A Deeper Dive into "Working Backwards"
Sohrab: Welcome, everyone, to our next "Agile Insights" conversation session. Today, I'm again hosting Colin Bryar, whom I first talked to, I think it was in the summer of 2021. Since then, almost one and a half years have passed. And I'm super happy to be able to have Colin with us again. And we will obviously be talking, again, about his book, which he co-authored with Bill Carr, "Working Backwards."
Now, before I go into what we will talk about today, I'd like to quickly hand over the stage to you, Colin. First of all, welcome. And maybe you can give us a brief introduction about who you are, what you've done in the past, and then everyone knows what we will be talking about.
Colin: Okay. Well, thanks again for having me on the podcast. And so, just a little bit about me, I worked at Amazon for a little over 12 years, from 1998 to 2000. So, when I started at Amazon, there were about 100 people in the corporate area, and got to see it grow from that to the big behemoth that it is today. We just sold books just in the U.S And by the time I left, we had AWS pretty much selling anything everywhere, and just got to see all that growth in digital, video also.
And so, I co-authored a book with Bill Carr, who worked at Amazon for 15 years, and we had just been reading articles, watching interviews, and reading books about how Amazon worked. And we felt that the stories were either incomplete, inaccurate, or just didn't really get to the essence of how Amazon works. So, we decided to get together and write a book called "Working Backwards," which really is an inside out looking view of how does Amazon do it. And so, we wanted to write down some principles and practices that Amazon used, really to help the next generation of business leaders.
And since the book came out in February 2021, we've been working with organizations who want to implement one or more of these business practices, which we view really as advances in management science that Amazon has created or modified. And helping different companies implement one or more of these processes that have made Amazon successful.
Sohrab: Thank you, Colin. And I think you nailed it when you said help other organizations to some extent discover these advancements in management science, because so many organizations, organizations that you work with, or organizations that I work with still apply a lot of those things that used to work in previous generations, previous also business environments. But maybe today's business environment requires completely different approaches also to the topic of management. But we will go there.
So, you and I talked briefly before we kicked off or before we started the recording. And what I want to talk about today is this book is split in two parts. The first part is being Amazonian. And in that part you talk about the different leadership principles and a few other things. And that is one of the topics that we will also cover in today's conversation. The second part is the invention machine at work. And in that you cover stories about the Kindle, ultimately all the way up to AWS, which became also this not only big revenue generator, but also a big cash generator for Amazon as a company.
That second part I won't dive into today because the second part of our conversation, I want to actually focus on the work that you mentioned you've been doing for the past two years, helping other organizations apply lessons from Amazon to bring back their innovative age. Because every organization at some point was a startup. I assume most organizations, maybe other than those who were initially run and operated by a government, and had a monopoly, like telco companies, were to some extent innovative. Otherwise they wouldn't have become big. But many of those organizations also lose that ability to consistently innovate. So, being that invention machine, and that's the part where I'm interested to learn more and more from your experience, what has worked based on the work that you've done in the past two years.
The Amazon Leadership Principles
Sohrab: And you said this would be okay with you, so we'll dive right into this. Now, you start the book with the leadership principles, and if I remember correctly, it basically starts with how the leadership principles at Amazon were initially created. And what an active role Jeff Bezos himself as founder, and back then CEO, had in the creation of the leadership principles. Can you walk us a bit through this process, A and B, point out some of the most important of those 14 leadership principles from your perspective?
Colin: Okay, yeah. So, Amazon had core values before, when I started, and it was a very long list, and then management principles. And it was back in the early 2000s where Jeff Bezos, he really just wanted one set of leadership principles that all leaders in the company could look at and use to help make tough decisions. And at first when it started, I thought, "Oh, here comes another HR, human resources exercise." But I was just impressed with the energy and effort that Jeff put into reading through each and every word and principle. Going into this, we didn't know how many principles we would end up with.
We did end up with 10, actually. It went through 1 revision several years later to go to 14, and now there are actually 16 leadership principles at Amazon. So, two revisions after the initial one that came out in 2004, 2005. So, pretty durable, but they do change over time. And so, that is one thing to notice. But what Jeff really wanted to do was to create a set of first principles that everyone in the company could use to help make tough decisions. And one of the things that I've noticed about Amazon is how consistent the decision making is across these industries and businesses that are quite different. AWS is a B2B business. You can have the e-commerce business, digital streaming [inaudible 00:06:38]. Amazon has an airline fleet.
But if you look at one of the things that they have in common is they all start from the set of Amazon's leadership principles. So, Jeff wanted to have principles that people could rely on when you need to move fast, when you don't have all the information that you'd like to have. Jeff would often say that for most types of decisions, you want to have about 70% of the information that you need. If you wait to collect 100%, you're probably moving too slow, and your competitors are going to get there first. And also, he wanted these leadership principles to help when senior executives and Jeff wasn't in the room. So, if you read through Amazon's leadership principles, some of them are pretty quirky, but they're all actionable. They actually help you rely...you can rely on them to make judgment calls. And so, they're not just posters on the wall.
The last thing that Amazon did that was notable about the leadership principles, it's not just enough to write them down and to memorize them. Actually, right now there are 16 leadership principles and over 800 words. So, I don't know anyone who can recite all of them verbatim. But what you have to do is you have to take those processes, and then weave them into every major principle. And then weave them in every major process you do at the company. So, you don't have to memorize them, you just act them out every day. And then it becomes part of who you are.
And so, that was one of the, I think, really noteworthy things that Amazon did with its leadership principles, is that just by doing everyday activities about how you interview and hire people, how you conduct and measure things, how you decide what products to build next, all these processes are built from the ground up with Amazon's leadership principles in mind. So, you act them out everyday versus memorizing them. So, those are a couple of notable things.
And any organization that has been around for any period of time, you already have some set of leadership principles, whether you've written them down or not. And so, at Amazon, we were really uncovering and discovering them versus creating them. And so, that's another important thing as you embark on a journey, and your organization to really codify your leadership principles, it is more of a discovery exercise than saying, "Here's who we should be." You, first of all, want to clarify who you are, and then you can have some aspirational principles layered on top of that.
Sohrab: Now, you mentioned, well, in this short period of time, so many golden nuggets. One was not posters on the wall. You don't memorize them, you act them out. I think this is super important because, probably you've seen it, I've seen it, so many organizations go, and you mentioned your own fear, "Is this going to be an HR exercise?" They run these exercises. They come up with a set of core values, or leadership principles, or however they want to phrase them or call them. But then it's just a poster on the wall. It doesn't weave into the processes that exist. And based on that, they're not acted out on. Maybe there's a few people who can recite them because it's not always 14 or 16, as you mentioned today. But it's less. But people don't act based on them. And the action piece, I think, is what ultimately results in the culture of the organization, being shaped on a consistent, on a daily basis.
The other part that you mentioned was what I found particularly interesting that last...what was that last piece again? I just had it on my thought and then I lost it. Doesn't matter. I'm sure we'll come back to that later. So, let's go into this. It was this. It was about uncovering, discovering rather than creating them from scratch. And when I read this book, I really felt you took me on this journey on how these leadership principles were uncovered, were discovered. But I think once you managed to uncover them, codify them, really make them transparent to everyone, you probably, and correct me if my assumption is wrong, you probably also created alignment around them, which resulted in more consistency within the S-team, but also other leaders within Amazon's leadership team to apply these leadership principles on a daily basis. Is that assumption correct?
Colin: Yeah. So, what we did is we took a look at who we thought the top leaders at the company were. Ask them, "What are the things that you do that make you successful in your job?" And then we asked other people, "Who do you admire as a leader at Amazon?" And we talked to them. And so, you can think of some of these leadership principles as personas of the people that we talked to, or elements of the persona.
But an example of how we wove it into a process, and you want to have consistent behavior. The way that you can rate or review your peers, your managers, manager, or people who work for you. Some of the questions during whatever cadence you have for reviewing people are, "Tell me the top...give examples of three to five examples of when this person has embodied the leadership principles well." And so, you have to name a leadership principle and give examples.
And then the other set of questions is, "Well, give three or five examples where this person has some room for improvement, or did not embody leadership principles." So, that is the lens of how you evaluate other leaders, and employees, and peers at Amazon. You have to walk the walk, and talk the talk with these leadership principles. And you get feedback, "Oh, yeah. I guess I could have done a little better. And using this leadership principle to help us through this tough situation."
And then senior leaders, Jeff once told me, "Unless you're sick of hearing yourself talk about these principles, you're probably not saying it enough." And that's because...well, I spent a year reading through these things when I was working with Jeff. And a new employee comes on, they may have...look at him once or twice, but it's a leader's job then to point out, "Well, here's why we're going to make a decision this way, and walk the room back." Because if we start from these principles, which you cannot compromise on the leadership principles, you start from there and then you logically walk through to, "So, here's why we're making this decision." You kind of explain to people along the way about how you can use these leadership principles to make the tough calls, because you have to move fast and business is hard. But you have to have some constants. And the constant we never compromised on at Amazon were the leadership principles.
Sohrab: Yeah. So, now you mentioned a few of the things that Jeff contributed to this. For example, saying, "If you don't find yourself talking about this too much, you're probably not talking about them enough." You also mentioned earlier that he played a very active role, and you were surprised by this, right? How the CEO, the founder CEO, in this case, was actively involved in doing this. Now, can you walk us a bit more into the details on what he specifically did in shaping or uncovering, to use your word, this leadership principles?
Colin: Well, the whole exercise probably took about seven, eight months, maybe even close to a year to come up with it. And it was Robin Mensinger, she was an HR executive, was driving the project. She would go interview a bunch of people, and then come back with, "Well, here's what we found." And we'd sit down with the CFO at the time, Tom Szkutak, Jeff, Robin, and I would sit down, and Mike George, the head of HR. So, a small group, we would look through, "Here are the findings," and start to craft these leadership principles. And some of them were, as I said, kind of quirky about it. But we really wanted it to be more to help you and direct behavior to make challenging decisions.
And so, when we came up with this list, we would keep asking do we have a necessary and sufficient number of leadership principles to accurately assess how leaders should operate. And a great example is, hire and develop the best. When we work with organizations, and they have their core values or leadership principles, those are synonyms. So, some call them core values, some call them leadership principles. Same thing as far as we're concerned. One that's often missing is something about how you hire, retain, and develop talent. And any growing organization that is a key function that a leader has to perform.
So, you need to have some type of leadership principle or core value that says, "How do we do this?" If you don't, then what you have is you have a set of leadership principles, but then you have another set of important activities that aren't encoded in who you are in your leadership principles. And you somehow have to kind of put these together.
So, what Jeff really wanted was one full set, again, to have a necessary and sufficient set of first principles to guide leaders on how to do everyday activities at Amazon. And so, when we came up with 10, we thought, "Well, 10 seems like a lot, but we can't trim it down to 9, and we don't...there's number 11," there were other ones on the cutting room floor. Well, this isn't really always true because you want your leadership principles to really always be true in whatever situation that you have.
And so, 10 was what we ended up with and we thought, "Well, it's a little bit more than most companies, but this is who we are, and so we're going to stick with it, and modify it over time." And as I mentioned, there were 2 modifications over the next 17 years. So, they turned out to be fairly durable.
Sohrab: So, a few were added but the initial 10 are still there?
Colin: When that first revision came, a couple of them were merged, and then one or two of them were added. Learn and be curious was example of one that was added in the revision in the 2010-ish, 2012 time frame. But yeah, similar merge between a few new ones. And then there were 16, there were actually 2 new ones that were added right at the time Andy Jassy took over as CEO.
Sohrab: Yeah. So, what you also mentioned... Thank you for sharing this so far. You mentioned that these leadership principles have to be woven into the existing processes, or new processes have to be developed with the leadership principles in mind. And I want to go into four areas. One is about organizing work, the other is about communication. The third one is about working backwards, which is the title of the book. And it basically means how you involve customer perspectives into the work that you do. And the fourth one, the last one is metrics. And for each of these, you have a subchapter in this book, or a chapter is part of the different parts.
Sohrab: And let's always start also with a few leadership principles because there's not only one leadership principle that is supported or covered in the aspect of organizing. There are several. So, one of them for me, you bring up the topic of single-threaded leadership. And for example, one of the leadership principles, if I remember correctly, is dive deep, right? So, leaders have to be able to dive deep.
Now, can you first share with us what single-threaded leadership means in the context of Amazon? And then we can connect it with some of the leadership principles that you have. And later on, we can maybe see to what extent this can be applied in other organizations.
Colin: Sure. So, single-threaded leadership, it was a concept that evolved over time at Amazon. It's what we describe as some of these processes in the book. Sometimes we describe the journey on how we got there, where we started off, and we were trying to solve very specific problems. And the solution, we ended up with...took a while to get to. And that's the case with single-threaded leaders.
So, the problem we were trying to solve with single-threaded leadership, which would become single-threaded leadership, was we were adding more people to...we were growing like crazy, but we were not getting enough work done. So, we were spending too much time coordinating, talking about doing work, resolving dependencies, asking other groups, "Hey, can you get a line so we can get this initiative out the door?"
So, we were slowing down. And if you extrapolated where we were going to be, the bureaucracy was going to crush us over the next couple of years. And so, there was a fork in the road. And at that point, a lot of companies can choose one of two things. You can say, "We're going to build some collaboration and communication tools so everyone knows what everyone else's working on." You have got all these pretty dashboards. And so, you can figure out what anyone is doing in the company, and will make it easier for people to collaborate. A lot of companies choose that path.
The other path, which is the one we chose, was, "Let's view coordination and communication really as a defect." So, how do we eliminate most of that and allow groups to not have to talk to and coordinate with a bunch of other groups to get their work done? And we first started off with this concept of two pizza teams, which are small teams with very encapsulated set of activities. We realized that over the years, the thing, the driving factor wasn't really the size of the team. It was having a single-threaded leader was what made the difference.
Because when we took a look at, "Well, which projects or initiatives are stuck and why?" We would ask a couple of questions. Jeff would or any leader at Amazon would ask, "Well, who's the most senior person working on this initiative and no other initiative?" And if you're trying to build $1 billion business, $100 million business, you can scale it up or down, if you have someone who doesn't really know what that looks like, who's never done it before, if you have a junior product manager who's in charge of this, and that's what they're working on full time, we found more often than not, you're not going to build that...achieve what you want.
So, we would approach...if an initiative is big enough to be worth doing, we would make sure that we had a single-threaded leader who knew what the end goal looked like. So, he or she could grow the organization, and organized in order to get there. Often that meant taking someone off from a big job with lots of people in their organization. Someone's managing 100,000 people to say, "Okay, here's something that's just starting out. You've got...you're it in the group, or you've got five people. And we want to build this into a $1 billion business. Now it's your turn to go do this."
In a lot of companies, that's kind of a deal breaker to say, "You've got this big operating job, now go to a small team and try to build up a business," which not likely, but has a decent chance of failure, because you're moving fast. And some of these worked out and some didn't. That would be a career breaker at some companies. But invent and simplify is an Amazon leadership principle. If you really want to be an inventive company, you also have to expect that some of the times you're going to fail, because if you don't have any failures in the company, you're probably not pushing the boundary moving fast enough.
So, for any initiative, we would just say, "Who's the senior most leader working on this and nothing but this? Do they have the organizational skills and capabilities to take this idea and make it a reality? And do they have the necessary resources to achieve their goal?" And what that last question is, is really, everyone doesn't have to report to that single-threaded leader, but have they been allocated resources, and have they removed enough dependencies so they have control over their own destiny?
So, unless you have a name, and a yes, and a yes to those three questions, we did not have single-threaded leadership. And that was really why, the root cause of why some of these projects just never came to fruition. So, we would appoint a single-threaded leader, make sure they had the right resources, make sure they had the right set of capabilities to go make this thing a reality. And more often than not, that would actually allow the project to eventually become a very large business.
The last thing, I'll end, there's a great quote by Dave Limp, who runs the devices business at Amazon, SVP. And he says, "The best way to fail at inventing something is to make it someone's part-time job." And new businesses are hard to get off the ground. And if it's number three or four on your priority list because you're already managing a big business, it's usually not going to get off the ground. So, you really want teams focusing on something that's worth doing and thinking about that full time every day as they get to work, "How can I improve the customer experience in this area," versus, "I've got a big job, I've got my day job, and I'm supposed to do this side project, which may or may not be...it can be big, but I don't really have time to get it done." That's what single-threaded leadership is about.
Sohrab: So, you started off... Thank you for that. You started off with the context. So, you first had the two pizza teams, which then became single-threaded teams, and then you moved into single-threaded leadership. Let me understand quickly. But the single-threaded teams were still in place. You just added the single-threaded leadership on top of that, is that correct?
Colin: Yeah, correct. And it's not an org structure. Sometimes those teams can be like a Spotify squad model. But the thing is that you have to dedicate those resources, and you let the...the team now has the authority, and the responsibility, and the latitude to make those tough calls. You don't have to go beg, borrow resources from other organizations, or ask permission to go do something. If you have the right team and the right leader in place, they're the ones who can help make those tough calls.
Like an example is do we use another service from either a third party, or an internal service within the company, or do we go build our own service to move fast and launch this initiative? And one example is with digital payments. So, there is a single-threaded team around as we were building Kindle. Well, for the first 10 or 15 years at Amazon, our payment system, when you bought something, we shipped you a product. So, the payment system assumed, "Well, I'm not going to charge until I know the address of where this product is going," and you charge when there's a shipment.
Well, with the digital product, there is no shipment, there is no address. And you have 15 years of assumptions built into this payment system, which is already doing a big job for Amazon, billions of dollars passing through this. And the small digital team says, "Well, can you just relax the constraint that payments don't need an address?" That's number 18 on their priority list. And so, you figure out, "Well, do we have to build our own payment system for digital?" Which is also a hard thing to do, and would require a lot of resources from a very small team.
So, you need a senior leader who can make that trade off. Can I use another third party service? Maybe I'll get 80% of what I need, but I'll be able to move faster. Or is this something that's so important that we have to build that on our own because it differentiates the service I'm trying to build? And there's not always one right path in there, but you need a high judgment individual leading that team who can make that right call.
Sohrab: Yeah. This is important. So, when I now look at a lot of the companies that I get to work with, I mean, the reason they bring us in is something's not going well and they want to move faster. They want to become more innovative. And there are so many out there where they, A, lack the concept of single-threaded teams. So, people are 30% on this project, 20% on another project, and so on and so forth. You've seen this, I'm sure. But what they lack even more is the concept of single-threaded leaders.
So, if you look, is there a senior leader who is fully dedicated to this project or this initiative, the answer is usually no. Or you cannot have one name. You sometimes have a steering committee. And in those committees, no one is basically really responsible for this thing. They try to make decisions collectively, especially if we go from Germany to Switzerland, which is a bit more consensus based. And things just move very, very slowly. And I try to always... And we will get back to this as part of the second part of this conversation. How do you really convince them to make that change?
Because this is a huge change for many of these organizations. And what I see as a challenge is, A, sometimes the organization doesn't want to appoint an individual. And sometimes nobody wants to take that responsibility. So, both are present in many organizations. And I'd like later to explore with you how we could probably navigate around that or challenge.
So, that was on organizing. And anyone who wants to dive deeper, I mean, there's a whole section in the book, and I highly recommend to read it. Let's move to communicating. Amazon has...if I read it correctly, they banned PowerPoints in the company to avoid having bullet points on slides, to avoid having beautiful pictures, to help people really think something through, write it down as a narrative, and then share that with others. And then we have the famous six pagers, PR FAQs, and so on and so forth.
The Six-Page Narrative
Sohrab: Can you walk me a bit through the history of this? And I think if I remember correctly, this was one of those ideas that Jeff also participated in, but other leaders helped build out. And what the implication has been of the six pager, or the narratives within Amazon as it grew over time?
Colin: Sure. And so, Amazon, they still do use some slides for things like an all-hands meeting, or a broadcasting one to many, or even working with partners. So, it started out not as a ban on PowerPoint or slides, it started out in 2004 as we had...I was working with Jeff Bezos as his chief of staff. And every Tuesday, we would have a four-hour meeting with the senior leaders of the company, the S-team. Think of it as Jeff's direct reports. And different teams would come in and present, either make a decision, give updates on a business, or figure out, like, "Hey, should we move into a new geography or category?"
And we were doing what every other company seemed, at the time, that teams would bring in a slide deck, and walk through the slides, rarely ever get through the deck because you have everyone's peppering with questions, some were then starting the next slide. Sometimes there just wasn't enough information in the deck to answer the question. And part of my job as chief of staff was to really help make Jeff a more effective CEO.
And Jeff and I were studying something from...it was Edward Tufte, a professor at Yale, who wrote a great essay called "The Cognitive Style of PowerPoint." And Edward Tufte proposes, "We'll use narratives, not slides." He gives examples of where you can have breakdown in communication. And so, it was Jeff who...it was one day in June of 2004, he said, "I'm going to ban slides for this weekly S-team meeting. And everyone you know from this point forward needs to come in with a written narrative. That's how we're going to evaluate, get information, and make decisions at S-team."
And so, one, it was...I talked about these processes, we didn't really create the Amazon operating system in theory, and then impose it through the company. It was we had a very specific problem we were trying to solve at the time. We were going to try something. Some of them didn't work, some of them changed. But they were meant to really address specific growing pains that we were having in here. At S-team, the stakes were getting higher. The dollars and the magnitude of the business, and the decisions were getting bigger. We were getting less information each week. And the meeting was still four hours. So, we needed to figure out a way how can we make higher quality decisions at this S-team meeting.
And so, Jeff's idea was, "Let's try narratives. The worst case, we could go back to using slides after a couple of weeks." So, it was a reversible decision. That was one thing. But no exceptions, starting next Tuesday, the teams have to come in. A couple of years later, Jeff would call switching to narratives one of the best moves that Amazon has ever made. He does not say those things lightly, I can guarantee you. And a couple of reasons why it really became a great competitive advantage for Amazon, I can tell you why narratives proved so powerful for Jeff staff meeting.
One example of a great process is it's scalable and repeatable, and it spreads through the company without you having to audit, and make sure everyone starts...is complying with the process. Once we tried it at the S-team meeting, other people realized, "Hey, this is a better way for me as a manager to help me make decisions. I'm going to start using narratives." So, the narratives just kind of spread out from there. There wasn't really a mandate other than that meeting to ban slides. It just spread through the company after months and...a couple of months where people realized that, "Yeah, this is just a better way of making decisions."
And the reason why Jeff called that one of the best moves Amazon made is narratives are harder to write than slides. And so, it really forces the presenting or the writing team to get their arguments straight and make a cohesive, well-written narrative. It takes a long time. But pushing that requiring, some worked on the presenting team when that does, is when you get to the meeting, you just have a much richer set of information to evaluate and make these tough decisions.
And so, narratives, what they did for Amazon, all in, we got about 20 times as much information within the same unit time to make the decisions. And so, the way a narrative meeting works at Amazon, let's just say it's a one-hour meeting. A team would come in with roughly a six-page narrative plus appendices. The reason we say...you heard the term six pager, that's just the length of a narrative. It's not a type of narrative. But we figured out the right information density. It would take about three minutes to read a page of an Amazon narrative. If you can read it quickly, it probably doesn't have enough information in it or.
And so, 3 minutes a page, that's 18 minutes. So, the first 20 minutes of the meeting is reading, and the next 40 minutes is silent reading and making comments, and the next 40 minutes is discussion. So, it's kind of strange if your first Amazon narrative meeting, you've prepared this document, and you're waiting for questions to come in, you're just looking. The next 20 minutes are silent where people have shared a document and they're entering questions and comments.
But if you were to look at it through another spectrum, you would just see a massive amount of information being transferred from the presenting team to the leaders in the room, which allows you then to have 40 minutes of data-driven high-quality discussion. So, you leave the meeting with either a much better idea or a higher quality decision. That is why it's so powerful for Amazon, is we just got so much more out of those 60 minutes, that 60-minute meeting, and we made higher quality decisions based on what we were doing before it was using slides.
Sohrab: Yeah, and it's interesting. If you think, a lot of leaders spend all of their eight hours, I mean, many work more than eight hours, but eight hours in meetings. So, 8 times 60 minutes, if each of those meetings on every single day, every single year is more effective, that adds up. And not just for one leader, but for every leader in the organization.
And you mentioned something and I tried this, that for people that are experiencing this the first time, it's very weird. And you're probably familiar, I'm sure you're familiar with Jeff's shareholder letters. And there is one particular interesting one. I mean, all of them are interesting, but one from 2015 where he also talks about leadership principles, where he talks about reversible and irreversible decisions. And I will get back to that. But I use that particular shareholder letter when I teach an Agile Leadership course. And this is usually the second day of our course. And it's right after lunch. And I hand it out, everyone sits silently and has to read it. And those shareholder letters are usually also something around six pages. So, it's very similar, and it's very silent, and everyone reads it. And then we have 40 minutes conversation around that.
And after that, I ask them, "So, how was this one hour for you?" They were like, "Oh, man, that was incredibly productive for a topic that we had no idea about. Where we sat down, it was complete news to all of us, and we could have a very deep conversation about it." I was like, "You just experienced how a meeting is run within Amazon." Many of them are like, "Okay, I need to replicate this in my company." Now, I don't know how many of them actually do it because, again, it's hard, as you pointed out, to build this habit of creating narratives instead of slides. But I think, and I've experienced it myself, it's very, very powerful to do this.
One-way door, or a two-way door type decision
Sohrab: Now, on decision making, you mentioned that this whole thing about rethinking how you communicate, and the narratives, came up as a specific problem on how to make higher quality decisions as part of the S-team meeting, and then later on in other meetings as well. Amazon does something remarkable, and I'd like to get your perspective on this. They distinguish between reversible and irreversible decisions. And based on that, based on that distinction, they also deal with this decisions in a different way. You gave an example of saying, "Hey, starting with narrative, sort of the six pager instead of slides, that's a reversible decision. So, we can just try it. If it doesn't work for two, three, four meetings, we just go back to how it was before."
With irreversible decisions, that's my understanding, you deliberately take more time. So, it's then not so much about the pace of the decision, but the quality. Whereas with reversible decisions, maybe pace is more important than quality, right? You always want to make a good decision, but pace, and then regular feedback cycles ultimately helps you to get to a good decision. Can you elaborate a bit about that distinction in decision making?
Colin: Yeah, so, we call them one-way doors or two-way doors. Two-way door, you go out the door, you see what's there. If you don't like it, you just turn around and go back. A one-way door, you go through the door, it's locked. It's very expensive to go back. And so, before we approached making a decision, we would ask, "Is it a one-way door, or a two-way door type decision?" And the reason you ask that right when you start out is because you have a different set of tools and processes for one-way doors and two-way doors. So, like a good example of a one-way door decision would be where do we want to build this next fulfillment center, or for city planning, or at a region like, "Where should the next airport go? Or do we expand a new runway?"
It's very expensive to...once you decide to make that decision, and you can't really go...you can go back, but you have to spend a lot of money. And so, to do that, or it takes a long time to reverse it. And so, you have to do a lot more research, contingency planning, scenario analysis, collect a lot more data. And a lot of large companies use that one-way door decision making for most of their decisions. What we found is that most of our decisions on Amazon were actually more like two-way door decisions. And that's where speed is important, and also the ability to course correct once you get more information. "Well, here's what we thought. Oh, it's a bit different. Let's course correct." And you'll get to where you're going quicker.
And so, for that, you're optimizing for speed and course correction versus the one-way door decisions, you're trying to minimize risk, minimize cost. You want to be a lot more certain about as we go through step one, go through this door, is it the right door to go through because it's going to be expensive to unwind.
So, just asking yourself, is it a one-way door or two-way door decision helps. And then saying, "What are the tools and processes we use, and how are they different for a one-way door versus a two-way door decision?" Because a lot of companies make two-way door decisions with one-way door processes. And that's where you get all of the cost, and the slow speed, and you don't get the benefit of being able to move fast and course correct. And so, just those simple questions up front really helped us move forward and choose the right path.
Sohrab: Yeah. And I think what you mentioned, right, most organizations use the one-way door decision making for almost everything. I think it's actually for everything. For a lot of things, they assume that it's just one-way doors. Now, my background, I don't know if we covered this in our first conversation, but my initial background is in medicine. And I always bring up this example, right? In medicine one-way door decisions would be if the patient is dead. So, you want to take your time deliberately, right? And figure out what is the best plan to go ahead. But there are also two-way door decisions.
And in business, many more decisions are two-way door decisions than in medicine. Actually, almost every decision in business is to some extent reversible. You mentioned it could be very expensive. It could take a lot of time. But there are so many two-way door decisions. And the course correction aspect of it is also tremendously important. So, I fully agree on this one.
Now, with regards to time, I think the section about working backwards and metrics, we can also cover as part two of our conversation. And I'd like to now make that segway towards we've gone through some of these elements of the Amazon operating system, as you phrased it, and you pointed out it was not built theoretically, and then implemented at once. Right? All of these things were created over time, always with the intention to fix a certain problem that you were facing as Amazon was growing incredibly fast.
Now, the past two years, and I think even before that, because when you left Amazon, you still worked at another company. If I remember correctly, that was sold to Alibaba or something like that. But then you started mainly working as an advisor, as a consultant for the past two, three, four years with organizations that want to bring in more innovation, more speed of decision making, maybe as part of that into their organizations.
And my question to you would be, how do we apply these lessons from Amazon, this advancement in management science from this one company? And I mean, there's many other companies out there that are doing incredibly interesting things, but let's focus on the Amazon principles. How do we apply them in other organizations? How do we actually help them take that first step towards becoming more innovative again?
How do we apply Amazon's Leadership principles in other organizations?
Colin: Sure. So, probably the biggest criteria for success is that the CEO of the organization really has to be behind trying to make this change. So, if we get someone from a company saying, "Hey, I'd like to go try narratives in my company, or use the working backwards process to get new ideas and decide whether to greenlight them." If the senior management is not bought into trying it, then it's probably not worth doing that at that point in time, because you can't...you have to...these processes conceptually, they're relatively straightforward to understand, but they do require discipline, and commitment, and you have to stick with them in order to get the value out of them.
So, we've noticed the single biggest success criteria is that there has to be a willingness to go try and implement some of these things from the CEO on down. So, we don't encourage every company to...we're not trying to sell that every company should act this way. But when we work with companies, we do ask and we talk to the CEO, "Are you bought into trying one or more of these things?"
The second thing is that you don't have to do all...there are six things we talk about, and then there's operating cadence. So, there's anywhere from 5 to 10 things you could do. We don't recommend doing all of them at once, so you can start with one or two of them. You can probably do about two at a time, and start implementing them over time. So, that's another thing. It's not you have to buy into the whole system and try it all at once. That's too much of...especially for any decent sized organization. That's too big of a change to happen concurrently at one time.
So, those are the two things. Pick what problems are you trying to solve as an organization? What are your biggest pain points? And start there. And then the second thing is make sure that CEO and the leader of the organization is committed to trying it, and making it work.
Sohrab: So, going back to the concept of single-threaded leadership, if you can't put the name of the CEO on top of this initiative, probably it's not worth starting this. Now, you mentioned there are six things, and then a few other things around operating cadence that you want organizations to address. And that they shouldn't do all six at once, because otherwise the change would be just too much. What are those six things
Colin: So, we talk about the leadership principles, getting those, hiring. There's an Amazon Bar Raiser process, and how do you attract and retain people who embody and reinforce your leadership principles? There's single threaded leadership. There's using and writing narratives in your company. There's the working backwards process, which is the name of the book, but it's a specific name process at Amazon on how you take new ideas, refine them, uncover the truth about them, and then decide whether you want to devote resources to go build that...turn that idea into initiative. And then the last one is metrics. How you measure things, input and output metrics. And different companies...
Sohrab: So, it's basically the six things that are covered in part one of the book.
Sohrab: A few of those we just discussed. I had that assumption because it was, again, six things. I just wanted to be sure that I'm not assuming the wrong thing. Cool.
Colin: And so, different organizations, younger organizations, the first place that we look at when we're working with companies, often people will say, "Well, it's not like it was a year ago. Our company is fundamentally different." And so, we ask a couple questions, one, "Show me your leadership principles or core values." And if the founders have not written them down, and codified their thinking about what has turned their idea into such a success, that's the first place you have to start. And then the second question is, "Tell me about your hiring process, and how do you ensure or maximize your chances that the people that you're bringing on board will embody and reinforce these leadership principles or core values."
And if you don't have a deliberate process to attract people who really comport with who you are and how you make decisions, that's how your company changes. If you grow from 20 to 50 people, or 20 to 100 people, over the course of a year, most of your company is going to be new a couple of months later. And unless you're really careful about bringing people on who agree with who you are, and how you make decisions, that's how your company is different a year. It's not like it was. It's because you attracted people who brought their own culture, and their own values from their prior organizations.
And you get a culture whether you like it or not. It's just do you want it to be a deliberate culture, and cultivate that? Do you want it to be the one that you want? And so, that's what the Bar Raiser hiring process does, is it attracts people. You map candidates on to your leadership principles and core values. That is a big part of the Bar Raiser process. So, those are important for younger and growing organizations, and a lot of founder-led organizations to make that transition.
And then these other processes come into play about as you're growing, how do you scale and still move fast, and innovate, and push the envelope while you're getting to be a bigger company, series C, series D. And we work with public companies, too. How can you remain nimble, and fast, and true to your roots? Jeff in one of the shareholder letters says, "We wanted Amazon to be an invention machine where we could take advantage of our scale, but also have that nimbleness of an entrepreneurial startup." And then he went on and talked about something else.
We just tried to define what that invention machine was in our book and in our practice. And there are some things that companies can actually pick up. You don't have to be a $100 billion multinational company to do it. You can be anything from a small company up through...you can be a $100 million plus company to implement some of these things, but they're nice, scalable, and fractal that they work for small organizations, large B2B, B2C, nonprofit, government organizations, too. So, we found them to be pretty darn near universal, and they scale up well, too.
Sohrab: Yeah. Now, you mentioned some concrete things that you recommend to smaller companies, or companies that are going through the different series of investment. And many of those companies are, at least to some extent, founder-led. And I can imagine being a founder myself that a lot of the messages that you share, that Jeff shares in his shareholder letters, that they are well received by these founders. And these companies being nimble, already doing innovative stuff, can apply many of the things, put them...weave them into their processes, or build their processes around these things. And through that become more successful, or in a systematic way, structure their organization going forward.
Now, what I believe is a bigger challenge is if we take companies that are, for example, 150 years old. And we have some of those in the U.S., we have many of those in Europe. A lot of old companies, the founder's not there, long dead. Even his kids, grandkids. Nobody's in the company. They're publicly listed. Maybe there's not even one major shareholder who calls the shots. There's a board of directors. And there are hired CEOs who usually have three, four, maybe five years as CEO in those positions. How do you help them? And maybe the answer is there is no way to help them.
But how do you help them to apply many of these principles that you shared with me today? And help these organizations with their cultures that are very difficult to shape based on the longevity over which they were created or manifested? How? What would be your recommendation? Because in some cases, I work with them and I'm like, "Oh, if I only knew what to do next."
Colin: Well, first of all, I have a ton of respect for companies that have been around for 150 years. They do have to change over time. You may not seem like it, but 150-year-old company has to do a lot of things right to stick around for 150 years. You take a look at the top universities, for instance, that list is relatively static over the decades. But you take a look at the S&P 500, the Dow Jones, the FTSE, whatever that exchange, and you look at who are the top 20 companies from 100 years ago to today, it's quite a different list.
So, people that can stick around, they are doing something right. So, one is recognizing what has made them special, and you don't really want...often don't really want to mess with that or change that. So, that is one thing to ask about, "Well, what are they doing and what don't we want to change?"
But it comes back to what I alluded to earlier, when we work with some of these companies, is what problems are you really trying to solve, because if they want to move faster, or innovate, or inject new ideas into the company, there are processes for that. So, really identify...being very clear on what's the issue, or problem, or the challenge that we're trying to solve. And then here's the solution. And then just agreeing to try something and stick with it.
So, for instance, when we switched to narratives at Amazon, I probably read at least the first 100 narratives that were written at Amazon, and presented to Jeff Bezos and S-team. First few of them weren't all that good, especially compared to where they are now. But it was something that we tried with, and we were going to stick with. A lot of the first narratives were, "Let's just take our PowerPoint presentation, and put bullet points in paragraph form." Well, that's not really a narrative, but it took a while to know this is what we're looking for with clarity of thought.
It also took a while too, for the people reading the narrative, there's a huge job on them because the second...the part of the meeting, the 40 minutes is really where the leaders who read and comment on the narrative, they're helping the whole group leave with a higher quality decision, or a better idea than entered the room. There is a huge job on the people reading that and commenting on the document. So, to do that, it just takes some practice.
So, really, it's being clear on the problem, and then saying, put a stake in the ground, "We are going to try this. It may be different. We're going to learn some things, and we'll adjust as we learn." And that, more often than not, and then you have to have the change management in place for larger organizations, works pretty well. And you don't have to do it all company wide either. You can say, "We're going to try it in this specific group or division, and we'll learn about how to do this. And then, if it works, we'll roll it out to the rest of the organization."
So, those are a couple of tactical things that you can do for companies, that we've done for companies that have been around a long time. But Amazon hasn't been around 150 years. But 150 years, if Amazon is still around, there will be some things that they learned, and changed, and adapted to. There'll be some things that were remarkably consistent over that period of time.
Sohrab: Yeah. So, I fully agree with you. The respect that a lot of these companies need to get, right? Because it's always easy to go and say, "You are doing this wrong. You're doing this wrong. You're doing..." You have done a lot right to be there over a period of 150 years, 100 years, even if it's 30 years, or 20 years. Many companies die, and don't even scale to a certain size if they're not doing something right.
But at the same time, what I see is that a lot of organizations that used to have great business models, let's take the automotive industry where Germany is particularly strong in, with the regular internal combustion engine. This business model is dying, right? So, they need to figure out a way on how to become more innovative again in the space of EVs. And they're not only competing with Tesla who has become incredibly strong due to their market cap, and all of the cash that they're also now able to make, but they're also competing with Amazon who is in the EV space. They're also competing probably at some point with Apple. They're definitely competing with Google. Organizations that have deep pockets.
Thirty years ago, Amazon didn't have deep pockets. I mean, 25 years ago, they didn't have deep pockets. But now they do have deep pockets, and they are this invention machine. And none of these organizations, if I look at Amazon, Apple, Google. Apple, maybe even more so than the others, but Amazon certainly does not look at industry boundaries. You mentioned at the very beginning you started in 1998 selling books online only in the U.S.
That's not what this company is doing today. Right? And it's not only the everything store. If you look at AWS, it's a completely different business. So, these organizations being these invention machines, having this huge amount of capital accessible to them, they can go into every industry. And I think this means, even if an organization has been around for 150 years, they probably need to change faster than they currently are.
And for that, I'd be really interested. You mentioned a few things, and I agree with everything that you mentioned. Not doing it company wide, starting small, identify what is the problem that we actually want to solve, and then do something around that. Which is completely contrary to what the big management consulting firms try to sell. They always want to do company-wide transformations. They always have a blueprint that fixes every problem. That's not going to work. I think you and I are aligned on this.
But I do believe that probably the way leaders show up, the way leadership happens, and number one on your list is leadership principles, needs to significantly change so that these organizations can change. What is your experience around that, and what are maybe the one, two, three key things that leaders have to do differently? If we look at the big companies, the ones that are not founder-led, the ones that have been around for a long time, the ones that are not that innovative currently, but have to become more innovative fairly quickly soon.
Colin: Well, I mean, it's a big question. I can try to answer a couple of those things. But every company is going to get disrupted at some point. And if you have a line of business, and you're at a company, you have a choice about is this current business going to get disrupted by another company, or will we disrupt ourselves by starting up something new that is a better solution to a customer problem than our current business?
Now, the thing is, you don't know the time frame over when that disruption will happen. You don't know if it's going to happen next year, or 5 years, or 10 years from now. But most businesses, it will happen. And so, you have a conscious decision to make about are we going to disrupt ourselves? Are we going to let someone disrupt us?
And this is where a single-threaded leadership can help. You don't really want the person in charge of the big business that's going to get disrupted to also be in charge of disrupting themselves, because there's just this inherent conflict of interest. There's also an aversion of risk. You don't want to do anything that messes with the cash cow of the company that's paying for the bills.
And so, time and time again at Amazon, if we had an idea that could disrupt the business, we would have a single-threaded leader who could move fast, and experiment, and try things. And essentially, their job is to disrupt this other big business at Amazon. We did that with digital. At one point, 75% of Amazon's business was shipment of physical books, CDs, and DVDs, because that's how you read, listened to music, and watched movies. That was 75% of Amazon's business. And we knew this digital media was going to disrupt a lot of that business.
We didn't know whether it'd be...in what order, and also how long it would take. But we did not... And a lot of the suppliers are the same. The publishers, the record labels, the movie studios, they make those same movies just in different formats. So, you could argue that, "Well, we have all these relationships. Let's let the person running the big media business, physical media business for Amazon also figure out digital." We did not do that. We actually took the leader of that business, Steve Kessell, and said, "Instead of managing 75% of Amazon's business, you now have a team of 5 people, and you have to go figure out what to do with digital. And your job is to essentially put this big line of business, which pays all of our bills, out of business."
Took a long time to do, but it would not have been successful. And if you take a look at the companies who've tried to make that transformation from physical to digital, there are more failures than successes. And the reason why I think a lot of that happened is people were not making the bold moves, and making the right organizational moves to say, "We are going to disrupt ourselves, and we're going to invest the capital, and do what it takes to do that." And that's one thing about how you can be more innovative.
Colin: The second thing, and this ties into an Amazon leadership principle, is customer obsession. So, instead of figuring out, "Well, here's what we're good at. So, here's what we are going to do next. Here's an idea we have, and we're going to try to put on the market." Amazon really inverted that, starting with customer obsession, with the working backwards process. So, I think one of the reasons why Amazon has been able to get into all these businesses, they don't really ask, "What are we going to build next?"
They ask, "What is a big customer problem that we can go try to solve? Are we one of the few who can solve this, and be the best in the world at it? Whether it takes a month, or a year, or seven years, that doesn't..." Long-term thinking is another part of Amazon principle that doesn't scary Amazon, but asking, "What's the next customer problem that we're going to try to solve," versus, "What's the next idea we want to go build or do next."
One's a company-led, skills-forward approach, and this is a customer-obsessed customer-led process. That's how you go down into paths, surprising paths that unlock a lot of value. And it looks like innovation from the outside in. But really, what you're doing is you're letting the customer lead you to solutions, and then you decide whether you want to invest and capitalize on that.
But that's one of the reasons why Amazon has a large portfolio of very different businesses. It's because we looked at, "Here are big customer problems, and we think we have a pretty good idea. And we think if we don't have the skills, we can acquire them somehow, or we have the skills and we can go move into this area." But it's a customer-led approach. So, those are a couple of things if you're looking at, "Well, how do you get companies that really need to innovate faster and change, or you have been stuck in a rut?" Those are a few things that you can...it's a different mindset to try to get there.
Sohrab: Yeah, I think both are super important, especially working with industrial companies. Like, industrial goods and services. In Germany, we have chemicals, we have all of this. They're really not used to looking at it from a customer perspective. It's always a, "Oh, I found this new chemical and we can build this, and then maybe we can find a business model around it." Instead of really doing that shift in your mindset and starting out, "What is the problem that a customer is facing, and how can we, with our capabilities, or new capabilities that we buy or build, basically solve the problem?"
The other is, and you mentioned it as a side note, you took the leader of 75% of the company's revenues, and put him on a 5-person team to build something completely new. In most organizations, that would have been career suicide. Because chances are, and you mentioned this as well as part of our conversation today, that the new business is not going to work, or it's not going to work in that period of time. Or maybe, despite having all the relationships to the publishers, the movie studios, etc., maybe Steve didn't have the necessary skills to lead a fully digital business. I don't know. Right?
But there are certain risks associated, especially as a senior leader moving from such an important position to a position which is unknown, and could become very important. How do you create, or how was the environment created that someone like Steve said, "Yes, I'm going to do this."
Colin: Well, it comes down to the leadership principles, invent and simplify. If you want to invent, you have to be...if you want to place bets, you want them to be bets that you think are going to work. But some of them will fail, so it doesn't really guarantee success. Two more examples. Andy Jassy, he was Jeff Bezos' chief of staff. He could have had any job he wanted in the company. He went to this risky area about web services, "Maybe there's something here. Maybe we could build something," which eventually turned into cloud computing and AWS. Unless you had someone like Andy leading that, it wouldn't be what it is today.
I mean, I can name countless examples of where Amazon leaders have gone on to... You want to put your leaders on your biggest opportunities, and that may be different from your biggest business. It may be the same, by the way, but if it's different, you want to...your management bandwidth is one of your constraints. And so, you put your top leaders on your biggest opportunities, independent of what the size of the business is today.
Another thing how it worked at Amazon is with compensation. Some industries, like the finance industry on the East Coast in the U.S., or movie studios on the West Coast, a lot of their compensation is tied to bonuses for annual events, or short-term events. How did this portfolio of movies do, or am I going get my year-end bonus? At Amazon, it's long-term compensation, and the currency is the same. The head of AWS, and the head of Amazon's retail business, they get paid in the same currency and get paid over time. So, when you move around in Amazon, it's not like you're missing out on a big bonus, "I need to stick around until I hit my bonus in this area."
So, all these things have to fit together in your organization, because you can't move top leaders around, you can't pay people to do one thing, and then ask them to do another, because typically do what you're paying them for. And so, if you have a long-term compensation where your reward is growing the long-term value of the company, then you're comfortable moving around in these things.
And Fire Phone was an example of a thing that failed. But you have to... When that failed, Jeff said, "I hope we're working on even bigger failures because our company's growing. We need to place these types of bets. Some of them are going to work, and some of them are not." But if it's a well thought-through experiment, that's okay. So, those are a couple of things how Amazon was able to do those things. But it really ties back to its leadership principles of ownership, invent and simplify, customer obsession. Those are three areas that really help people move around.
And the last thing I'll add is it's not always...it's hard sometimes to say, "Well, why do I have to give up my big job and move to this?" And so, you have to explain to the leader, "Well, here's why." And it's not always a bed of roses to do that. But if you want to be successful over the long term, and create these disrupting ideas, you do have to go down that path. And this is why it's an easy concept. It requires discipline to stick with. So, I can't say it's easy. What I explain is easy to understand, but it's hard to do in practice.
Sohrab: I mean, I mentioned earlier I'm from the medical world initially, right? Being healthy is an easy concept. Eat well, sleep well, work out well. But we all know it's hard to implement, especially over a long period of time. And I think the same applies to here "Working Backwards: Insights, Stories, and Secrets from Inside Amazon."
Colin, we talked about constraints earlier or you talked about constraints. We're hitting the end of our time box. And I want to use this opportunity to, again, thank you so much for joining me now for a second time, walking me and others who will view or watch this video at a later point in time through all of your experience. And I truly hope that our paths will cross again. I have so many more questions that I want to ask you. But thank you so much for being here.
Colin: Well, thanks again for having me. Great questions. And thanks again.