This is a transcript of episode 217 of the Troubleshooting Agile podcast with Jeffrey Fredrick and Douglas Squirrel.

Dragan Stepanović joins us to discuss emotions as system signals, why intervening at the emotional level has less value, and why you should get back on the horse.

Show links:

Listen to the episode on SoundCloud or Apple Podcasts.


Listen to this section at 00:11

Squirrel: Welcome back to Troubleshooting Agile. Hi there, Jeffrey.

Jeffrey: Hi, Squirrel.

Squirrel: We have a guest today, Dragan Stepanović. Jeffrey, tell our listeners why Dragan is here today.

Jeffrey: Well not too long ago we had a episode on ‘mobbing for safety’ that was inspired by a tweet Dragan had put up. After that it was suggested to us that we would should invite him on and that sounded great to us. So, welcome and thank you for joining!

Dragan: Thank you for having me.

Jeffrey: So your tweet had talked about mobbing and psychological safety, what are the things that you are thinking and working on these days?

Dragan: On social networks, conversations these days related to code reviews and pairing and mobbing use the term ‘co-creation’—I would be interested to learn who came up with that term—so I’m a big proponent of co-creation patterns and working together in a synchronous manner. I think veterans have a really tremendously positive impact on the team’s productivity when it comes to flow, throughput, and quality, broadly how the team works as a unit, which includes psychological safety.

Jeffrey: In your Twitter bio, you have a phrase that I don’t think I’ve seen very frequently, which is ‘socio-emotional technical systems.’ ‘Socio-technical systems’ I see more frequently, but you really bring the emotion part into it. How did you end up talking about the emotional element with systems thinking?

Emotional Data is Valuable Feedback

Listen to this section at 02:39

Dragan: Over the years I picked up on a lot of the patterns in organizations you see with social behavior and structures, depending on the way of working and things related to the technical architecture, etc. But I also noticed the effects that these have on the emotional level, on the individual level, and that there’s lots that can be discovered by understanding the feedback we get on an emotional level as individuals. So that was the reason for including that. So not only social structure, but also, you know, on the emotional level and on the on the individual level. So the idea behind it is there is so much feedback that we get from the systems in which we are working on and working together in that I think if we try to tune in, we’re going to discover a lot of the submerged feedback that hasn’t been surfaced up until this point.

Squirrel: Well, let me try to channel some of our listeners right now. A lot of my clients, when I talk to them about emotional engagement at work they say, ‘I just want to be professional. What I’d like to do is make sure that I stay rational because that would be the best thing to do, wouldn’t it?’ I wonder how you react to that, I’m sure people say that to you.

Dragan: Yes, exactly. I think it boils down also to two systems thinking, quoting Deming comes in handy here, that famous quote that 90 or 95 percent of the behavior in the system is governed by the system itself. We tend to try to react based on the events that we see without understanding the patterns and balancing and reinforcing feedback loops that happen in the system. So we tend to end up with very wrong intervention points in the system, trying to intervene on the individual level and also on the other side trying to correct the system based on the individual behavior. I think we can do better than that. There’s this interesting thing in one of the papers I was reading on causal diagrams, which is a tool from systems thinkers. There’s this really important distinction between intervention points in the system and the monitoring points in the system. Intervention points in the system are the points where a small change in the system can affect great change—there’s high leverage—and the monitoring point of the system is where a small change can be detected very quickly: there’s high sensitivity. That tends to get flipped when people are trying to intervene in their systems, they try to intervene in the monitoring point instead of understanding the points of intervention or the leverage points. To give you an example, think of a shower. The point of intervention is the tap itself, determining if the water is cold, right? Where the water exits is the monitoring part. Imagine trying to intervene in this system in the shower head instead of the tap. That’s the kind of analogy I’ve got in my head and I try to get others to when I see these mistakes in the wild.

Squirrel: Tie that up to emotion for me because I’m not sure I quite followed. How do emotions help me to do either monitoring or intervening? I’m not sure I follow.

Dragan: Yes. The emotional part is a monitoring point. It’s the feedback we get from the system, right? The system bubbles up on this level as well. And this idea, ‘just be rational’ or ‘try to be professional’ or whatever is trying to intervene at this monitoring point on the emotional level of the individual in a sense, right? So going back to this Deming quote, this feedback is super valuable, and we should be very attuned to it in order to understand what is wrong with the system, so we can make a change in some other part of the system, in a structrual intervention point that is going to lead us to to desired results on the emotional level.

Squirrel: I think I might follow. I think you’re saying that we should pay close attention to the emotional input we get, but not necessarily take action on the individual level of the person having the emotion, but rather take action within the system. Did I hear that right?

Dragan: Exactly. So there’s this thing that we tend to say and hear a lot. ‘People matter,’ or ‘people are the most important part of the whole system,’ and I believe in that. But trying to intervene at this point in the system is not going to help us with our goals.

Squirrel: I get it. What an interesting point of view; I often find that my clients either say, ‘I don’t want to deal with the emotions at all. I don’t want to monitor the emotions, I just want to monitor the system,’ and that doesn’t tend to work out very well. Or if they do hear some emotions, they say, ‘Oh, that’s an HR issue. We need to make sure that person’s happy. We need to address something, give that person a pay rise.’ And those things don’t work out very well. You’re giving us a systems thinking way of analyzing that second intervention. Have I got that right?

Dragan: Yes, exactly.

Squirrel: Got it. Jeffrey, what do you think about that?

Jeffrey: I think it’s great. There’s been a lot of research on things like psychological safety and engagement—both individual attributes—which say you need high engagement and psychological safety to have good outcomes. In a systems thinking diagram as you draw it out, you’ll see that there is a causitive link from psychological safety to your outcomes. But the important insight here is that psychological safety has to be built somewhere else. You have to find other intervention points that will impact psychological safety. I feel like we’re kind of teasing people a bit because the natural question becomes, ‘what are some of the system intervention points to help impact psychological safety?’

What Familiarity Enables

Listen to this section at 09:52

Dragan: Yes, we talk about psychological safety a lot these days, and it’s definitely important, but it’s important to understand it as a monitoring point in the system. It’s kind of a snapshot of the system that you have. Trying to intervene at that point isn’t really going to help or is going to have a limited impact, but trying to understand the whole environment in which these things happen is very useful in order to understand where exactly we need to change the structure so we can have a bit more of a behavior that we would like to have, and bit less of a behavior that wouldn’t.

Jeffrey: That was behind the previous episode we did; you had a very simple diagram about the connection between mob programming and psychological safety. I think this is one of those co-creation patterns. Why would mobbing or other co-creation patterns be an intervention point that impact psychological safety? How does that happen?

Squirrel: And could you remind us what mobbing is? Because not all our listeners will have heard the previous episode or know what it is.

Dragan: Sure. I’ll try to use a quote from Woody Zuill about what mob programming is: it’s all the people that are needed are from the team working in this same environment at the same time and all on the same thing. So it can include two, three, or usually three plus people that have all of the needed cross-functional skills in order to deliver a given functionality or to provide and value to the customer. When it comes to co-creation as one of the intervention points in the system, it’s very interesting if we think about the definition of psychological safety that came from Timothy Clark, which is that it’s an environment of rewarded vulnerability. So if we don’t have a reward when we’re vulnerable there’s lack of psychological safety, but also if we’re not able to be vulnerable that also means that we’re not really be able to achieve the levels of psychological safety we would like to have. So I see it as a kind of dualism in that sense. Then it’s a question of how we can change the way we work in order for people to be able to be vulnerable in front of others. Working together is one of the really great ways to do that. With pair or mob programming, if someone is working with me all the time and is able to provide me continuous feedback, not only after each line of code change, but even after each letter that I type, it means that someone is going to be seeing all the mistakes that I make along the way, right? We need to build this trust between each other through acceptance of each other. That is kind of the prerequisite for building psychological safety. So I think working together together synchronously tends to be a very powerful leverage point in the system if we want to establish higher levels of psychological safety.

Jeffrey: What’s interesting for me in that description is, I think about the two people in the room during a mob session or pair programming, there’s the person who’s typing at the keyboard and then there’s the person watching. Our listeners may have viscerally experienced what it’s like to type in front of an audience, which is to say it generally feels bad. In my experience, I do a lot of demos and talks in my career, and I’m always very careful to make sure that I am not typing when I’m on stage. I will have things copy and pasted, or I have multiple links open, because it feels very uncomfortable to type in front of other people. To me that resonates with that idea of being vulnerable. Is that the kind of thing you have in mind there?

Dragan: Definitely. There’s an interesting thing that happens because when we get together and first try pairing, it feels awkward until we build this acceptance and trust. Once we get over this hump, it gets way smoother and more enjoyable. Lots of teams also don’t get the chance to get over this hump. I often hear that pair or mob programming is the one to be blamed, but actually pair and mob programming helped to elevate this feedback on the surface, right? So this feedback was always there, that people were not able to work together. But now we have mechanisms to expose it. And, you know, maybe we should try to do something about it instead of just dropping it.

Squirrel: Yeah, you know, ‘I have this horse and I’d really like to get places faster and I’ve tried being on the horse, but it keeps throwing me off, so it must be the horse’s fault. People can’t ride horses because the horse doesn’t like it. That’s just not going to work so we should give up. I tried it for an afternoon, but no more.’ That’s an example of a general principle, right? This is why we do small batches in our continuous integration, is that if the process of releasing is painful, you should do it more often. If the process of interacting with other people and typing in front of them and being vulnerable with them is painful, then you should do it more often, just like you should get on the horse more often if you fall off it a lot.

Jeffrey: There can also be discomfort on the other side, the people who are watching. When it’s the first time that people pair or mob, it can be not always natural for someone to speak up while someone else is typing, I think there’s an innate sort of deference. There’s another definition of psychological safety which I think speaks to that side which I heard from Amy Edmondson on Chris Clearfield podcast, she was saying that the term psychological safety had probably set the field back. She was regretting her choice of of words; Amy Edmondson being the person who brought the term to attention. She said that a good phrase is ‘felt permission for candor.’ I think that speaks very well on the speaker side of the pair. The people in the room who are commenting on what’s happening are experiencing psychological safety when they feel comfortable speaking up. So you have the two different definitions of psychological safety speaking to both sides of what’s happening in the room, the experience of the person at the keyboard being vulnerable and the people in the room feeling comfortable speaking what their view is. When both sides are comfortable doing that, you’ve now achieved psychological safety.

Squirrel: Yep. We’re running low on time Dragan, would you be willing to consider coming back again? Because because I want to bug you about that notion of batch size, and making sure that your work is something that you can iterate on frequently so that you can get to psychological safety faster. I know that’s another area you’re really interested in. Would you be willing to come back next week and talk to us more about that?

Dragan: Definitely. I would love to.

Squirrel: Fantastic. We’ll be back next week with Dragan and Jeffrey and me. See you then. Thanks, guys.

Jeffrey: Thanks, Squirrel.

Dragan: Thank you.