Dude, you've got, you look so good, man. You too. Yeah. You haven't changed at all? For everybody's listening, this is what all guys in their mid 50 say to each other when they haven't seen each other in a little You look so good. It's awesome. So we do talk and you talk about heuristics and mindsets and mental models with great leaders. And you're a member of this s team at Amazon.
Can you tell us what the s team was and then maybe give us a sense of some of the main things that you learn being part of that and some of the mental models that you were using to achieve this level of excellence. Well, the s team is just the senior team. It started out as the direct reports to Jeff. So the longest serving members of the s team when I departed in February Pete me and Jeff you know, because I was basically hired as a direct report.
I worked for Joe for that year, and then the president layer was kinda gone. Was what you would imagine it to have been in the early days. It was the operating committee or team that ran the company.
So, you know, head of tech, head of HR, head of legal, head of ops, head of Pete, And, you know, when you go into the mid 2000s and we start to invest heavily in AWS and in the devices business, my team led by Bill Carr, Bill Prime Video, and then I asked Jeff Blackburn to take it over because we got to a place where most of the work was identifying, you know, scripts that were gonna be winners, and I just knew that wasn't gonna be me. And I thought Jeff would be a terrific honor. He was.
He and the team built that business amazingly. But when we got to a more diversified set of businesses, the role of the s team started to change. It was less about operating the individual businesses and more about deciding on the mechanisms, and we have a spell. Pete to it in a second. We have a very specific definition of mechanisms, but it was about What should we be doing together to make sure that this culture thrives that we stay day 1 versus day 2?
You know, meaning that it feels like a startup even if big and that we continue to attract and retain the very best people in the world. And we did things together like build the leadership principles I was really adamant that we should have a single set of principles that define what it means to be a leader at Amazon. And we're all leaders.
We talked about should the the only for people that have direct reports, or should they be for every Amazonian, and we decided they really should be for every employee of the company, and they start with customer obsession, and it ends with deliver result. They're public. So, you know, anybody can take a look at them, but the s team was the steward of the leadership principles, maybe 1st and foremost, it ran a couple of mechanisms. Who who wrote them?
Did you write them back in 99, or did you wait until 1002 when you were firmly in Scotts there? Or 2002, we started there was a team led by a woman named Robin Menziger who was in HR and later, my George, who was an HR leader too, worked with me and Rick Beller, who was our CTO, mostly because I think Pete sort of the most passionate members of the s team about this.
Rick was a West Flint Grad and understood leadership in a way that, you know, that few people do, and and we saw eye to eye on a lot of things that were leadership related. So You're right. He was from West Flint, and you were from Tiger Inn, and so there was a lot of Right. Just kidding. Right. It was perfect. It was perfect. Yeah. Right.
And we both played 150 football So, you know, we were on opposite sides of the line a few years apart, but, you know, we thought it was important to codify what made this place special. And there's a lot of Jeff in the language, the first draft, you know, clearly there wasn't an it wasn't good enough, and I think Jeff was right to push back on a bunch of things. And through that iteration, we ended up with a set of Flint that were pretty enduring.
We made maybe 2 or 3 other changes in their 18 year life, including adding learn and be curious. A few years ago, but they really did stand the test of time and they proved to be a inoculation against a drifting culture. Yeah. So a lot of companies do this. Right? And I know with all of our companies, we actually run them through a program to actually develop these things. Very few of them make the impact that yours did.
Is there something about what you said or about who you were, about how you said it that really allowed it to come to life? Well, I think one of the things we did write was we use the words in the leadership principles everywhere. I mean, we use them when we communicate each other. We use them when we write documents and Amazon's as a writing culture that meetings start in silence as you read the document. Mhmm. We don't use PowerPoint.
And, you know, I'm saying we all the time, by the way, that's gonna be a hard habit to kick, but Yeah. Yeah. That's forgive me. We understand. But, you know, these words were pervasive. They inform how you give feedback to people. They inform how you set your goals and how you measure against those goals. They inform how we talk to the investment community and customers.
And one of the things that the little tricks that I was lucky enough to keep persistently in our focus was not letting the words get diluted with more explanations. So what I mean by that is the leadership team will often craft a single page of principles and values. And if they've done their work well, you could argue that those words are the only words that a new person needs to understand the culture.
Mhmm. And if you need more words than that, in other words, if you end up having the urge to create, you know, Wiki pages to define and further elaborate on each of the principles, you may not have written them crisply enough to start with.
So I kind of made it my mission over the years to seek out diluted, and I would argue much less effective explanations of very crisp principles and try to remove them from the corpus so that when, you know, you're forced to teach somebody the leadership principles, you're forced to go to the primary document, have the person read it, and think deeply about it, and then ask them, what do these mean to you? How do you think you'd employ them in your work?
And that, it produces very different results than I would argue less Beller written prose that's too pervasive and too copious. Too many committee, too many hands on it, too many compromises. And so I'm actually hearing that you would go around and move language around in document. You would talk to people about their language. Yeah. For sure. And you were the enforcer, and other people on the senior team were enforcers of this.
They agreed with you that we needed to enforce a way of speaking in a way of communicating because that ultimately produces a result. In human cultures and a businesses, create cultures, the oral tradition matters a lot. Spoken word is very powerful. And, you know, I think we sometimes fool ourselves into thinking that the artifacts that we create that are in PowerPoint or in Pros or in binders sitting on a shelf really define the models people have of the company and the way people interact.
And my experience has been a little different. I think that there's the way you wish it was, and then there's the way it is. And if you can make those 2 be as close to each other as possible, you're gonna succeed. But it it does say cleaning and effort all the time over years, every weeks, every consult. Constantly.
Yeah. It was one of the most important jobs that I and others had, I think, was And I gotta dig on this because as new people come into the company and they move up into position of leadership, they want to put their own stamp on things. And Pete, you're saying, no, don't move that over there. It's right here. We need to keep it right here. I can't let you do that. Was that disappointing to people?
Well, we have a little phrase that we would put in parentheses around things that were slow changing like this. And it says, unless you know Beller. And so these are our leadership principles, unless you know better ones. And from time to time, people would propose better ones, but it's not enough to just say, well, these are better because we're gonna ask, why? Why are they better? And how can you demonstrate that they're better for us in this environment.
And we did revise those principles 2 or 3 times over 18 years. And it was after you know, some really careful, thoughtful pushing by people that thought we were missing certain things. But I think one of problems that firms have is that when you bring something in from the outside, you know, you can bring the process and the understanding and the science and the data And you can also bring a bunch of stuff that's a little less crisp.
And I'd argue try to separate the 2, you know, get to the real root of insight and leverage and process and discard the stuff that, you know, that's sort of window dressing. Maybe the cultural stuff or the assumptions or Yeah. Cultural stuff's important, but, you know, the way humans describe culture is a very complicated mixture of precision and imprecision in my experience. And it's the imprecision that leads to a less effective culture, I think.
Doesn't mean to say that the leisure principles need to be prescriptive. I just mean that the words that you choose need to be carefully considered and very precise. Right. I remember one of the words that stood out to me from the leadership principles says this is one of the things that makes Amazon peculiar. Right. That word peculiar is not a common word. Right. And on FX, say, you know, if you're not weird, you're weird.
It seems like you guys were leaning into this peculiarity, and you weren't leaning into difference. You were leaning into peculiarity, and there might be synonyms in the source, but the choice of the word peculiar was peculiar. Yeah. It really is celebrating diversity of thought.
And, you know, one of my favorite clauses in the leadership principles says leaders don't believe their body odor smells of perfume that you could imagine some Currier ways to say this, but I like that clause because it's a reminder that we were sort of a young startup. Like, I don't see a dodgy day 2 company saying, you know, we don't believe our whatever stink. That's right. And at the time, you're, like, 3233 when you're joining and Jeff Bezos is 34, 35, somewhere in there.
So you guys were doing it. Did you think that startup founders or great leaders have to be peculiar in some way? I don't think they have to be. Well, look, I think all humans are unique in some way. And so if that's what peculiar means, then I think everybody has And as we go back to my mentor, Andre Tremper said to me, you know, don't hide all of your peculiarities. You know, you will make more authentic connections with if you have the courage to reveal you. Mhmm. Really good.
Hey. You had made a comment about mechanisms, and you said you'd come back to it. What are the mechanisms you were thinking of? Well, it's a more precise definition of people use mechanisms as substitutes for process all the time or as a synonym and tended to think a little bit more about it at Amazon, a mechanism is a tool or a process where you achieve adoption. So you actually do the work to make sure that people understand the tool or met or process.
And then most importantly, you periodically inspect to make sure that the tool or process is being used as intended or can't be improved. And then if you find that it can be improved, then you improve it, get adoption for the new version, and so on. And this seems simple, but what I find over the course of decades is that most tools or processes that companies implement eventually go off the rails with a couple of failure modes.
And the funny ones are you go back to something 10 years later, and people are still doing it, even though the reason to do it is long gone. What may be worse or maybe just a different failure mode is you go back and there's a real need for the process, but it has mutated into something that adds no real value, but yet again, people are doing it very diligently. And, you know, and you can't blame the people that whose job it is to do the process.
You actually have to inspect the process and ask Is this optimal? Is it working? You know, is the data still as accurate as I thought it was? You know, one of the things we would always say is uninspected data is always wrong. And this sort of obsession with inspecting process really led to some of the operational excellence that was vital for things like prime. Got it. So talk to me about Prime and, you know, you guys kept inventing even though you were getting bigger and bigger.
When you got there, how many people were there. When I got to Amazon, there were roughly a 1000. And now how many other? Over a million. So While you're growing, while you're getting all this girth, which normally syncs all processes and all modes of thinking, frankly, you guys are inventing AWS. You're inventing, Alexa. You're inventing Prime, and you were involved with all of those things. Unbelievable track record, man. Unbelievable.
So how do you invent when you've also got all these people and all these systems in the process in place? Yeah. I mean, there are a bunch of you know, tricks, none of which are overly complicated, but the hard part is to stay focused on them over the course of years. It's one thing to implement something new, but I've met too many CEOs who, you know, they read a book. They listen to a podcast. They have an idea. It's great. They implement it. And then they're off to the new thing.
And they don't really come back in a couple years and ask the question. Is that insight we had a few years ago still adding value in the way that it was? That's hard thing to do when you're get your energy from creativity, but it's essential for creativity to actually compound value over time. Otherwise, you just end up with a pile of, you know, what was once creative that is now, you know, just weight.
And so the things that allowed us, I think, to scale without that weight included careful Morgan. So the idea that as often as possible, you when you put someone in charge, you make them single threaded. So when Jeff wanted to start AWS, he made Andy Jassy single threaded. Andy didn't work at anything else other than the business plan Morgan. And when we started devices, Steve Kessel, who was running the books business at the time, moved over and took over with no other responsibilities.
This nascent business went from running the largest business to running, you know, having an assistant and, you know, and nothing else. And so that's first thing. Single threaded leadership. 2nd, I would argue that and this is easier now with cloud services.
But at the time, being disciplined about software architecture was vital because, you know, you can slow down and end up with organization following a bad architecture if you're not careful or worse the architecture then following a bad Morgan. And we had a lot of work to do to create a service oriented architecture out of a more legacy architecture. That may have been actually the hardest management problem in those middle years.
We measured everything, and we focused on the inputs, not the outputs. James rarely worried about, you know, quarterly revenue. This is a classic example. When I took over the retail team, I noticed I printed out the inventory over time, and I noticed that we had a trough every March 30th or 31st, June 30th, September 30th December 31st. That's when our inventory was lowest. Well, why would it be at the end of every quarter? Why is that the optimal time for low inventory?
And then I looked at sales it turned out we were seeing small but perceptible sales dips in the 1st couple of weeks of every quarter and actually the last week of every quarter. And I'm like, Beller, clearly what's happening is humans are worried about a target for inventory turns, and they're slowing down buying so that they can hit the inventory. And then they ramp it back up once the quarter starts again.
And, you know, if you see that kind of a metric and it's got a period of a quarter in business, something's wrong. Right. Because this should be based on what the consumers are doing, not based on what your internal team is doing. The math should optimize how much inventory you have. So I said, we're not gonna do this anymore. We're gonna find out what the optimal is from the math, and then we're gonna leave it there. I told the investment community we're gonna do that. They thought it was nuts.
They were like, well, then you're not gonna have said, no. No. No. We're gonna have discipline, but we're gonna have discipline 365 days a year, not 4 days a year. And it turned out sales went up. Inventory turns went up, but it took, you know, a year or so, and we had to have the courage to do that.
So I got onto that by talking about focusing on inputs, not outputs, but if you focus on inputs, you're less likely to worry about the inter organizational output stuff that a lot of teams fight about. And there's a host of other things that I think Amazon did Beller. Highlight 1 more, which is I think when you choose to invest in so many different businesses, you know, many consultants would tell you that's crazy. You should pick a core competence and stick to it.
You have to make sure you protect the nascent businesses because the mothership is gonna wanna destroy them. And I think Jeff was brilliant at this. I think he carefully protected Andy and AWS. He carefully protected Steve and then Dave Lamp and devices. Even though it may not have felt that way if you were Andy's, you know, Steve or Dave. Yeah. Right.
But he made sure the 3rd party business that we built where, you know, we invited other retailers to sell against our retail team in the same store. That's nuts. And the only way that that sticks is if you keep the retail business from crushing the third party business, which Jeff did masterfully. We made sure that 3rd party, sellers were treated as customers from the beginning. So, look, I'm hearing such great perspective that you kept pulling back and saying, is the language right?
Pulling back and saying, is the process right? Pulling back and saying, is our architect our software, right, so that our company organization structure is, right, kept pulling back. And that's amazing, and that's a mental trick. But then you've gotta actually get the people who you're working with to accept the changes you wanna make to put things back in the right spot. So what you said was it took courage to do this.
Yes. So where do you create the courage for the people the way that you let them know when they failed that you're gonna separate the problem from the Pete. And if they're failing because you know, they're gonna be unsuccessful that you'll help them to move on to something where they can succeed. But if they're failing because of things that aren't within their control, that you're gonna work with them to remove those barriers.
So, you know, I think a lot of it is setting up a relationship with people that lets them know that you're gonna hold them to high standards and simultaneous support them if they fail. And I will say this about Larry Bosco. I learned this from Larry. Larry was the best person that I've ever worked with at making you terrified of performing poorly for him, but also even if you had failed and he was mad coming back for more.
And I have this great example where I was it was first business I was running. It was my first quarterly review with the CEO with Larry. And, you know, I was nerve about it. Business was doing okay, but not great. And my assistant comes to the office, and she's a little frantic. She says, Jeff, Larry's in the phone. Pick up the phone. So I pick up the phone as Larry It's Larry. Hey, Larry. What can I do for you? You said it looks like we have a review coming up tomorrow.
We're reviewing all of the chemical businesses. Big day for you. I went through the book. I've looked ahead. And, you know, I want you to know that I think the world of you, I think you're gonna be a fantastic leader. You're gonna have a great career, but tomorrow is gonna be a rough day and hangs up. Like, I didn't sleep. I was, you know, way overprepared. I, you know, got through the day fine, but I wanted to come back for more.
He was just great at setting a high standard and letting you know that he cared about you. Very interesting. So that's just a lot of time, a lot of personal relation ships a lot of being authentic with people and being there for them. Yeah. And the hard part is as, you know, I think there are these moments and you hinted at this earlier where cultures change and task of leadership changes.
And I would argue that if n equals the log base 10 of the number of Pete, that there are these distinct phases of leadership n equals you know, 1, 2, 3, 4, 5, and now Amazon at 6. You know, the one that you just described works best from 1 to 2. So when you have, you know, a hundred people or less, you can know everybody. You can manage by walking around. You certainly know your ten you know, most important people really well if you do this right.
But when you get to n equals 3 and 4, you're depending on others to create that same kind of environment where people will take risk and learn from those failures. And it's a harder thing to do. It requires some different kinds of mechanisms some different kinds of projection at nqual6, at actually 56, I found that I had to do a lot more video work for example, to project and to try to be authentic on video.
In fact, I had an early interview some number of years ago where I was sort of playing around with different video techniques. And I'm essentially memorized the script. I had an hour long interview as you often do in a fireside chat. You know, the questions in advance I had written out all the answers, and I memorized them like an actor.
And I delivered the talk, you know, because it was essentially a soliqua with the interruptions of the questions And when I was done, I showed the tape to a couple of people because I was super proud. You know, I nailed the dismount. It's great. I didn't miss a word. I was super proud of myself, and I showed it to a few Pete. And I said, what do you think of they said, you know, it's good. You know your lines.
You've said all the right stuff, but they said there was something missing that I've seen you do on stage or in video. That that was just missing for me. And I was like, what is it? What is it? And they said it was just authenticity that it was so carefully crafted. What they said is even if the words come out a little bit imperfectly even if you insert like I just did parentheticals, which would made written prose hard to read.
There's something about you know, your authenticity in an extemporaneous environment that's missing from this video. So the next one, about James length, it knew the questions, You know, I spent time prepping. I thought through them. I had practiced with somebody, but then when I got on stage, I just riffed. And I showed that video to the same person there, like, You should do this every time.
Yeah. You know, it turns out in the prose, the transcript, I was being a perfectionist because the English prose wasn't as perfect as I would like to write. But the impact on an audience that viewed it was overwhelmingly better. Right. What's what's that they say? They don't remember what you said. They remember how you made them feel.
Exactly. And that matters much Morgan, and, you know, 4, 5, and 6 than it does at 1 or 2 because at 1 or 2, you know, if you're recording everything instead of just being with your team, know, they're probably missing out. Right. The medium is the message there where that's kind of inauthentic, even if you're being authentic on video. But later on, when they know you can't be with them, to be authentic on the video is a good substitute.
Bezos and others have described you as an incredible teacher to everyone at Amazon. What do you think makes someone a great teacher? Well, I think competence matters. It's hard to teach something that you don't thoroughly understand. Right? And so when I go back to what we talked about up front where early career I was collecting experiences so that I could maximize option value.
In effect, I was achieving confidence in a bunch of areas that made it easier for me to teach, you have to be a lifelong learner to teach. So, you know, teaching isn't just getting to a certain point and then using whatever you've learned without any new learning. I think it's a iterative process. So, you know, the best teachers are also often the best learners. I think communication style and ability to relate to people in an authentic way helps you be a, you know, a better teacher.
And then finding mechanisms to project teaching broadly as an organization grows is Morgan. And that's something that I think I watch especially in this era of zoom. I watch great teachers do zoom and Coursera and, you know, that world where you have some teachers able to project to much broader audiences more effectively than others.
You know, it's a certain kind of skill to be able to use your voice and your face to project that what you're saying is you believe it and that it's coming from an authentic place. It's if you don't really believe it, and you're trying to do this on video, it's gonna be super hard for others to believe it to. And, you know, part of teaching, part of learning is believing the teacher. You can be skeptical, but you sort of have to believe that the teacher has confidence.
And so in a way, what makes you a great teacher is that you care about people and you're a great learner. I hope so. And then I hope I spent the time to treat teaching as a task of paramount importance. Mhmm. That's right. So As opposed to something that you do reluctantly. Got it. Helping fix those processes, helping fix the language, helping fix the communication so that it all works for people and teaching folks around you that in a consistent way. Amazing. Here's an example.
I led a weekly business review for the North American businesses for, call it, 10 years straight. So, you know, 500 meetings. With the top 80 to a hundred people in the retail business. And we would go through a deck of 50 slides in an hour. And I would not look at the slides before I walked in the meeting.
And every single one of those 500 days, there was some nugget presented to me just by concentrating on the slides where I had a chance to you know, expose a potential trade off that people might make where they might choose poorly and help them build their mental model and how Amazon thinks about making that trade off. And it just every single meeting, I found a few of those, and those we would call teaching moments. And the sum of them over enough occasions adds up to relearn Right.
And so it's a seminar Morgan. Even if it's not my mistake, I can see someone else make a mistake. You correct it, and I can then not make that mistake myself next time. And I don't have to Exactly. And it's tricky. You have to not embarrass the person who made the mistake. You have to enroll them in helping everybody else learn and but if you set that environment up right, boy, it's powerful. Yeah. Good for you, man. So, Wilky, I gotta congratulate you on this huge epic run at Amazon.
And now I know you can't stop building and inventing. You're just built that way. So What are you building and inventing now? Well, for the last 6 months or so, I've been working on a company that we launched in January called rebuild manufacturing. The CEO is a classmate of mine from Grad School at MIT.
I'm the chairman, the non executive chairman, We have a great small collection of investors and a really good board, and the mission is to build a multi division industrial company that will build US Factories, employ US workers.
We're starting with the number of acquisitions of companies that either have physical technology So material science, physics, mechanical engineering and so on, or have established product market fit and are generating, you know, significant revenue, but don't have the latest technology and combining them into an operating company with a healthy dose of computer science and automation that we think can compete effectively against global
competitors and in businesses that example, NFX is gonna be in where you'll have some of the companies, you know, wanting to produce physical products but yet not having, you know, really any idea how to do it at scale perhaps, or if they have an idea how to produce it at scale, don't have the Pete tier supply chain necessary to provide components that we would that rebuild would turn out to be a really good American partner for those kinds of firms.
So I can imagine over time in, you know, renewable energy and EV of all kind and you know, and vertical farming. And, I mean, all these things that are new that have physical componentry to them are gonna need manufacturing supply chain, and we're here to help amazing. So is this going to be geographically centered somewhere, or is it gonna be all over and then connected through a hub of transport?
The first three acquisitions include Flint, in Denver and the Carolinas and the headquarters is Boston. You know, we'll have locations wherever the first pennies that we buy are. And then after that, We'll start to build, you know, greenfield sites. And, you know, we'll probably end up doing that across the country also. Mhmm. We'll do it to be near customers. We'll do it where there's talent that we can tap into and so on. The goal is not this isn't a fund.
It's not a a loose collection of companies where you know, they keep their individual focus and don't really integrate. We want this to be a lasting Morgan real company, and we wanna build a culture. In fact, one of the first things we did was write the leadership principles for rebuild manufacturing. They're called the rebuild way. We actually posted them on rebuildmanufacturing.com. And, you know, we're hoping that that has the same cultural effect in some small way that it did at Amazon.
Well, it sounds like they're gonna have to find a Jeff Wilkie who's 32 or 33 to get this sucker really moving. Well, I'll tell you, the CEO miles has the energy of Jeff Wilkie at 32. He was that way in MIT, and he's still that way. So I'm very grateful. That's amazing. That's amazing.
So if there are startup people out there who are listening and you wanna join a great culture and learn about manufacturing and learn about maybe IoT and other things, and how this whole world of software and manufacturing is gonna come together, then you know where to go, rebuildmanufacturing.com. Thank you for that. We would hope to be one of you know, of many American firms that kind of revitalize this important part of the economy.
I think the pandemic and everything that the US struggled to, you know, to produce that was sorely needed brought this problem into relief. Awesome. Well, Jeff, it's been great to talk to you, pal, and I hope to do so soon again. Always a pleasure, James. I'm glad that you got a chance to connect, and I can't wait to raise a Pete. You've been listening to the NFX podcast.
You can rate and review this show on Apple Podcasts, and you can subscribe to the NFX podcast on Apple Podcasts, to modify, Google podcasts, or wherever you get your favorite podcasts. For more information on building iconic technology companies, visitnfx.com.