharing their own.

Fredrick says everyone understands that getting all the information in the room is what they want: more information gives more choices and helps make better decisions. Squirrel jokes that now everyone knows what to do, so the talk can end.

Fredrick says the problem is the gap between espoused behavior and actual behavior. People believe they make decisions this collaborative way, except when the stakes are high. When something important is at stake and they might feel loss if it goes wrong, they behave differently. Squirrel models this by saying that in this special case he already has the answer and just needs help convincing everyone else.

Fredrick says that when something is important, people switch from collaboration to wanting their own ideas to win. They see differences not as strength but as threat. As a result, they are not curious about what others think because, if they are right, why be curious? They also do not share everything they know because it seems obvious. Even though they espouse curiosity and transparency, in the actual decision they try to win, do not share what they know, and are not curious, undermining trust and respect.

Fredrick says building trust and respect is easy but difficult: easy because it is simple to understand as doing what people already know they should do, but difficult because actually doing it is hard. Squirrel compares this to weight loss being easy because all you have to do is not eat.

Squirrel says they will teach steps the audience can follow in the session to begin changing behavior by the end. The necessary equipment is a piece of paper folded in half and a pen.

Fredrick introduces the Four Rs: a general set of steps for improving conversations. In this talk they will use them specifically to improve transparency and curiosity; in the book they use other tools and dimensions, but those still come back to the Four Rs. The technique is simple enough to learn in the remaining time.

Squirrel tells the audience to fold the paper vertically to make two columns. Fredrick lists the steps: Record, Reflect, Revise, repeat the reflection after revising, then role-play, and role reversal. Squirrel jokes that there are six Four Rs and they are helping with conversations, not math.

Fredrick begins with recording. The audience should think of a conversation where they were frustrated and identify the key exchange, not the whole conversation. Squirrel says just a couple of lines is fine. On the right-hand side, they write what was spoken: what they said, what the other person said, the things that would show up on a video camera. Squirrel asks what to do if he cannot remember every word. Fredrick says exact memory is not required; a naturally paraphrased, faulty, incomplete memory is enough to work with. Squirrel calls this part of their no-excuses policy: there is no reason not to begin working on conversations today.

Fredrick says the right-hand column is the visible world; the left-hand column is the invisible world: your inner dialogue, what you thought and felt while the conversation happened. Squirrel asks whether, if he has developed telepathy, he may write what the other person thinks. Fredrick says yes only if he really has telepathy; otherwise, write only your own thoughts and feelings. The right-hand side alternates their dialogue and your dialogue, but the left-hand side is all about you: your thoughts and feelings as you react and while you talk.

Fredrick says the record must be written down. Squirrel asks why it cannot all stay in his head. He explains self-distancing, distinct from social distancing: when something is on paper, it feels almost like someone else's work. Because the brain is conditioned to focus on how other people react and what they think, seeing yourself as another person allows better analysis. He tells a story about a teacher named Benjamin who recorded conversations, played them back, and shouted, "Benjamin, stop doing that," because the recording helped him see himself from a distance. The audience does not need to shout at themselves, but they do need paper and pen.

Fredrick moves to the second R, Reflect, the scoring section. The basic question is whether you are being transparent and curious. To assess curiosity, Squirrel says you cannot achieve curiosity without asking questions, so the easiest scoring method is to circle the question marks. In the Norbert example on the slide, Norbert asked one question.

Fredrick says the next question is whether the question was genuine. Squirrel explains that a non-genuine question is a leading question, such as, "We should speed up because we're getting close to the end of the talk, right?" It is really an opinion with a question mark at the end. In the slide example, Norbert advocates KVM and says, "Why don't we ask about KVM?" but means, "Let's use KVM." Fredrick says a genuinely curious version would acknowledge that KVM sounds interesting and there are other alternatives, then ask how to explore them. Squirrel says you can only tell whether a question is genuine for yourself because you need to know the actual thought and feeling behind it.

Squirrel says Norbert's first reflection is that he was not really curious. Fredrick then asks about transparency: look at the left-hand side, the things you thought and felt, and ask whether there are things there you are not sharing. Probably there are. The challenge is how to get those thoughts onto the right-hand side in a productive way. It might help Norbert to say, "What you're saying doesn't match what you're doing, Quinn," but saying, "What a hypocrite you are," will not produce a positive reaction. The useful idea is to express the gap between what you hear and what you see productively rather than unproductively.

Squirrel says repeated cases reveal triggers: items that tend to cause difficulty in living up to espoused behavior standards. You can plan for the next time those triggers happen and decide how to behave more constructively. Fredrick says that anticipates the third R, Revise.

Fredrick says Revise means coming up with another set of actions: things he could have said that would be more similar to the behavior he wants to create. In the Norbert example, Norbert felt he had not been curious or transparent and had been triggered. Squirrel and Fredrick then role-play the revised dialogue. Squirrel as Norbert says that almost everybody already knows KVM, he can check to be sure, and asks whether that is a good next step. Fredrick as Quinn says it is good to get the information but not to let the team think the choice has been made; he cannot leave business-critical decisions like this up to the team. Squirrel as Norbert says that does not sit well because he thinks they need more autonomy, not less, and asks to talk more about how they make decisions.

Squirrel says this dialogue is not perfect, but it has much more challenge and productive conflict than the first version, where Norbert more or less knuckled under to Quinn. Norbert asks questions that get more information and might let him change the situation. Quinn might still be unwilling to move, but Norbert will be surer about that instead of being in the dark and assuming he already knows Quinn's thinking.

Fredrick says they have just role-played a conversation that was not theirs. After someone creates a revised dialogue and scores it again, that repeat leads to a revision they like; then it is useful to have a dialogue with someone else. Something can look good on the page but feel wrong when spoken aloud: it may not sound like how you actually talk. Role-play helps make the words workable. Role reversal can also help because hearing someone else say the words back gives more self-distancing and lets you reason differently.

Fredrick says those are the Four Rs. If people start analyzing conversations and producing different behavior, there are predictable, demonstrable changes: improved relationships, trust, and respect. Improved trust and respect are the foundation for better conversations and collaboration, leading to better decisions and ultimately better leadership. This is simple but not necessarily easy.

Squirrel emphasizes that the process is something the audience does. The challenge is to take the paper with the revised conversation, try the techniques, and repeat the process. If they make those changes, they should see others behave differently toward them. It is not a process someone else does; it is a process you do, and the result is improvement. If they do not believe it, they can try it, because they have just done it during the talk and have an opportunity for a different conversation next time.

Fredrick says every day offers chances to improve because people have multiple collaborations daily: meetings, one-on-ones, email, Slack. Each is a learning opportunity to ask whether you are being transparent, curious, and behaving the way you would expect people to behave when collaborating and making good decisions. If you do conversation analysis, you will learn more, improve your own performance and behavior, and change the behavior of the whole group.

Squirrel points to more material in the slides: the five conversations from the book that they help address: trust, fear, why, commitment, and accountability. These problems are usually not hard to diagnose but are hard to know what to do about. They give steps like the ones just taught.

Squirrel closes with a challenge. If someone says there is too much fear, nobody understands why they are doing this, or nobody trusts them, first try the techniques described. He compares it to bugs: today someone would ask whether you have tried automated tests. He wants people to ask themselves and others whether they have tried conversational analysis, because it just might help when a team is afraid of its boss.