Most legacy companies are organized around hierarchies that worked in the 1980s — but won’t necessarily be effective today.

MIT Sloan Management Review: We read some of your thinking on distributed leadership, and I was wondering if you could describe this concept and what it looks like at your organization.

Maile Carnegie: If you look at the way business is run in a modern 21st century company, whether it’s Google or Amazon.com or others, it is organized in an agile way. Decisions are made by the people who are closest to the customer. This means you get a lot more speed in delivery and speed to value. Employee engagement levels are also much higher in these agile 21st century companies.

Most of the traditional 20th century companies — such as Procter & Gamble, which I’ve worked for, and now ANZ — are still very much organized around functional, classical 1980s hierarchy. They tend to run excruciatingly slow.Employee engagement is on the decline.

Distributed leadership is about moving to agile. While there is the quite orthodox application of agile, I’m thinking more of agile being a significant change in leadership, skills, and capability. So, distributed leadership is fundamentally about shifting leadership, skills, and capability, and moving to agile — and that’s what we are doing at ANZ.

As you move toward this more distributed leadership, do you need to provide employees training to take on these more leadership roles? Are they eager to step up? What have you had to do to help make this distributed leadership happen?

Moving to distributed leadership requires a combination of training and new skills. Some of the habits and tools can be taught quite easily. Some new skills may also need to be brought in. In some cases, those skills can be trained. But in other cases, it’s going to require bringing new people in, for new skills that haven’t been called on before, like data science, software development, and the DevOps space.

Many traditional 20th century companies have outsourced their software development. However, newer companies wouldn’t outsource this skill. It would be near and dear to their hearts. They wouldn’t outsource it to save their lives. People in the organization may have a good background in software development, but they haven’t had their hands on the tools for a while. So, that’s a great opportunity to train them and brush them up. But in many cases, these skills would need to be brought in.

The other thing we’ve done to encourage distributed leadership is to fundamentally recast our definition of good leadership. To enable people to make really good decisions that are closer to your customers, you need to have much, much better leadership in the organization. We’ve recast what it means to be a great leader. In the past, a command-control leadership style may have been what was expected. But now the leadership style is quite different when you’re running an agile and distributed leadership organization.

What are those different skills in terms of distributed leadership you’re looking to cultivate, train, or find in people?

The two big skills we are looking for are technical mastery and leadership. For the technical mastery, we need more software engineers. We need people who have still got their hands on the tools, have a craft, and are excellent at their craft. Unfortunately, in most 20th century companies, success has been defined by actually no longer doing the work but managing others who do it. You then have people who were at one point a great marketer or a great data scientist or a great software engineer. They then got promoted, and the promotion meant they were no longer actually doing their craft. So, the first thing to focus on is to get people back at being excellent at their craft — their technical mastery. The second skill needed to be a great leader in today’s context is to be inherently curious. Today’s leaders need to lead through influence rather than through command and control. That’s quite hard for people who have really only had one quiver in their leadership bow, which is command control.

What does that training program look like for these leaders? Are you bringing in in-house training to help them develop more quivers or external programs?

It’s a combination of both. It starts with the recruiting process. Recently at ANZ, we asked people to apply for roles that interested them. We didn’t just slot people automatically into where we thought they would fit best. We then screened the leaders against the number one criteria of values and leadership. Not everybody had all the attributes we were looking for in our new agile world, but we believed that if they had the right cultural values and leadership attitude, they could develop the other skills relatively quickly.

To me, it’s essential to make sure you have people in those leadership roles who already have the “new” skills or whom you believe can rapidly acquire these new skills. After that, we followed a regular capability deployment method. Throughout this process, clear communication is essential as it’s a big change from what people are used to.

From a training standpoint, a lot of the training will be done internally, but we will bring in external staff and skills as needed.

What have the challenges been around building the capabilities you need in a 21st century world?

Many organizations have been guilty of training their management just to manage, rather than keeping their skills sharp in their craft.

Let’s just pick on marketing for a minute, where I started my career. When I grew up in marketing back in the early 1990s, I was actually doing the work of marketing. Then, typical of what happens in any 20th century company, I got promoted. And every time I was promoted, I got further and further away from the real work. One of the things that shocked me when I moved to Google was how little I actually understood about digital marketing because I had been so removed from it. Today, you can’t imagine anyone saying they’re a proficient marketer if they don’t know how to do digital marketing, yet I was leading teams of marketing people at P&G and I did not deeply understand one of the most important tools that a marketer has.

This can be quite confronting for people. It was now very transparent about the technical mastery that was required to genuinely do your job. That’s pretty scary if you haven’t had your hands on the tools for, let’s say, a decade. What’s important is how you can coach through that. We are finding that people are loving getting their hands back on the tools and they are enjoying being truly accountable for their results.

Some of the focus of this year’s report is on legacy companies and their unique challenges. I read something where you talked about incremental versus unachievable aspirations. Can you tell me a little bit more about what you mean? How is that a particular problem for legacy companies?

There’s a lot of fun poked at legacy companies, so I understand that it can feel disingenuous if you’re not working in one of those organizations. In a company like Google, its purpose is to literally change the world. The company holds itself to a lofty and unachievable mission. If you compare that to a lot of legacy companies, they have achievable and incremental missions. One of the obvious outcomes of that is if you’re striving for something that is incremental, you’ll hit your goal but you are also going to get small, incremental results.

One of the really interesting things you see in these more contemporary companies is that they’re failing every single day to achieve their purpose, and they’re comfortable with that. That comfort with failing starts with the audacity of their mission, whereas with a lot of legacy companies, you can see the fear of failure baked right into the beginning, which is how incremental their purpose is.

Is it a problem in digital because those unachievable aspirations are necessary? Do these companies underestimate the magnitude of the change they’re facing?

The magnitude of the change may be underestimated, but culture needs to start from the leaders. It needs to start from the top. I’ve rarely seen a company culture revolution start from the bottom up. If you are in an organization that has an unachievable purpose, and the leaders are demonstrating their comfort with their failure to meet that, it’s a good thing. They will feel relentless energy to go for it. The really important cultural aspects of failure and of curiosity come when you’ve got a leader who is willing to recognize that you’re not living up to the purpose that you’ve set for yourself.

What’s ANZ doing as a result of that? Are you making changes so that the company is trying to fail or is more comfortable experimenting?

ANZ has committed to a new purpose relatively recently. Our purpose is aspirational — it’s to shape the world where people and communities thrive. That’s not a purpose that says, “We just want to be a solid bank with an adequate shareholder return.” We have definitely taken a leap of faith in creating a more aspirational purpose. But we’re not really feeling anxious about it because we want our actions to speak louder than going out and promoting the purpose externally.

But internally we’re doing a lot to bring our purpose to life. Our CEO, Shayne Elliott, absolutely starts with our purpose, and it’s being felt throughout the organization. We’re now starting some of the organizational changes we need to make — some are capability changes, and other bits and pieces — because it’s all going to anchor around the purpose.

We have found in our data this year that the presence of experimenting does not really separate companies with what we call lower digital maturity from the more mature companies. It’s more what they do if those experiments are successful. Do they use those successful experiments to drive change across the organization? Does that connect with your experience? If so, how do you do that? How do you scale experiments to drive transformative change across the organization?

I 100% agree with you. When I came to ANZ, and when I was working with other legacy or 20th century companies, you go in and you say, “Oh, gosh, it would be great if we could do X.” And those who have been around for a while say, “No, no, no. We’ve done that. We’ve tried it.” These companies have often experimented in so many fabulous ways, and then they don’t go anywhere. One of the observations I made with a colleague of mine, Gerard Florian, who came in to lead the technology group, is that we got sick of identifying these green shoots. We’d say “If I see one more green shoot, I’m going to scream.” The reason is that we need those green shoots to actually grow into something. To me, this comes back to having the right capabilities and the right structure.

To incubate something requires a different set of capabilities than it does to scale something. Together with our CEO at ANZ, we are creating a group that is going to incubate. We’ve also got some ideas that we’re talking about scaling now, and we’re going to take them out of the incubation group and create a different set of capabilities. You need the management systems to be able to identify those green shoots or those incubation products, but you also need the different capabilities to scale them. But I completely concur with your observation that there are endless experiments, a lot of them successful, that don’t go anywhere in legacy companies.

This goes back to the distributed decision-making, where you almost need numbers to help drive influence through the organization, because leaders can’t possibly be in direct contact with all the employees out in the field. Have you changed your metric program, or how do metrics come into play here when it comes to distributed decision-making?

One of the differences between a legacy company and a modern company is that while people at a legacy company will say that they drive their business through numbers, a modern company is doing that on steroids. The mathematical understanding of a business that an Amazon or a Google will have is, quite frankly, light-years away from what many legacy companies have. When you look at the strategies of a 20th century company, there are a lot of words and relatively few numbers. If you look at the strategies of a more modern company, it’s largely driven through clear numerical goals.

So, one of the capabilities we’ve been working on as we move to distribute leadership is that we need the leaders of each part of the business to be able to deconstruct it so that they deeply understand the math of what drives customer acquisition and the math of what drives retention and loyalty. If they understand that, when they’re cascading it, they’re cascading really hard, crunchy numerical goals through the group. We have made a lot of progress on this in the past six months.

Maile, I want to thank you so much for your time. As we wrap up, are there any questions we should have asked that we didn’t?

I think an interesting one is the role that the importance of ethics and regulation plays between these legacy companies and some of the more modern companies, and the different standards that both the regulators and the customers have for some of these older legacy companies and the newer companies. It’s going to be really interesting how both sides play that out.

Are you saying that the legacy companies are held to a higher standard or a lower standard?

To a higher standard. Let’s look at open data. A company in the U.K. and in Australia has to comply with a regulation from the government saying that it needs to open up its data, which means that as a customer, you can request the data that the bank has on you and it’s initiated to a third party. Recently research in Australia said customers are assuming that if they demand that banks share their data with Joe Blow organization and anything bad happens because of that, the banks are responsible and will compensate them.

One part of it is that most legacy companies, particularly ones in highly regulated industries like banking or energy, need to have their risk and governance framework retrofitted to enable some of the things we’ve talked about, and to enable that risk-taking.

Actually, some of the invisible work I spend a lot of time on is with our friends in risk and legal to say, “OK, how can we make sure that some of those invisible parts of an organization are either massive enablers or massive robots? How do you rebuild those to support it?” That needs to happen now. I do quite a lot of work in that area. That’s all the internal stuff, but again, it’s going to start hitting pretty soon, if it hasn’t already, with government, regulators, and consumers.