Roche is one of the world’s leading biotechnology companies. It employs thousands of scientists and other R&D professionals around the world, dedicated to discovering and developing medicines for some of the world’s most intractable diseases. While the company’s approach to innovation has always been deeply collaborative—and remarkably productive—in 2018, it pivoted to agile, the organizational construct that is sweeping across industries. In March 2019, two leaders of Roche’s agile transformation in product development—Frank Duff, senior vice president, and Malte Schutz, vice president—met with McKinsey’s Aliza Apple and Steven Aronowitz to discuss the origins of the work, its progress, and its potential. An edited version of the conversation follows.
McKinsey: We spoke last year with Roche about its distinctive senior-leadership program on agile mind-sets. But let’s go back to first principles. Why agile? What motivated you to explore agile?
Frank Duff: A combination of internal and external factors led us to agile. Internally, our people describe excessive complexity in processes, systems, and governance that makes it challenging to get work done. This negatively impacts employee engagement and leaves some feeling removed from the mission that brought them into the company. Conventional approaches to address this have been insufficient in moving the needle.
Externally, digitization in healthcare and the insights associated with advanced analytics compel us to evolve. Development timelines are accelerating, and competition from traditional pharmaceutical and new technology players is rising—and so are the expectations of patients, regulators, and other stakeholders. We need to think very differently about the way we develop our medicines.
Malte Schutz: One quick story on our internal complexity: About a year ago, I was working in China. One of our physician scientists came up to me to discuss a new project. He said, “It started really simple, but then we complexified it.” That’s a new English word, but it totally describes how we operated. And the consequences are profound.
We serve patients with life-threatening and life-altering diseases. That’s an incredibly strong purpose. Unfortunately, we had built up internal structures that moved people away from that purpose. With agile, we want to remove a lot of the complexity in the organization so that that link to the patient—the reason why we do what we do—is a lot clearer and more direct.
McKinsey: Does agile help you operate like a technology company?
Frank Duff: We are exploring how to best apply the constellation of mind-sets, behaviors, and methodologies associated with agile. The emphasis placed on each element is likely to differ across industries, but we believe that these new ways of working support our ultimate objective of bringing innovative medicines to patients faster.
Malte Schutz: We have a long and successful tradition, and we recognize that agile has to build on that. And we realize that it’s a journey. Agile has to fit our organization, and it’s not a “reorg,” where we go from one state to another and arrive at a new steady state. Working agile means continuing to evolve, improve, search for alternatives—whatever it takes.
Frank Duff: Someone on our team said that if we get this right, we won’t need another big transformation initiative in five years, because we will have never stopped transforming.
Adapting agile for biotech
McKinsey: Let’s talk about the ways you have customized agile for your company and your industry. Is there a macro approach to it? How are you thinking about it?
Frank Duff: If there is a macro, it’s that many of the mind-set and behavioral shifts associated with agile are readily transferable to pharma. We’ve encouraged these shifts across the organization, and the impact has been positive. We’re trying to be a little more focused in the way we experiment with agile tools and methodology.
Malte Schutz: The sequence is also important. We started talking about an agile mind-set long before introducing agile tools. If we had immediately started trying to do everything in scrum, we would potentially just reformat our work and not gain any of the true advantages of agile.
McKinsey: You use tools to apply the mind-set, as opposed to using tools for the sake of a new process—is that right?
Frank Duff: Yes. We’ve been promoting that idea pretty heavily, that the mind-sets and behaviors have to come first, then the other supports, and not the other way around. We talk sometimes about how easy it is to fall into the trap of “fake agile.”
Malte Schutz: That’s when you walk into a room and say, Let’s do this in scrum and Post-it and kanban and whatnot, but you haven’t really asked questions like, Are the right people in the room? Do we have the right objectives? Do we have the right customer orientation?
McKinsey: So, it comes back to mind-sets. Can you tell us more about the mind-sets, old and new, and the ways they differ?
Malte Schutz: One of the key pieces is empowerment. In the past, our teams interacted with multiple review committees. Teams worked something up and took it to a committee for discussion, input, and decision making. That meant handing over some of the ownership, because it’s the sign-off of the chair of that review committee that you are after, as opposed to truly owning the clinical-trial design, or the overall molecule strategy, or whatever the team is working on.
We had to flip things around. The mind-set now is not “I work on this and then I have to go to this committee,” but rather “I need to achieve X within a certain period of time, so I have to ask for input from Y to do that.” For the committee members, their roles have changed from waiting until a review meeting hits their calendars to becoming providers of advice and guidance when approached. Their mind-sets change to include a strong service orientation.
Principles of agile design
McKinsey: How have you put the agile mind-set into action?
Frank Duff: We’ve tried to reflect the new mind-sets in our operating principles. Greater empowerment and accountability across the organization is certainly a key theme. Another is “fit for purpose,” in terms of both how people are working in teams and how they approach their individual deliverables. We want to get away from unnecessary perfectionism and instead get to a “good enough” deliverable for consideration by our stakeholders. To do that, we rely on evolutionary design, a critical element of agile.
Malte Schutz: In evolutionary design, a technique of making incremental improvements to a basic design, you have to ask yourself up front, What do I need here? And in our industry, it’s frequently very close to perfection. It would be really unwise to open enrollment for a clinical trial whose design isn’t as refined as it possibly can be. But we can use evolutionary design as a better way to arrive at that optimal study design. Take one of our principles: advice seeking. The team uses rounds of internal and external feedback on its initial idea, and the design matures through stages of its evolution. I’m convinced that working this way is more likely to deliver an optimal study design than the old process, in which a team would hunker down behind closed doors, spend a lot of time, and only show the study design to the world once it was “perfect.”
McKinsey: What about safety and quality? Are they compatible with evolutionary design?
Malte Schutz: If your deliverable needs to meet a high regulatory bar, that might be the equivalent of the version-four or -five release of a new software program. We are still doing iterative design, but we are keeping early versions in the project team and are learning from them. We apply learning loops that are characteristic of evolutionary design, and we don’t implement the earlier versions. To some, that begs the question, Why do you even do it? We do it because with these earlier versions, it’s easier to get expert input. It’s easier to socialize and discuss them, as opposed to developing something entirely within a team and only showing it when you think it’s perfect.
McKinsey: If we could press you a bit on that: critics routinely say that agile is not possible in a highly regulated industry, like oil and gas, banking, insurance, or, indeed, pharmaceuticals. What would you say to them?
Frank Duff: Certain things that we do are regulated and have to be done in a very standardized way. But much of what happens in a large global organization is not regulated; much of the process complexity we experience is self-generated and self-imposed. If you separate the things that require regulatory rigor from things that are just ways of operating your business, you’ll find many areas where agile can be applied without concern.
McKinsey: You’ve mentioned the principles of advice seeking and fit for purpose. What are the others?
Frank Duff: Another is customer-centricity. The patient is our ultimate customer, though not our only one. We’re encouraging teams to better understand and focus on what the customer needs in all aspects of their work. Yet another is being integrated and global: thinking beyond your team or function and considering the impact of your work on the broader organization, with the bigger picture in mind.
Malte Schutz: These principles are so important. In the past, while we would have overarching principles, we would follow them up with standard operating procedures that described in great detail who did what, when, and how. Now, people have to interpret the principles for themselves and for their own work. Some are really happy with that; others are still asking for the template or the playbook. To them, we say, Well, the principle is, for example, to seek advice, and you have to decide when you reach out to advisers. You have to decide. You’re empowered.
McKinsey: And what’s been the effect when you point people back to the principles?
Malte Schutz: I believe that it unleashes creativity and independent thinking. People will focus on outcomes of their work, as opposed to required process. We used to build timelines back from committees’ decision dates; that is now a thing of the past. Today, a team leader asks first, Can I make the decision myself? Or, in the context of the broader portfolio, Is it a decision that affects other programs? If so, then I probably need to take it up a level, look at impact and interdependencies on neighboring programs, and get those people involved.
Toward an agile culture
McKinsey: You mentioned some people were struggling a bit. What are you doing for the people who are hesitant to jump on the train with you?
Frank Duff: We understand that when you’re shifting culture and encouraging new ways of working, people are going to be in different places on the change curve. Some, the early adopters, will be way out in front; a proportion will be in wait-and-see mode; and others will be hesitant and even resistant.
We’ve taken a multipronged approach to help people move along the curve. We started with clear and consistent messages from leaders across the company about the direction we are heading and why. We conducted listening tours with different stakeholder groups to address people’s concerns and do some myth busting. We also established a network of ambassadors across various functions who serve as trusted internal resources to support the change. Additionally, we launched a series of well-subscribed interactive webinars, some with over a thousand participants, and shared pioneer stories about teams that are making it work and how they feel about it. These stories make it real for people. All of that said, we know we’re not done; this will continue for months as we encourage the entire organization along.
Leading agile change
McKinsey: How does senior leadership support the transformation?
Malte Schutz: You need to offer people strong and unwavering senior-leadership support and create psychological safety. That’s when people feel comfortable to take smart risks. In part, what that means is that senior management has to stay out of some of the details.
McKinsey: Support by staying out—can you explain that?
Malte Schutz: Leaders need to clearly signal what they expect, their desire for transformation and change—and then live it. You can’t empower a team and at the same time ask it to report out every minute detail of the project.
Frank Duff: Another important element is changing the mind-set of leaders, from ultimate decision makers to coaches and facilitators who ensure that those making important decisions are asking the right questions and thinking about the right things. In the new model, we strive to serve those making it happen. We facilitate and remove roadblocks. We also try to make it clear that the unnecessary perfectionism is no longer required—and to model that consistently. People need to know that they can come to us early in the ideation process and bounce ideas off us without concern.
McKinsey: The beliefs that underpin leadership must change, too. Many leaders feel like the way they provide value is to command people and control the organization. That’s quite different from seeing yourself as enabling and supporting. What’s the difference in what you have to believe?
Frank Duff: It’s less about seniority or the size of your organization: it’s about our ultimate purpose of bringing new and innovative medicines to patients. I gain tremendous personal satisfaction supporting our people as they pursue this goal.
Malte Schutz: I keep coming back to humility. If you work in a traditional, top-down organization as a leader, you probably have more experience—and you might think you know more—than the people who report to you. In that situation, the chances are that you would display very controlling leadership behaviors. But if I believe that the people who report to me are an incredibly diverse pool of thought, talent, and ideas, then it is my job to bring out the best in them. If I am clear that there is always more innovation and creativity in a group of people than in my own head, then I arrive at servant leadership.
McKinsey: People often ask about the business outcomes of an agile transformation. Are there specific things you measure as markers of success? And are there any early results?
Frank Duff: We can measure many steps in the drug-development process and are doing so to ensure we are having the intended impact in delivering for patients. Staff engagement and the extent to which our employees feel connected to our mission are also important. We’re monitoring those as well. Early results are encouraging, but it will take some time for the full picture to emerge.
Malte Schutz: It’s interesting that, while we continue to have a strong drive for perfection, for many scientists here, perfection didn’t include the timeline to launching the drug. People would lose sight of the fact that too much perfection in the daily work would delay programs. If a drug gets to a patient six months later than it could have, that is not perfect. Now, timeline is part of our mind-set and one factor on which we assess our impact.
McKinsey: How about your day-to-day work? Have you seen big changes in the daily routine?
Malte Schutz: We used to have meetings for hours with 20 to 30 people. Now we have smaller working groups and fit-for-purpose meeting invitations. To keep everyone informed about what’s happening in these smaller groups, we gave people software tools and guidance about how to use them and how to keep our data safe, and we asked them to experiment. Nearly 40 teams in parallel started to experiment with the tools. After six weeks, we paused, found some creative communication examples, and disseminated those to accelerate the process of team-to-team learning and iteration. After about three months, we came out with some really good insights about the types of communication that were most helpful and those that didn’t work so well.
Frank Duff: The resulting transparency is an important element. We want people to think beyond their molecules, to think about the broader portfolio, so we need to provide tools and information that allow them to understand the full-portfolio context, all the trade-offs and decisions being made throughout the organization. That level of transparency is new.
McKinsey: That’s important, we’d say, because Roche is not a small company. Not only do you have a broad portfolio, Roche has product-development sites in five countries and is part of a larger global organization. How have these messages of empowerment, daring, and transparency been received in other parts of the world?
Frank Duff: I’ve traveled extensively to explain the “why” of agile and the rationale for our transformation. The response across our sites has been very positive. The pain that people experience from unnecessary complexity is shared across regions, and anything we can do to reduce that and focus our people on their important work for patients is well received.
McKinsey: Thank you for taking the time to explain your work.