Transcript for:
Agile Decision Making Strategies

Welcome everybody, my name is Arnaud Chevalier, I'm a professor of strategy at IMD and today we'll talk about agile decision making, so deciding when you don't have all the facts. Sometimes I call this session as well decision making under ridiculously high uncertainty and maybe you feel that's particularly relevant in this day and age. as we are all dealing with COVID. So just to give you a heads up of what we're going to talk about, first we'll look at an example of a decision going bad and then we'll identify some tools that you can use to avoid making the same mistakes. Then you'll have a chance to reflect on these tools and to identify which ones you think might be most useful for you.

and then we'll have a chance to interact on these tools. Throughout we'll get also a chance to do a Q&A. Francesca is in the back here helping me with the chat and Q&A so if for whatever reason I'm missing some of your comments she might come online. Sounds fair? And hello to everyone who keeps on joining us.

Wonderful to have so many of you here. So let's get started and to get us started I'd like to for us to go back a few years. So we're now in May 1944 and Nazi Germany has a stronghold on mainland Europe with the Allies being stationed in the UK for the past couple of years now, the US, Canada.

England and obviously a number of other countries have been amassing troops in England and everyone knows that an invasion is soon to happen. The Germans know that as well. They just don't know where it's going to happen. They are strongly suspecting that the Allies might invade in the north part of France, close to Calais.

and it makes logical sense. It's where the English Channel is at the narrowest. But also, they receive tons and tons of evidence pointing to activity going on in that side of England.

There they have a spy network informing that lots of troops are massed in this side of England. They also have conducted their own aerial reconnaissance and they have monitored radio traffic. So lots of evidence pointing to Calais being perhaps the way where the Allies would invade. But to be fair, they're also suspecting that the invasion might occur just as far south as the Cotentin Peninsula, or for that matter, anywhere in between.

So they have to prepare for this, and not knowing where the invasion will happen, and facing limited limitations in as much resources they can dedicate, they have to decide how they're going to deploy their limited resources. So they decide to do a hybrid approach. They build fortifications, static positions all along the coast over here. So they have all those static points over here. and throughout.

They're building that famous wall of the Atlantic. But they're also saying the way to combat the Allies when they come in is to have reserve troops, tanks, or panzer divisions in the back that can be ready to be dispatched wherever that invasion might happen. So the idea here is to The moment they know where the invasion is happening, they can dedicate, send all the pensions to where they must need it. Okay, sounds fair enough, right? Sounds indeed like an agile way to approach the problem and dealing with a problem that's broader than the amount of resources we have to handle.

Now we are late on September 5th. and local commanders in Normandy are reporting that they're hearing lots of planes over Normandy. So they inform the German high command and they request that their high command release dispatch to the reserve troops. German high command considers it and denies it, saying, nope, these are probably just bombing runs, we're used to that, nothing to worry about. Fast forward a couple more hours and early on September, on August 6, sorry, August 6, 1944, again local commanders reporting power troopers landing in Normandy.

Again they request the release of reserve troops, again their request is denied. A few hours later, we're now in the first hours of June 6th, about 5 a.m., and the invasion fleet is spotted. And again, the local commanders are requesting that the troops are released and again being denied.

And even in fact, after the invasion started, their requests for our reserve troops got denied again. The rationale for the German high command was, hold on, this is only a trick, this is only a decoy, the real invasion will happen indeed to the north over there and we need to be ready to dispatch our reserve troops over there when they are needed over there. Fast forward a few hours and In fact, by the end of August 6, the Allies have a foothold in Normandy. And by the time the Germans finally decide to dispatch their troops, it's too late. In fact, they've lost some time.

They had to dispatch them during the day. This is where the Alliance fighter bombers can come and pick them up. So they're significantly weakened. the panzer divisions on their way to Normandy and by the time the reserve troops arrive in Normandy, it's too late. The Allies have established a foothold from which they will eventually liberate all of mainland Europe.

So the Nazi had possibly a good chance to stop the invasion, but as we all know, they failed. So what happened? Well, what happened is that facing uncertainty, they wanted more clarity.

They said, we don't have enough information now, we can't act now, let's wait. And by the time they had sufficient clarity, it was too late. Their window of opportunity had disappeared, and whatever action they took might be too late.

Fine, okay, so that happened to Nazis during World War II, but we see this example again and again, also in the corporate world, right? Kodak and their response to the digital cameras, Nokia and how they responded to smartphones, research in motion, Palm Pilot, and again, how they responded to the iPhone, the paper industry facing digital changes, or even Boeing reacting to how they should have dated that 737. way back in 2007. All of these companies delayed decision making in the face of extreme uncertainty and by the time they were committed to make a decision it was too late. So how do we avoid this?

Essentially today we're going to walk through a three-step process that can allow you to do this. And it starts from a very simple point, right? The realization that we can't predict the future, but we can plan for it.

And that's the part where we should be focusing on. To do this, three steps. First, we need to define the futures with an S.

Second, we need to prepare for the futures. And third, we need to roll out our coping strategies, adapting as we... as is needed.

So let's walk through these three steps and to do that let's pick an example. Imagine that you're the CEO of an airline so using the amount of passengers that you transported one year ago, 12 months ago as a baseline so we'll give that a score of 100. Chances are you're now quite a fraction of that maybe 20 maybe 30 of that Okay, so how many passengers would you expect to be transporting in September 2021? So a year from today.

Just type in in the chat box what estimate 80, 120, 55, 40. So we are seeing here a bunch of estimates ranging as low as 25%, maybe, and as high up as 100 or even 110%. And that's fair, right? So you might be saying, actually, we do not know where we'll be. Well, we know, however, that we'll be in between. to extremes.

The worst case scenario maybe at 25% and the best case scenario maybe at, let's go with 80%. No big deal. But what we've done now is that we've book ended the future.

We've said, okay, there's a continuum of possibilities from a worst case scenario to a best case scenario. And well, We don't know where the future will be around this scenario. We know it will be a discrete point.

We just don't know where it is. However, we know it will be in that range. And that's important because now it allows you to move out and not worry about cases that are beyond this range.

So very simple first step of bookending the future, but still... enabling you to bring in some level of less uncertainty, if you like, in your decision, in the sense that there's some extreme cases you can exclude, and we have now a better idea of what we're dealing with. Great.

Next, I need to identify the forces that shape the future. So the forces that will lead the future to be trending towards one extreme or another. as well as their leading indicators. So again, let's imagine we are CEO of a airline, we want to predict what the future will look like at say 12 months from today, so clearly we're dealing with COVID.

What kind of forces do you think would be dictating how the future will oscillate between worst case and best case scenario. So whether we have access to a vaccine, say Bogdan and Stéphane, yeah, absolutely, I agree. What else?

So Hugo, maybe climate change. Okay, maybe. What else? Unemployment rates. Okay.

Border closure, economic recovery. Beautiful. I love it. And I agree. So there's a number of things.

As you can see, the chat keep on coming and now we have dozens and dozens of possibilities. So here what you might want to do is identify, okay, so there are lots of driving forces. Can I identify the most important ones?

And maybe you might want to say, The most important ones might be first the longevity of COVID. The time we'll be dealing with the virus. And in the worst case scenario would be if we had infections and death continuing for over a year, or if we had isolation policies remaining in place indefinitely.

That happens really, me as an airline means I will be transporting fewer people. On the other side, if the virus dissipates within a short time frame, if the isolation policies are lifted quickly, then it would be trending towards the best case. OK, so first, driving force might be that virus longevity. Second, I might be looking in terms of the global mindset. Things are changing.

And worst case scenario, maybe now we're having this ongoing. a match going on between the US on one side and China plus Russia. It's as well, it's unclear whether the EU will remain as strong as they were even just a few years ago.

All this might lead to fewer people taking place. On the other side, if we have countries starting to coordinate a response and we have trade deals in place, we have global supply chains, weaker border controls, all of this might lead people to actually travel more. And perhaps a third driving force might be the digital adoption, the ability of people to now replace what we used to do. face-to-face or in real life with online stuff. So again, if video conferencing becomes a thing, if we keep on doing these Zoom webinars, if we have new forms of entertainment replacing vacationing away, all this would lead me as an airline to transport fewer people.

On the other side, if we have an in-person meeting coming back, or remaining the norm, if we have that evocation in the way, bouncing back as people are saying, enough with being stuck at home. Now I want to take some exotic vacation elsewhere. This would lead me to transport more people.

So this might be the forces that I decide are important. And then I would be looking for their leading indicators. That is hints. that these forces are at play and also hints on the strength of these forces, the intensity of those.

And by the way, I stole quite a bit of this from my colleague and good friend, Professor Michael Wade, who's a professor of strategy here at IMD. So the point here, the point I'm making is that on this continuum of scenarios from a worst case to base case, forces are at play that will drive the future towards one extreme or another. Your job as a decision maker is not necessarily to consider all the forces, then it becomes way too complicated.

But rather to identify which are two, maybe three, maybe four forces, but not many more than these, because there's only so many you can track. You should be tracking. Again, it goes back to George Box's old saying, your own models are wrong.

but some models are useful and this is exactly what you're doing here. You're trying to model what forces are at play by considering only the most important ones. Okay, so let's review what we have so far.

We said we need to define the futures and that starts by identifying a range of futures from the worst case scenario to best case scenario and also identifying the forces that are shaping the future, that are pushing the future towards one extreme or another. Next, we need to prepare for these futures. So let's go back to it.

Let's assume there's an absolute perfect storm coming to us. So again, you're the CEO of that airline. Let's assume that 12 months from now, we are now dealing with the worst case scenario. What would we called success? What would be success then?

Yeah, maybe it's to survive. Okay, what else? Still be alive. So yeah, maintain 20%, not lay off staff.

So I agree. Any of those might be appropriate. But so in that worst case scenario, when we are transporting, let's say, 25% of the people that we transported in 2019, 12 months from now, maybe I define success as avoiding bankruptcy. And maybe I'm only a shadow of what I used to be as a company, but I'm still up, still ready to fight another day. Okay, so I can define success there, and I can define as well a success plan.

So to be successful, for me to be able to avoid bankruptcy, I might be saying, in this case, what I need to do is reduce my operating costs by so many percent. I need to find a line of credit for so many billion euros. I need to even find alternative revenue sources for so many billion of euros per year. Perhaps I need to. Exit altogether the transporting people.

business and get into an alternative business, maybe more cargo, maybe military, maybe get away from transporting. The point here is that I've identified even if a perfect storm happens, I know what success looks like and I can devise a success plan for it. I can also identify what I need to have done by when to reach to implement a success plan. If I want to, say, have a line of credit of so many billion ready 12 months from today, then that means that maybe nine months from today, I need to have finalized my loans. And then six months from today, I need to have secured my collaterals.

And then three months from today, I need to initiate negotiations with lenders, so on and so forth. In other words, I can create a path between where I am now and where my success plan would be again for that. worst case scenario.

Okay, I've done it for the perfect storm, I can also do it for the best case. What happens if indeed the best case scenario where we're transporting, what do we say, 80% or 110% whichever you think is the other extreme. I can do exactly the same steps, identifying what success would be.

In this case maybe success might be to be able to service all the demand that's coming to me, as well as my role at Tynine to get there. Right? So two extremes, two very different ways of being successful.

On the one hand, I need to contain my costs, I need to find alternative ways of making money. On the other, I need to have installed capacity to respond to all the people wanting to fly on my planes. And indeed those are just the two extreme scenarios, but you could develop a scenario for each for fragments or for where the future will be along that continuum of possibilities.

Good so far? Okay, so now what we've done is we've defined the future and then we've prepared for the future and in that preparation we identify what success looks like in the two extremes, but as well along the continuum, and what a success plan with its rollout timeline would look like for me to be able to be successful. Then the third and last step really is to adapt. What does that consist of? Well, we know indeed that the future will be a discrete point in that range.

You just don't know where it will be in that range. Okay, so how does that work out? Now, if we were to plot vertically here the probability of where the future would be, let's think about it.

So here I am, as of today, I'm monitoring my three forces, remember? I identified virus longevity and global mindset and digital adoption. And I'm seeing that, OK, COVID cases trends and continue to trends widely per countries.

Countries don't coordinate on opening borders. I might be seeing that the global population confines only to regional travels. And I might be saying, out of this, really, I cannot identify a high probability of where the success or where the future might be. And therefore, I have something that's.

fairly flat, right? Okay. And so your job at this stage is to continue monitoring it. And maybe three months from today, you will have, between now and three months, you will have gathered more information about how life looks like, right? You might find that the EU cases continue to decrease, that some borders are reopening, that some business travel resumes.

I'm making this up. But the point here is that all this information... All those indicators you're gathering are giving you an information of how big or how strong these forces that are shaping the future are at play.

And so that will give you an idea of where the future is more likely to happen. And you keep on doing this as you move along further in the future. six months from today, you might find that more forces, such as, you know, vaccine indeed is confirmed for late 21, or maybe early 22 now, that the EU is coordinating actions, that there's a little uptake in leisure travel, all this gives you some information of where the future might be. And eventually, you might have, well, eventually 12 months from today, the future would be somewhere, right?

But the point here is not to have a zero or one, a completely dichotomous approach, but really to form that, to get to that realization of where the future is by integrating that information and changing your mind, updating your thinking as new evidence surfaces. So as you do this, so this is the part where you're monitoring the future. You're monitoring the forces at play.

And as a result, you make up your mind as to what future is becoming more and more likely. But meanwhile, you also need to roll out your success plans. So since you're starting in an environment where you have no idea where the future is at, you first need to roll out essentially all your success plans.

And then... maybe 30 days from today, as you gain a little more information and you realize, hmm, it's now completely unlikely that the future will be the absolute worst case scenario or somewhere very close to the best case scenario. Since I know this is not going to happen, I can stop rolling out these corresponding success plans and focus instead all these resources into...

making the success plans that are more relevant to where the future will occur. And I can keep on doing this, right? As I gather more and more information, I can keep on restricting my operations, stopping these rollouts to focus on the green ones, on the success plan. that will correspond, that will allow me to be successful based on where the future will be, right?

So let's step back a little bit. And this is where really your job as an executive differs when you're making decisions under regular conditions, under regular uncertainty, and where you make decisions under, in this case, absurdly high uncertainty. Typically, we make decisions and try to be as efficient as can be, right? But what I want to do is not roll out a bunch of success plans, but rather I want to achieve my goals spending a little less effort, a little less resources to achieve them.

That's being more efficient, doing things with less. But really, when I'm dealing with complete uncertainty, I need to be effective. I need to be the thing that will allow me to survive tomorrow. I just don't know what the things are. So ideally, I would be doing both.

I would be both efficient and effective. But under extreme uncertainty, I don't know where that is. And so that requires a shift of thinking.

You can't be worried about being efficient when you're dealing with extreme uncertainty. Rather, you need to step back and be effective. Because the worst case scenario, is if you end up having deployed a case where you're prepared for a best-case scenario, you have all this installed capacity, you're ready to transport tons and tons of people, except what's happening is a worst-case scenario where nobody is coming to your planes, nobody wants to be traveling, and you have huge costs that you need to assume.

So your goal here is not so much to try to optimize trying to make this window as small as possible, but rather to minimize that distance between your rollout plan and the ideal rollout plan to correspond to the scenario you'll be dealing with. So in other words, you need to first focus on being effective and only then can you cut out some of the some of the fat that's not needed. You can discontinue some of the rollout plans and be more efficient. Okay, so let's summarize what we have so far. First, we have defined the future, again, as a range of possibilities between worst-case scenario and best-case scenario.

Second, we've identified which forces are driving the future towards one extreme or another, as well as the indicators, the signals that tell you that these forces are in play. Third, we've prepared for the future. We've identified what success would be like for the worst-case scenario all the way to the best-case scenario.

Again, different definitions of success, but success nonetheless. There is such a thing as being successful even in the worst-case scenario. And we've also identified what it would take for us to be successful, what needs to be in place for us to be successful, no matter what the future. looks like. And finally, we've adapted.

Here's the agile part where we have continuously monitored the forces so that we update our thinking as to where the future is going and doing so we have rolled out success plans focusing first and foremost on being effective and then on being Okay, so now is a good time probably to take a break and give you a chance to integrate this. So I'm going to release a poll. You should now have a window appearing. My question to you is, out of these six different actions, which do you think would be most beneficial for you in your case? For those of you who have already replied, my follow-up question, and we'll do this through the chat, but would be to ask you, okay, what concrete actions will you take to implement some of these ideas?

So think about this already, and then we'll go through that together. Meanwhile, if you have questions also, please remember there's a Q&A box. So remember to post them there and maybe upload the ones that you think are most relevant for you as well. You should see a box here with the end results. About a third of you replied that what you would most benefit from might be to identify what success looks like along the range of futures.

So anyone who replied this, care to expand on this? How are you going to do this? So, Velinda, thank you. You're saying draw the success scenarios as indicated, worst and best, create related actions for both monitor and modify as changes take place.

Okay. So Harold, yes, so I like your comment over there of monitoring forces rather than trying to predict them. And Maybe it's a little bit of both, right? Maybe you're mixing the monitoring and the predicting, knowing that it's really easy for all of us to fall prey of confirmation bias. We tend to see what we want to see, what coheres with our view of the world.

So monitoring or predicting, rather, is important, but it can also be dangerous. So we need to step back a little bit in this. So... Vercel, you're saying fast prototyping on idea and impact. I agree.

And I think this is the whole, in a way, this is captured in that flexible rollout at the bottom, where at the beginning, you really release all your success plans, but you're ready to adapt and to update in light of new evidence. All right, so back to your questions over there. So the most voted questions right now is Andreas. You're asking, how do you deal with trade-offs? For instance, for the worst case, you might plan for layoffs early enough.

But for the best case, you might need even more people to serve the demand. Right. So that's the real underlying pain, right? Some decisions.

are true folks in the road. There's a beautiful saying by Bezos like him or hate him, I have my opinion on the guy, but there's one beautiful way he has to put, to define, to clarify, sorry, to classify decisions. And he's saying some decisions are one-way doors, others are two-way doors.

The one-way doors are the one that once you've made that decision, it's really difficult to roll it back. Think quitting your job. or building a plant in one country.

It's really hard to roll those back. But you also have two way doors decisions where with a little bit of effort, with a little bit of resources, you can actually step back and course correct. And really, so to your question, Andreas, how do you deal with trade-offs?

You try to delay the trade-offs as much as possible. So you try to make these first decisions over there and you try to design your coping mechanisms, your success plan, so that these early decisions are two-way doors and only as you're gaining more and more certainty or less uncertainty on your future would you commit to one-way door decisions so the ones you can't come back to. Atif, you're asking what is the best way to define agile decision making, speed in decision making or timely decision making to have preparedness in order to have minimum or negative impact of potential risks.

So that's a great point, Atif. I think you need to start with a clear definition. Chances are you ask five different strategy professors of what's agile.

and I think you'd get seven different opinions. So what matters here is maybe that we don't all agree on what definition we use, but at least we clarify what agile means. In this case, in my case, my definition of agility is an ability to course correct as new evidence arrives.

So in other words, being Bayesian, you're thinking, I have a viewpoint, I have a course of action. new evidence arrives, hmm, I need to update this, I need to use that new evidence to update my thinking and course correct. Carmen, you're asking any tips to when making time-sensitive decisions when you don't have to monitor or adapt?

So, you are always making time-sensitive decisions, right? And this is a subject for a much... broader conversation and we should have a webinar on this soon but remember decisions you only have to consider three things for decisions you need to ensure you have a good question a good decision so make sure you ask the best question you can you need to have at least two different options ideally more than two but if you have just one option then there's no decision you just go with that one so you need to create the best set of options that you can, and you need to have good criteria where you're considering the things that really matter, not just to you, but also you and your stakeholders, so that in the end you can rank your options, deciding which one best coheres, has the least painful trade-offs.

Chances are there will be some painful trade-offs, otherwise we don't even think about these decisions. But the point here is to clarify them and try to go deep into understanding what these trade-offs are and elect the one that makes the most sense for you. So you might not be able to have a full-blown analysis for all the decisions that you make, but you can keep these three components in mind as you're making decisions. So it's a mental checklist. Do I have a good decision?

Good question. Am I really asking the question that I should be asking? Okay, do I have decent options? Do I have good criteria?

Okay, based on this, which one should I elect? Michel, you're asking what should the Nazis have done? So what should they have done? First of all, they shouldn't have been Nazis, right?

But if they have wanted to reply... To this case, my feeling is they should have released their Panzer Division, their reserve troops, as early as possible. Early on June 6, the Allies were extremely, extremely vulnerable as they were landing in the beaches of Normandy here.

Very vulnerable because they had no place to deploy all their superior forces. And so if the Germans had contained them there, they had a good chance of pushing them back. The moment they let that window of opportunity go away, then the Allies had more maneuvering space over here and enough space to deploy really all of their superior, at least in numbers, superior forces and now they could take advantage of this.

So in the case of the Germans, what they should have done is not wait so much to have more and more clarity, waiting, that craving of having more certainty before acting, but really move forward. Or in a spirit of being rapid prototyping, then maybe they didn't need to release all of their reserve troops to this area, but really release some of them earlier, faster to address the problem at hand. So Eric, you're asking the assumption of the model is that the leader makes a sharper assessment of the future compared to the general market.

So as an executive, I would say we plan as we normally do by working across the worst and best case scenarios. I cannot see yet how we can do better with the different methods. What's my view? So my view is that as a leader, you don't necessarily need to... The assumption...

It's not that I'm making a sharper assessment. Hopefully, as a leader, you are integrating the viewpoints of others, not just yours, but including the ones you gather from the general market. And by the way, prediction markets might be a good way to gather these diverse point of view of where the future is going. Integrate this. What matters most to me is that you as a leader have a range of possible ways.

to move forward depending not so much on the strategy that you had in mind before COVID, that's no longer relevant, or your preferred strategy. Your preferences are probably not really what matters. What drives is in this case when we're dealing with COVID under extreme uncertainty, what drives our decision making is externalities. It's what the environment is pushing on me.

I need to be prepared to respond to all those possible ways. And I need to roll out the coping or the success plans that will help me best do this. We're running out of time. So just one more question and then we go.

How do we rule out and remove what you rolled out without looking stupid like an executive? Your followers may think you don't. know what you're doing and the market will destroy the value of your share prices. So, Hugo, I love the question.

And this goes back to, I think, more of a psychological issues where we consider that being flexible and changing opinions is a source of weakness or it's an indication that we're weak. That we need to change, right? This is the fundamentals of Bayesian thinking. It goes back to my own definition of being agile. Sometimes you need to change your opinion.

You need to change your mind. And there's a beautiful saying by Keynes saying, when the facts change, I change my opinion. What do you do, sir? that idea that I need to always update my thinking in light of new evidence. So, Hugo, how do you do this?

You go back to your stakeholders. First of all, you don't start by setting them certainty. You are very clear in the early stages of dealing with your crisis that you are dealing with uncertainty. And you don't say, I know the way forward. You're telling them we are preparing for all the possible ways forward.

And then you keep on walking them through your thinking and say, we are evolving. We have now a better understanding. Things have evolved and we are taking these evolution. into consideration and as such we are course correcting as we move forward. So I wouldn't try to hide it.

I would try to embrace it. Having said this, there's also a time and a place for being flexible, right? A colleague of mine, Frédéric Danzas, and I, we are now interviewing top CEOs to understand how they make decisions and what type of question they ask.

And there is a consideration that keeps coming back where they're saying in some settings I need, I'm in problem solving mode and I'm working with my stakeholders to find the best way forward. In some other settings I need to have answers, I need to have solutions such as when I'm on the call with investors, by the way. So I need to, I have these two facets of my personalities. Sometimes I'm flexible. Other times I need to be bulletproof with my answers.

Okay. This is all we have time for today. So I'll let you go.

If you're interested in continuing this conversation, please do reach out. A couple of ways of doing this. So I'm on Twitter, if you fancy this.

my Twitter handle is down there to the left. We also at IMD are releasing a, we have an app to help people make better decisions. It's called Dragon Master and you can access it. It's free access at dragonmaster.imd.org. And if you're interested in problem solving and decision making, I have a blog where I every now and then talk about this stuff.

Thank you very much for joining us. It's an absolute pleasure to have you joining from all over the world. And I look forward to continuing our conversations.

Thank you. Hopefully, see you soon in Lausanne. Bye-bye.