Welcome And Change Resistance

Peter [0:04]: Welcome to Definitely Maybe Agile, the podcast where Peter Maddison and Dave Sharrock discuss the complexities of adopting new ways of working at scale.

Dave [0:12]: Hello, Peter. How are you doing? Great to hear from you. I had a question for you, just to kick things off. I've been working on organizational change and why it's difficult, digital, AI. We've talked about many of these things, but I wondered if we could explore the foundational bit, that immune response we see in organizations that makes change difficult.

Peter [0:35]: It's an interesting one. You and I have been operating in this space for a very long time, and as we know, especially as organizations get larger, it's harder and harder to implement change. Because there are more and more people, more and more individual... well, if we're in the AI world now, non-deterministic nodes. The people in your organization.

Dave [0:56]: Warm-blooded, non-deterministic nodes.

Peter [0:58]: Warm-blooded, non-deterministic nodes, yes. People are actually very, very good at change. It's just en masse, people who have been subjected to very rigid systems, or who've observed certain behaviors, or who've been doing things a certain way for a long time, find it harder to change. They conform to the system they operate in. So even though individuals can change, changing the entire system, the entire organization, can be much harder.

Dave [1:33]: I'd actually add to that: the systems, the structures within which you're trying to achieve change, those organizational practices and support systems, are expressly designed to limit change.

Peter [1:49]: Yeah.

Dave [1:50]: They're there to stabilize. Kotter's eight steps, step seven or eight, is "institutionalize change." Institutionalizing change means putting structures in place so the change is stable and we don't revert back or drift. We stabilize and solidify a change into a new norm.

Culture And Structures That Stabilize

Peter [2:12]: Yeah, and we need that, because we want things to be consistent and repeatable. We want things to happen the same way, or at least a similar way, over and over again. So you need consistency, and that's where systems, rules, and all the rigid pieces come from.

Dave [2:29]: As you're describing it, Peter, it reminds me of conversations we've had over the last few months around context: what are the stable pieces we're working on, what's the purpose, what are the goals. These institutional practices often stabilize around certain ways of working, certain norms within the organization, what we often describe as culture. They get locked in by how we report to one another, whether an organization has a culture of one-on-ones or more visible steering meetings, coordination meetings around projects, how they manage work. What structures are in place will often determine how easy it is to make changes.

Peter [3:13]: Yes. I've seen this in more than a few organizations, where it's almost not allowed to talk to the person sitting across from you, because you're not allowed to talk to somebody unless there's a project code to charge that time to. That kind of behavior stifles the ability to change, because people can't easily interact.

Dave [3:41]: It's expressly designed to limit the fluctuations you might see. Skip-level meetings are another practice that starts breaking some of that down, when you see conversations pushed beyond the normal lines of reporting.

Peter [3:57]: Yeah, and we know organizational change is hard. We talk about things like the organizational immune system.

Incentives, Permafrost, And Constraints

Peter [4:05]: They also describe it as the permafrost layer, where it's very hard for a message to get through the way you expect it to across the organization, because you've got all these competing incentives. Everybody's got their own pieces they have to be concerned about and ensure get delivered. If your incentives are misaligned, it's very hard to change anything.

Dave [4:27]: That's one of the reasons I appreciate this conversation. What we'll often do is come into an organization wanting to affect change, bringing new ideas to the table, and we see these things as barriers. But if you reframe it and understand that these structures exist because they generate the behaviors that make the organization work the way it does today, it's less about removing a barrier. Because when you remove those barriers, you often end up somewhere you don't want to be, since you've taken away the constraints. All of a sudden things happen that you didn't expect. It's more about adjusting and rebuilding them to support where you actually want to get to.

Peter [5:15]: Yeah. Changing the incentives guides people to start moving in the direction you want, creating space and permission for people to experiment or try new things that move toward that change.

Dave [5:31]: So you're creating the ideal conditions for the change to thrive. And I think one of the things we want to bear in mind is that we shouldn't make those changes lightly.

Peter [5:44]: No, and we shouldn't minimize how difficult this is. It's not an easy thing to do, as shown by any number of statistics that will happily tell you how hard transformation is and how often it fails.

Dave [5:58]: Yeah. As we look at something like this, a couple of things really jump out.

Boeing And The Cost Of Finance

Dave [6:05]: One is that if we're not deliberate about these changes, we can end up somewhere we don't want to be. There's no shortage of stories about this. Anyone going through an MBA, or any executive education program, will have case studies like this pulled out and discussed at length. The one I'm seeing mentioned a lot lately is the Boeing–McDonnell Douglas merger, which shifted Boeing away from an engineering-first, quality-driven mindset, where engineers making decisions were elevated in the organization and had a lot of influence. When the merger happened, McDonnell Douglas leadership brought a much more shareholder-driven, finance-driven mindset to the table, focused on hitting schedules, sometimes at the cost of product and technical quality. That's contributed to the situation Boeing has found itself in.

Peter [7:21]: One of the common pieces I see, partly from my DevOps background, is that when finance is the one making those decisions, redundancy in the system gets seen as waste. So costs get cut where redundancy is actually needed, redundancy that might have stopped planes from dropping out of the sky. There are simpler examples too. Why do we need two network switches? Because we need to parallelize access into two different systems. If we only have one, it's possible for someone to wipe out data with no backup. There are good engineering reasons to do things a certain way, but that doesn't always translate well onto the accounting ledger.

Dave [8:19]: And the consequences matter, because a lot of organizations will suddenly go into contraction or stabilization mode, managing everything from a financial perspective. The problem is that can squash or significantly shift the ways of working within the organization. In Boeing's case, even though they want to shift back to an engineering-first mindset, it isn't a switch you can just turn on and off. Once you've eroded something like that, it's a bit like trust. These human interaction elements, once eroded, take a long time to rebuild. It takes persistent, consistent activity aligned to rebuilding whatever's been lost.

Peter [9:09]: Yeah, we were talking about this last time, about leadership behavior being a key part of it.

Leadership Behavior, Trust, And Honesty

Peter [9:20]: If you've got new leaders, different leaders behaving differently, it changes the culture of your organization. And changing leadership behavior can be very difficult.

Dave [9:28]: It's interesting, because that moves us back into the organizational honesty conversation we started with. At some point, the messages you get from leadership and the behaviors they actually show have to be pretty closely aligned for you to build real changes in culture, in the organization's immune response, so new practices can be adopted across the organization.

Peter [9:57]: That's especially true with AI, since it's a huge disruptor at every level. You'll see different levels of adoption in different places, and it's not enough for it to just come from the top: "I want AI." I think we've mostly moved past that stage. I hope so. But there's still a need for the people closest to the work, the ones who can actually apply this to real business problems, to figure out how to rethink the way things are done using this new technology. Most of the use cases I've seen have been: here's the system as we've always done it, with all its slow, painful aspects, and we're just going to smear a layer of AI over the top of it. Nothing really gets better.

AI Adoption And The Smear Layer

Dave [11:08]: I kind of blame the Pareto principle here. Whenever we're making some kind of change to an organization, a process improvement or whatever it might be, we tend to focus on the 20% that delivers 80% of the value. The problem is the remaining 20% is often the hardest, the toughest conversations to have. One organization I'm working with right now has a tendency not to move people who are in the wrong place. People get promoted into a position, but once they're there, the organization doesn't have a good way of tackling the tough conversation of "this isn't the right fit, how do we rethink it." So they end up with people stuck in the wrong roles. It's not that they're bad at their jobs, it's that the organization struggles to have that difficult conversation, so it gets avoided. When AI comes along and accelerates or catalyzes change, it tends to surface these areas we've been avoiding, and forces the difficult decisions.

Peter [12:24]: Yes. It amplifies the weak spots in the system and makes them very obvious, very quickly. Then it's a case of, well, now you have to have that conversation, and most organizations will just walk away from it.

Dave [12:39]: That's been their history, they've avoided them so far, so it becomes even harder to deal with. When the change is something like a new way of working, and we've been involved in Agile and DevOps for many years, that pace of change is pretty pedestrian. It's human-paced. It can take months, and it ends up affecting next year's performance conversations, not this year's. So it's easy to put off the tougher conversations. But nowadays changes happen in days and weeks. There's governance and infrastructure needed to get changes into production, sure, but there are also tough conversations along the way that don't have a mechanism to move at that pace either.

Human Pace Versus AI Pace

Peter [13:30]: Yeah, we still have to move at a human pace for the human parts of the system.

Dave [13:35]: Right. Those are often the conversations about how we manage, how we lead, how we make tough decisions under uncertainty, without just rolling the dice. We're warm-blooded, non-deterministic nodes in the system.

Peter [13:49]: So, with that in mind, how would you sum this up for our listeners?

Dave [13:54]: One thing, and I don't know if we've formally defined it, is this idea of the organizational immune system, and recognizing that, just like our own immune system, it's there for a purpose. It's a positive thing, not a negative one.

Peter [14:12]: Yes.

Dave [14:12]: We often bang into it thinking it's negative, but we have to recognize it's there for a reason, it serves a purpose. We don't want to get rid of it. We want to understand it, work with it, and make carefully thought-through changes so the immune response isn't attacking the parts we're trying to change for good reason, the beneficial changes we're trying to bring in. That would be the first thing.

Peter [14:41]: That makes me think of value stream mapping as a way of making the system visible, so you can identify where those changes should happen.

Work With The Immune System

Peter [14:50]: One thing I hear a lot is, "well, we don't have any value streams." You do, you just don't know what they look like.

Dave [14:54]: They're intertwined. It's like a ball of wool you have to untangle.

Peter [15:02]: Right, and figure out where the dependencies are, where the mess is that's slowing things down, where the conflicts are, and what needs to be addressed. That kind of exercise can be incredibly valuable. It's not the only way to do it, there are others, but it's a good way of looking at what parts of the system you need to focus on.

Dave [15:27]: I couldn't agree more. Just having some shared terminology, some agreed steps, lets us understand things a lot better. And building on that value stream mapping, another piece is recognizing behaviors that need to be modified or changed. In many cases, we don't want people experimenting too much with certain things. We want repeatability, not too much variation. So we put a lot of structure in place

Value Streams And Tuning Variation

Dave [16:05]: to limit variation. What we're seeing now, with the push toward more innovation and more change, is that we actually want to turn down some of those immune responses that limit innovation and limit variation. We want to tune down the thing we've been actively muting.

Peter [16:25]: The interesting part is, if you've been aiming for consistency in the system, and then you arm everybody in your organization with something that's incredibly variable, purely variable in fact...

Dave [16:38]: Yeah, and that generates an immune response, and behaviors where people are unclear, pausing, not knowing what the next steps look like.

Peter [16:51]: Because it isn't as simple as pulling up a ChatGPT prompt and typing something in. There's a lot more to it, in terms of thinking about how these systems need to interact to actually create value within the organization.

Dave [17:06]: And where the opportunities emerge as a result. I've found one or two really interesting stories around this.

IKEA Story, Takeaways, And Wrap

Dave [17:15]: IKEA is one of them. Their customer service department brought in AI to automate the majority of customer service inquiries, something like a 47% reduction in inquiries coming in. But what they found was that the remaining inquiries were mostly about interior design, very human-centered discussions. So they retrained their customer service agents to focus more on interior design and answering those tougher questions. Rumor has it, at least as I remember it, about a billion dollars on the bottom line as a result. It's that interesting mix of removing the standardized piece and then bringing in some variability that actually has value for the organization.

Peter [18:04]: Exactly. I think we managed more than three takeaways this time.

Dave [18:06]: Yeah, we covered quite a few, but there were some interesting stories in there.

Peter [18:12]: Well, with that, I think we should wrap up for this week. Thank you to all our listeners, as always. Don't forget to hit subscribe and tell your friends about us. You can reach out to us at feedback@definitelymaybeagile.com. Until next time.

Dave [18:28]: Peter, thanks again.

Peter [18:29]: You've been listening to Definitely Maybe Agile, the podcast where your hosts Peter Maddison and Dave Sharrock focus on the art and science of digital, Agile, and DevOps at scale.