Safe to fail

Edmondson: Leadership behaviour really matters: by leadership behaviour I don’t just mean senior executives. I particularly emphasise the importance of team leader behaviour. Research on that is pretty clear – the behaviour of those with more power, authority or status in any given workplace plays a crucial role in influencing the level of psychological safety. This is one of the reasons that, in almost any organisation, you will find substantial differences across teams in the level of psychological safety; it’s because it is so influenced by local leadership behaviours. But even if you’re not the leader, your behaviour matters. What you do affects your peers’ and even your manager’s psychological safety.

Edmondson: There are three things leaders can consider: The first is setting the stage, by framing the work and making a logical case that everyone’s input really does matter. This creates the rationale for psychological safety. It is a leadership responsibility to help everyone understand what they’re up against, the nature of the work they’re doing. If the work is highly interdependent, or complex, or uncertain, then it stands to reason that people need to speak up, or else things might go badly wrong. In a workplace characterised by interdependence or uncertainty, leaders have to care about psychological safety because they need to hear from people. The work itself necessarily requires people to be open, transparent, and frequently communicating. The work cannot be done well unless people take these kinds of interpersonal risks. Secondly, leaders have to invite participation proactively by asking good questions. You let people know that you understand your limits by saying things like, «I might miss something, I need to hear from you,» or «what are you seeing out there, I’m not an expert in your area, I’ll welcome your comments.» Most leaders think that they have implied that they are open to comments. But that’s different from being proactive and soliciting thoughts. Most of us will respond positively when someone asks directly for our thoughts. Thirdly, it is important to respond appreciatively – particularly to bad news, concerns or questions. And to remain fundamentally positive when someone has the courage and/or takes the effort just to speak up. 

«What you do affects your peers’ and even your manager’s psychological safety.»

Edmondson: I think you can transition an existing team. And I think it’s helpful to be as explicit as possible about the transition. We may or may not have gotten off on the right foot, but if we aim to do spectacular work on this project we’re going to need to engage everybody’s hearts and minds. So leaders need to say things like, «I propose to start the transition by welcoming your thoughts on how we’re doing and where you’d like us to be.» 

Edmondson: People are sharing their ideas, pointing out results that didn’t go as well as they had hoped. They are seeking feedback, sharing information, and asking for help. They are proactively exchanging opinions and assumptions as they try to get to the bottom of impossibilities in the project. 

Edmondson: Coaches can play a crucial role as an objective bystander who can see functional and less functional dynamics in a team. They can do two important things: Firstly, they can observe the patterns, which are helpful or unhelpful. People are open to their observations as they are not a team member, boss, or support net and have that clear objectivity. Secondly, they can offer suggestions how to improve the dynamics that they see.

Edmondson: The team effectiveness literature has long emphasised a structural perspective: if you get the structures right, it will be likely that the processes will go well and the team will be effective. What that research observes is that many times in organisations, people expect a team to do well but haven’t adequately considered if it has the right people, resources, or goals. What I found two decades ago, and added to the prior research, were two things: firstly, the mechanisms through which good structures help produce good outcomes matter. Secondly, there is still unexplained variance in similarly structured teams. For example in a hospital setting, where they have the same skill level, the same clear goal and resources, they can perform very, very differently because of differences in the team climate, beliefs, and in particular psychological safety. That suggests the possibility that if we can alter the psychological safety we can enable teams for higher performance that haven’t been performing well because of a poor climate or poor leadership.

Edmondson: When we talk about DNA in organisations, I think we usually mean culture, which describes the behaviour or routines that occur in the absence of a formal rule. The subtle programming that shapes behaviour or routines in an organisation in the absence of some kind of formal rule or task obligation among employees. Shaping culture is mostly about modelling behaviour, setting an example, and providing direct coaching and feedback, which also helps to build a culture that is needed for an ongoing learning. In this regard leaders shape the culture of an organisation, primarily in terms of how they show up, their own modelling. What they do is more important than what they say. The sent messages matter as much as the emphasised values to help shape the culture. For example, consider an organisation committed to advancing human health through the development of leading edge medicines. If leaders in that company constantly refer to and reinforce the need for experimenting, which involves many failures along the way, then they send a clear signal that you need to take risks, to fail fast, and learn fast. This helps to create a culture where people know what’s expected and what’s desired.

Edmondson: First it is important to understand the nature of expertise. Experts are those who master an area of knowledge or discipline. They have a knowledge reservoir and deep understanding of how cause and effect works in a realm, e. g. medicine or the law or engineering. It is a kind of false expertise that is reluctant to admit or acknowledge mistakes or gaps in one’s knowledge. The true expert understands that. True experts are in fact those who are perfectly comfortable with failure and errors. In fact, they have the confidence to recognise an error or a failure that a non-expert may not recognise. They also have the self-confidence to speak about errors or failures and learn from them. This is how they help the organisation figure out what needs to be changed to go forward and perform better. If you are in a, what we now increasingly think of as, VUCA world, a volatile, uncertain, complex and ambiguous world, it’s quite absurd to think about getting it right all the time no matter how much expertise one has. Nonetheless, sometimes we learn in very subtle but powerful ways to cover up, and not admit, and to fail to disclose or discuss when things are going wrong. Of course that leads to all sorts of problems, because eventually everything does come to light. But in such circumstances it needs to be reframed as an opportunity that we value and reward expertise, the needs for constant learning and change in a constantly changing world.

«True experts are in fact those who are perfectly comfortable with failure and errors.»

Edmondson: On the one end of the process knowledge spectrum is work where a great deal is known about cause and effect. For example the work at an assembly line is well described, such that if a small task is carried out at right time it will produce the desired result. On the other end of the spectrum is work that requires constantly experimenting because the state of knowledge for how to achieve a desired result is limited. If you are a cancer researcher you face many diseases you don’t yet have a solution for, and the only way forward is through experiments. You expect many of them to must fail. Consider the difference between the two ends of the spectrum. It would be absurd to imagine that the same way of working and/or of treating expertise would work in these two very different work environments. In concrete terms: if you are on an automotive assembly line you execute the task precisely as specified. You celebrate that precision. You need to speak up if you see that a task is being done incorrectly; this is nothing shameful. It just needs to be quickly identified. In contrast, in a laboratory, you literally need to celebrate every thoughtful failure, because it is the only way to move forward and get to new knowledge and new solutions. Often people misunderstand the true meaning of expertise, and think it means not admitting errors, never being wrong. This is not the case. Expertise means having the judgment to use failure to learn and improve and innovate. In many organisations, managers do not appropriately modify their responses to suit the context, that is, to match the nature of the work. For high uncertainty work we should develop good hypothesis, try them out and learn fast from the results, celebrate intelligent failure and make it discussable. For low uncertainty work we should celebrate successful adherence to an expert formula, along with successful identification of deviations that might lead to harm.

Edmondson: They are very different, of course. Technical risk involves something that we need to accomplish and today’s solutions are incomplete. We know there is risk, and we talk about it and manage it with expertise and good diagnostic processes. Interpersonal risk is more invisible, but just as problematic in organisations. Interpersonal risk means that I’m reluctant to say something or do something because I’m not sure whether you will think less of me if I say it. This can create business and technical risk for the organisation. Clearly both matter, but the way you treat them is different. For helping people take thoughtful interpersonal risk we set the stage by framing the work. We remind them of the complexity and interdependence of the work we do, so that people understand how much their input is needed. To reduce technical risk we pool the information and expertise we have, so that we don’t take stupid risks. We only take technical risks because we must – because no one has the full solution right now and so we have to learn together through experience.

Edmondson: In many ways the principles remain the same And the stakes are higher, the chasms and expertise chasms are greater. The potential interpersonal risks and concerns about image are greater. Leaders in this domain – project leaders and especially team leaders – have to be explicit about the challenge, the risk, and the opportunities this way of working offers. Everyone needs to sign on to a process of learning as we go.

Edmondson: The main difference is that our bodies of knowledge, our areas of expertise are likely to be further apart. We have less shared understanding. If you are an engineer and I am a marketer in the same company, we speak very different languages, but we both work for and understand our company, what it is all about, what matters around here. But if you are working for a city government, or in the education sector, and someone else is coming in from infrastructure and engineering, you have little in common. And so, it calls for far more deliberate process and more careful process management, and more explicit attention to building thoughtful experiments and cultural and psychological safety.

Edmondson: Yes.

Edmondson: I think coaches have an even greater opportunity to be helpful. Because they can insist on a good process and good structures to make sure we are communicating as carefully and thoroughly as we can about who is doing what. Additionally there is a particular need for what I call interface management: on the one hand there is some work that you and I can do on our own within our area of expertise, on the other there are certain parts of our work that come together where we have to make sure that our different parts are going to sync up together.

«A coach can be in a great position to step back, see the bigger picture and recognise where people may be at risk of missing something.»

Edmondson: It is a combination of project management, facilitation and coaching. It does have to recognise that there is a high potential for misunderstanding and details to be missed. A coach can be in a great position to step back, see the bigger picture and recognise where people may be at risk of missing something.

Edmondson: I think it is really more of the same. But maybe the need for structure could be higher because at least they have been described as having a shorter attention span, being easily distracted by competing demands and opportunities. Maybe they need even more coaching and support for staying on track. I think we have a tendency to incorrectly view all millennials as the same, but I’m not sure that it is accurate to think we need to have a one size fits all millennial’s approach.

Edmondson: I like to help leaders think about the combination of producing high motivation and engagement along with high psychological safety. Trust refers to relationships between two entities. It could be two individuals or it could be an individual and an organisation. Psychological safety, however, characterises the climate in a given workplace – such as in a team, a branch or a unit of people who are working in some proximity or interdependence with each other. In today’s world of high uncertainty and high challenge it is essential that people feel their work matters, are motivated to do their best, and know that they will not be punished or humiliated for saying something, asking a question or making a mistake. That their colleagues will welcome and value them as fallible human beings. We are all fallible human beings, and we err when we think we are supposed to be perfect.

Edmondson: Yes. And I think the opportunity for anyone to do that lies in three things: Firstly pointing out what we are up against – that is, the challenge, the uncertainty, and the interdependence. Anyone of us can say that is really interesting, or important. Secondly, ask questions. The more you ask genuine questions of people, the easier it is for people to speak up. A question is a direct invitation for input. And that’s obvious and profound at the same time. I think that as we ask genuine questions of each other, we are demonstrating respect for each other, assuming we show genuine interest in the answers. The third thing is to be willing and able to come forward with our own thoughts, concerns, and vulnerabilities.

The interview was conducted by Tina Dörffer.