Clay Christensen: Hi, this is Clay Christensen, and I want to welcome you to a podcast series we call the Disruptive Voice. In this podcast we explore the theories that are featured in our course here at HBS, Building and Sustaining a Successful Enterprise. In each episode, we'll talk to alumni of our course, and others, who are trying to put these theories to use in their lives and in their organizations.

It's great fun to hear from them. And I hope that you find these conversations inspiring and useful. If you have an idea about a topic or a speaker that you'd like to hear more about, or if you'd like to comment on our work, please reach out to us here at the School.

Derek van Bever: Hi, I'm Derek van Bever, and you're listening to The Disruptive Voice. Even more than most episodes of this podcast, today's conversation with Andy Jassy, the CEO of Amazon Web Services--and a 1997 graduate of HB--is dedicated to the memory of Clay Christensen, and to all those who love him. Clay asked us to reach out to his friend, Andy, back in 2017, when we were researching what Clay called the Capitalist's Dilemma those of you who are familiar with our research will perhaps remember that the question we were trying to answer was why established companies were shifting their innovation focus from long-term market creation to short term "sustaining" and "efficiency innovations": cost cutting, not growth.

This was particularly concerning to us from a job creation perspective or more commonly for large corporations, a job destruction perspective. We identified lots of drivers of this shortsighted behavior. The way innovation pipelines are formed, the metrics that we use to evaluate innovations, the difficulty of steering capital to truly longterm projects, stock buybacks, managerial incentives. It was a pretty hefty list.

An anomaly in the large corporate economy to this short sighted activity is of course, Amazon: 26 years young and still in day one, as they describe themselves. So we interviewed Andy and learned a ton on about how Amazon identifies evaluates and pursues new business ideas. Amazon Web Services is one of these: a market and world changing innovation to stunning effect.

Clay was hoping to meet with Andy to follow up on our conversation, but his health didn't allow it. We reached out to Andy after Clay died earlier this year to ask him if he would share with our listeners, some of what had so impressed Clay from that conversation. And he's graciously agreed to spend some time with us today.

So Andy Jassy welcome to the podcast.

Andy Jassy: Thank you for having me. I appreciate it.

Derek van Bever: Let's begin a little bit chronologically. You've been with Amazon since leaving HBS. Could you share a little of your career path with us? You graduated from Harvard College. How did you choose to come to HBS and then upon graduation to take this crazy career risk with a startup bookseller out in Seattle?

Andy Jassy: Yeah, well, I had been living in New York in before business school. I originally--when I got out of college in 1990--I thought I wanted to be a sportscaster. And, I pursued that for a while. And then I worked at ABC sports and FOX TV for a little bit, and I decided that I just didn't have the patience to put in all the years before you got a chance to really get your shot at a big market.

I went and worked at a consumer, a company called MBI that did all of its products via direct marketing. It was a very entrepreneurial place where I learned a lot; got a lot of general management experience and then I left that and started my own business with somebody I'd worked with at MBI and I'd gone to college with. And that was a great experience, but we knew that what we were building was not what we both wanted to do forever. And we decided to close that down. I went back to HBS. You know... going to business school, I would say this, and I'm not sure people believe it. And I'm actually a little bit embarrassed to admit it, but I'm deciding where to go to business school was probably the single hardest decision I've ever made. It created such a life conflict and turmoil for me because it involves so many other things. And just where I went involved which part of the country I lived in. I had a lot of people giving me input on both sides, but I eventually got into both Harvard and Stanford. Which I didn't expect to have a choice like that. And it just, it kind of dug up a lot of issues, but I ultimately decided to go to HBS

Derek van Bever: Well from that point on, I mean, it's just the weather is the deciding factor, right?

Andy Jassy: Well, you know, I went to HBS because of the respect I have for the institution, in part, because I, you know, I really wanted to have the largest possible network of classmates, to kind of really through the work experience with over time. And in part, because I knew that the threat of the "cold call" would force me to work harder. And at that point, having been out of school for five years, I was little worried about whether or not I would be able to recommit myself academically, the way I knew was needed. So I ended up choosing HBS and which was a great decision for me. And my wife who's from--then she was my fiance-- now my wife is from the West coast. We lived on the East coast together for awhile, and she really wanted to go to the West coast to be closer to her family for a few years, we agreed. We, we go somewhere for two to three years and then moved back to New York. That was kind of the agreement we made on the napkin of a bar when we made our decision. And that was 23 years ago of course, and so, the statute of limitations is probably expired.

But I came Amazon. I took my last final exam at HBS, the first Friday of may in 1997 and I started Amazon next Monday. No, I didn't know what my job was going to be, or what my title was going to be. It was super important to the Amazon people that we come that Monday. So we did, it was three weeks before the company went public. And I joined this marketing team, which was about eight people at the time. And we just split up all the areas. All the areas were too big for each of us, but we just tried to split them all up.

I was supposed to manage competitive intelligence and customer retention. And I did that for a few months and then I was pulled off from that to work on what they call a SWAT team project, which was to explore what other product categories Amazon should get into before the brandsmith was a "books only" brand.

So it really was like writing a business school case. I took music and another HBS classmate of mine, Jason Kilar took video. Another HBS classmate of mine, Victoria Pickett, took software. And we all you know, tried to evaluate whether Amazon's customers would buy those products from us. We offered it and who was in the market and who would be successful and should we enter? Should we you know build there? Or should we try to buy it? In typical Amazon fashion, we decided to pursue them all.

I went back to marketing and built out our customer relationship marketing team, which I did for about a year and a half. And then I went back to music, I ran product management, and then was the general manager of the music business.

And then I went and spent about a year co-leading our marketing for the company. And then I left that to do -- it was a very unusual job -- which we then back then called, the Shadow Job where I'd shadow our founder and CEO, Jeff Bezos. It was really like a chief of staff role. And that was very interesting. I did that for about 18 months. And coming out of that job we had this idea, that maybe we could build a technology infrastructure platform that would both allow Amazon to develop and build things more quickly, but also lots for the third-party companies. So I left that Shadow Job to go work with a team on what we called a "vision document" for what a business might look like there. And that's what turned into AWS.

And that was back in late 2003. And since then I've been managing the AWS business and other parts in my world--I manage our payments capability, then an external payments business we got and ultimately moved that to our eCommerce platform. And I took on a bunch of digital pieces, like our app store and our browser that we built and our gaming efforts. So, you know, along the way, things that work on as well, but most of my time, the last 17 years.

Derek van Bever: So I don't need to tell you that the career arc that you just described is the dream, the absolute dream of our students and your fellow alumni, to be able to join an exciting company early in its history like that. And to get your hand into so many different areas. You know, I was curious: so you graduated in 97 from HBS? Clay was so young then. I mean, he was, this was probably just pre-Innovator's Dilemma publication. So you never had him as a teacher, I don't think. How did you and Clay get to know each other?

Andy Jassy: Well, it's impossible I think, to be an HBS student and then to be somebody who studies business and not know Clay. I mean he is arguably the most impressive business thinker of all time. And as somebody who is as passionate and focused on entrepreneurship and innovation as I am, and there are a lot of other people I know who have similar passion--I mean, just reading how he thinks about it and the things you got to think about as an innovator and the things you have to think about as a larger company, who has innovated, but has had some success. And you know that there are so many young hungry companies who want to innovate and take away market segment share from what you're doing.

I mean, his books have really been the Bible, I think, for all of us. So I had a lot of respect for Clay and, you know, I didn't know Clay. I had a couple of times I came back for HBS reunions, where they teach those classes which I really enjoyed, and I came to Boston every once in a while for business and I just cold emailed Clay one day and just asked him if he might be willing to spend some time with me. It was relatively early in the AWS journey. I think it was maybe around 2008. And he very quickly and graciously said yes. And I had a chance to spend an hour with him.

And you know, I learned a lot in just that time we spent. And fortunately at that point I had lost a lot of my inhibitions that I had in my first 10 years out of business school about feeling like I was asking dumb questions. And I just asked lots of questions, many of which were probably naive. And he just was very thoughtful in both giving the principles and constructs to consider, but also not telling me what to do.

I just really appreciated that. And we stayed in touch over the years. You know, I continued to be very appreciative of the time and the work that he's done.

Derek van Bever: Clay lived for those moments when former students and people with interesting questions would come through his doorway and spend an hour with him.

And it's remarkable since he's died, as we talk to folks whom he got to know, how he was always so busy. And yet your story of--you know--he was there, we spent an hour kind of on the fly that happens so much. He was so there for other people and always there to help it. So we miss him every day.

Andy Jassy: I also thought, incredibly, unbelievably humble. I mean, there were so many amazing things about him, but just the humility he had, the curiosity he had. It'd be easy for somebody who achieved as much as he did, to kind of sit almost and tell you what you should do, and to kind of opine, but he was much more curious about the challenges we were facing and how we're thinking about things than he was trying to dispense advice.

And I just was very admirable person.

Derek van Bever: Part of Clay's secret sauce was that he was always learning. You know, for somebody, as you say, who was as smart as he was he was, a great thinker maybe because he was so humble, you know? He would just, he would always say, you know, "shoot holes in my ideas", you know, help me, help me make them better. Very rare.

Speaking of rare mentors, so relatively early in your career, you got the opportunity to shadow Jeff Bezos. What did you learn about management or innovation, management of innovation, perhaps, from, being up close with Jeff at that juncture in the company's history?

Andy Jassy: Well, he's such an unusual thinker and, I don't know if I've ever seen anybody with the mix of skills that he has. He is very inventive. He is an unusual mix of both having strong technical competence, but also very strong customer empathy. You don't usually find those two in a single person.

People would bring him ideas that I felt were very, very inventive and very big. And he had a way of making them bigger and sometimes managers will ask for more and in asking for more, people turn them off because they've asked for something that just is unrealistic and people just say, "yeah, that's not viable." But he had a way of really building on people's ideas and making them bigger and just at the point where they were viable and people didn't believe they weren't possible.

And so you's great at taking big ideas and making them bigger. I thought that I had very high standards before I started that job. And I think most people would, would have argued one of my strengths was that I had high standards. And then, in doing that shadow job, I realized that my standards weren't high enough. Jeff has unbelievably high standards. And I think when you're managing a really large organization, it's growing at a pace that, many of them are at Amazon where you can't be in all the meetings anymore, you know, one of your best ways to have leverage is building the culture and the expectations and standards. You want us to have very high standards and make sure people understand what high standards are.

I learned from him and from being involved in a lot of innovation that innovation is messy. You know, I think a lot of times people think you can schedule innovation. Or, you know, we need to brainstorm about this new idea. Let's do this in 60 minutes and we're going to get it done then and we're going to have the right people in the room.

My experience is that it's not that clean. It's messy and it takes a lot of detours and you go, you end up at a lot of dead ends. And people have to argue with each other and people debate passionately and they riff on what each other are doing, but it just is messy and it's okay. Either we've been taught, before I did that job, you start meetings exactly on time and finish meetings at the scheduled ending of meetings. And that is often times the most efficient way to run a business. But I think when you're inventing, you have to just let the conversation take you where it does. And sometimes I would find that you would, we would go an extra hour or 90 minutes because we felt like we were trying to invent something.

If we weren't there, we felt we had, you know, some glimmer of hope and what we were doing and we wanted to keep pursuing it. And that was actually a struggle for me as a relatively young manager at that time. I think the other thing too, is that you; if you're going to invent a lot, you have to be open to learning. And you have to be willing to challenge your most closely held beliefs.

And Jeff has very strong opinions, but when he's in brainstorming mode, when he's inventing, he is very careful in listening to other people's ideas and viewpoints. And then trying to build on top of those. And I think there's some people who really want to be the person who can supply the answer. And the reality is as leaders, as managers, our jobs are to get to the right answer, whatever those are, whoever asks, then we have to kind of build a process and facilitate a conversation and move it along. So you get to that right answer is that really does change the equation for customers. He's very good at adjusting on the fly as the conversation is going on.

Derek van Bever: We in Clay's course--that we teach at the School--Building and Sustaining a Successful Enterprise, the term that we use, I think it's from, Henry Mintzberg is "emergent strategy." This idea that you have to be open constantly iterative learning and understanding that, you know, the processes that you've developed that build this, that sustain you, that take you up along the sustaining trajectory are not of much use when you're inventing new things and you've got to kind of separate those processes to make sure that you're not getting captured by, as you say, what you believe is correct. You have to be open to that, to that discovery. You know, where Clay's ears really perked up in our conversation was when you were talking about --you told us a really interesting story from your early days with Amazon, about how you all changed your new business investment proposal process over time to focus away from measuring NPVs and, you know, metrics. If Clay had the chance to really dig into the next chapter of research on capitalist dilemma, he was super interested in how metrics can in some ways be innovation killers. Um, could you share that story with us?

How did you measure before then? And how did you, and I guess it was you and Jeff said, Oh, this isn't working very well.

Andy Jassy: Well, we have this process and the company, as we started to get bigger and there were so many new initiatives we wanted to pursue with limited resources. And so we, we built a process of the company called NPI, which stood for New Product Introduction.

And if you had an idea that required technical resources across multiple groups, you would have to present your idea that this NPI meeting, that we had every six months with this big group of people from all the technology teams. And you would present your idea. And then at the end of it, each person from those different technical teams would hold up a card with the number of "people-weeks" it was going to take to get done their part of the technology. It was actually quite dreary because first of all, you knew was kind of your one shot to maybe get this thing funded. So there was a lot of pressure. There's a big group of people. It's like this hall, this kind of big rectangular table. And there were so many good ideas, so much clever innovation percolating inside Amazon, and we would get done with it.

And we'd only be able to fund them like two of the ideas. We'd have 20 good ideas and only be able to fund two. And then the way we would choose often times was overly reliant on the projected NPV. You know, and of course, if you do something new, it's so hard to project what the NPV is and you know, what ends up happening.

The only NPVs that are really sure bets are the things where, you know you're going to take, you're going to take fraud out of the system so it's a very predictably high NPV, or, you know you can take costs down in some fashion, but with all the new innovation ideas where you're not sure how customers are going to react to them. It's, you know, the, the bowl people put high end NPVs and the more modest people put more modest NPVs and we would often make decisions based on the NPV. And, you know, in the time when I was shadowing Jeff, one of the things that we talked about was just our frustration and how few of these projects we were funding. And when we looked back at the last several NPI sessions, we looked at the NPVs of the ones that we funded and the ones that we didn't and others who funded ideas like them and we weren't so happy with the choices we made. And so we, we kind of did, we decided two things. One was, we weren't going to use NPVs as the primary measure for whether we fund something. Just because--I think if you, if you fool yourself into thinking that you could predict the NPV for something that's new, you are doing just that you were really fooling yourself.

And so it's false precision. So we're weren't going to rely on that. The other side was the root cause of the problem for us was that our software was too coupled. And so if you wanted to build something that required multiple teams, you know, each one was a project where each technical team had to dedicate a certain number of technical resources. And it was really hard to manage it. We decided what we wanted instead was one of each of those technology teams to take their technology and to build well-documented hardened Application Programming Interfaces or APIs so the teams could use their services, but without ever having to talk to them or coordinate with them. You know, really almost like services. And that would remove this needs to have an NPI process, because if I needed the search capability or browse capability, or an item catalog, I just use their APIs and then scale it to whatever use case I want it. And so, that took us several years to address, you know, getting all the teams to have these hard, well documented APIs. But that allowed us to, to kill the NPI process so we could move it at a much faster rate.

Derek van Bever: You know, Clay loved that idea. That image of you and Jeff looking at two piles of projects, and saying, "here are the ones that got through our process, here are the ones that didn't; boy I sure wish we were working on the ones that didn't because those look cool." And, I imagine a lot of our listeners can really identify with that challenge that the NPV gate is going to always favor the closer-in more-certain, as you say, you know, efficiency innovations where the result is predictable and you miss all the fun that you might have on the other side of it. How did the idea for AWS come about?

Andy Jassy: Well, we had a few things that were happening around the year 2000 or so. And the first few years, I'll say, of this century, that were making us believe that there was something there. You know, the first was, we had this business that we tried called merchant.com and it was-- we enabled other third-party eCommerce merchants to use our eCommerce technology, fuel their sites.

These were things target.com or Marks & Spencer. They use Amazon's e-Commerce technology underneath their websites and to deliver that capability to them to be a lot harder and a lot slower than we ever imagined, because of what I was saying a little bit earlier, where we had jumbled together parts of our platform at Amazon and building so quickly the first eight or nine years, that to deliver to them the functionality so they could customize, you had to deliver all the individual pieces through APIs. And so that process being so hard gave us religion inside of the company that we really wanted to build on the service oriented architecture faction, where the services were loosely cobbled with hardened APIs we could use however we want. So we really, around the year, 2000, we became a services company very quickly and with no fanfare, but totally changed the way we thought about software.

The thing that was happening was, we were adding a lot of software developers in 2002 and 2003. And yet all the projects were still moving slower than we thought. And yet for a long time, when things didn't get done we'd say, "well, if I had more people, this would get that faster."

Andy Jassy: And as we had started adding people that just wasn't happening. And at the time I was in the shadow job, working for Jeff, and I would spend time talking to the product leaders. And they would say, "well, it's really great that you guys think this should get done in two to three months, but we spent two to three months just on the storage piece or just on a database piece, just on the compute piece and nothing we're building scales beyond our own initiatives. And all my peers are doing the same thing and reinventing the wheel."

And that was very interesting to us. In retrospect, it was probably fairly obvious, but it wasn't at the time to us that there was this real thirst inside Amazon to build a reliable, scalable cost effective set of infrastructure services so that people can move much more quickly.

Because when we ask people, what percentage of your time is spent on what differentiates your initiative versus the undifferentiated heavy lifting of the infrastructure, about 80% of their time spent on the undifferentiated infrastructure. So we wouldn't want to try and flip that equation on its head. That was the second thing that we realized.

The third thing was we were doing this offsite at Jeff's house and we were evaluating different businesses. We have this one call on a whiteboard that said "consumers and all the capabilities we're building for consumers", then we have another column that was, "sellers and all that different capabilities for third party sellers."

And then we wrote down words, "developers/enterprises". Then we asked ourselves if we believed that they would build applications from scratch using infrastructure services, if the right selection existed, which we believe they would, if it existed. Well then the operating system became the internet, which is really different from what had been the case the last three years.

And then we asked, "if there was going to be an internet operating system, what were the key components of what's already been built and what--given what Amazon's good at--can we meaningfully contribute to? When we looked at that mid-2003, we realized that none of the key components of that internet operating system had been built.

Andy Jassy: And when we thought about what we were good at, and Amazon was always a technology company at its heart, we realized we could provide all of those components in an internet operating system. And with that, we decided to pursue this idea, which I mentioned earlier, I left the shadow job to go explore with people, to build a technology infrastructure platform that other companies could built all their technology applications on top of. And that's, that's really what became AWS. And in some ways, the way we look at new businesses enabled us to pursue this. It's also, it's similar to a lot of the teaching of Clay around thinking about the jobs that need to be done. You know, when we moved away from looking at NPVs as the measure for whether you should try to do business. We moved toward this model of looking at five criteria:

[1] The first is if we built it and it was successful, could be really big and really big in a way that would move the needle at your company? Whatever that level is for your company.

[2] And then we look at: is it being well served today?

[3] And then we look at: do we have some kind of differentiating approach to it?

[4] We look at: do we have some competence in the area, or if not, can we acquire it really quickly?

[5] And then if we like the answers to those first four elements, then we ask, can we put a group of single threaded focused people on this initiative, even if it seems like they're overwhelming it with strong senior people, if you try to add really busy people do the existing business and the big new idea, they will always favor the existing business because it's surer bet. So we want to peel people away from the existing business and put them just on the new initiative.

And so when we looked at those elements in AWS at the time, we thought the idea could be very large. You know these market segments that we could address with AWS, which is really instructional software, hardware, and data center services globally. It was trillions of dollars. So it could be really big. We thought it was being extremely poorly served at the time. We had both ourselves as a customer, as well as a number of other third party developers who just hated the pricing model, hated the commitment model, didn't feel like the services catered to smaller developers of companies, only to really large enterprises willing to spend a lot of money. They felt like too much was coupled into each service that they didn't need or did want to pay for. And they thought like the progress was so slow. So we felt like it wasn't that well served and we had a differentiated approach in what we were going to do at AWS. We had a lot of competence because we had to get good at running infrastructure services and data centers to scale Amazon the first eight or nine years.

And so we decided to peel off a number of people, from the core business, who we thought were deep in this area and could just focus on it. And we decided to go pursue it. You know, I think it's quite similar to how Clay used to always talk about thinking about, you know, what is the job that people want to hire you for?

And what's not working today and how can you do the job better and then build an experience that way. That's what we try to do.

Derek van Bever: Oh, Andy, I mean, what you just delivered was a graduate seminar in Clay's thinking, right? Jobs-to-be-Done. RPPs. Asymmetric motivation that gives you an advantage over the incumbents. Setting up an autonomous unit. Hybrid disruption with new market and low end disruption combined. Unbelievable. Unbelievable prescience and a great testament to what Clay taught us. And you told a great story, just to close out, you told a great story about one of the problems with doing NPVs as all of our students coming out of consulting firms know is that you can basically make the NPV whatever you want it to be if you just leave me alone with Excel and some assumptions for awhile. So how did-- did you do an NPV on AWS or how, how did you, how did you sell this to your peers and management and then to the board?

Andy Jassy: Well, it's a very good question. And it has kind of a little bit of an interesting story to it. So we modeled AWS every which way to Sunday. And we there's various sophisticated, complicated models. It was very different from the rest of what we did at Amazon at that time where we were a retail-only company.

And when we looked at it, you pretty quickly realized there were a half dozen metrics that totally changed how big the business could be and how profitable business could be. And frankly, we just had no idea how to project how many customers would use it. How soon? To what extent? We had some ideas about how highly utilized we could make the hardware in the data centers, but, at scale, you know, some of those things you don't quite know. There were a lot of unknowns, but we had a model and we had sensitivities for a model but we spent very little time, collectively, as a senior leadership team on the model. It was really, you know, we look at new products then how I look at it today when we make new investments in IR businesses. You really just looking to see, "can it be big?" And can it be profitable and sustainable? And that's really what we looked at, but when we presented the idea of AWS to our senior leadership team, which we call our S-Team, first, and had a very good animated discussion about it and decided that we were going to pursue it.

And then we presented it to our board and, you know, we don't do PowerPoint presentations at Amazon. We only do narratives. Six page narratives. We spend the first 20 or 30 minutes of the meeting reading that narrative and then people ask questions and make comments. We have a discussion. And when we did the board conversation, the board read the document, and they said, "I think you may have left out the pro forma P&L because it wasn't included in documents." And I said, "Actually, no, I didn't unintentionally leave it out. I purposefully left it out just because it's so hard for us to project how customers are going to react. And there's a half dozen variables that could make this a million dollar business or a billion dollar plus business, not profitable at all or meaningfully profitable. And I went to the whiteboard and kind of showed a little bit of what those variables were and what they could be. But, I think it was unusual at that time. It was unusual for the board and we were making a very big investment in AWS. But I didn't have it in the doc, just because we spent hardly any time looking at that as a senior leadership team, only to know that it could be big and it could be sustainably profitable.

Derek van Bever: That anecdote -- that illustration -- so many times metrics are used to kill a new idea and particularly, you say it's obvious in retrospect, and of course there was a ton of risk around it, but you know, you kind of held hands together and said, "If this works, it could be really big. There is an unfulfilled job to be done here that Amazon is super well-positioned to provide in a very economical way to a new segment of customers who have been priced out of this in the past." And that's textbook Clay.

The researcher who originally reached out to you, a young woman named Lior Eshkol, she was researching job creation, which was what she was what she was interested in pursuing with you, originally. And I wonder, do you track job creation at Amazon? Is that an important metric to you all?

Andy Jassy: We do track it. And I think we're really proud of the number of jobs we've created, you know, in the world and particularly in our country. I mean, we have a million people work in Amazon. It's kind of crazy. We have hired more people over the last five to 10 years than any other company. And I think we're proud of that. But anyhow, we don't really think of job creation for the sake of job creation. What we're really trying to do is we're trying to build amazing customer experiences in a number of different areas and in areas where we feel like, either there are brand new whole cloth inventions that create customer experiences that we think people really want and will really help customers. Or in areas where we'll look at a certain customer experience and realize that it's not what customers want and it appears broken. And then we talk to customers and they confirm that. And then we try and build something there. And if we have success, it just, it tends to lead to us needing to hire a number of builders in a lot of different areas. So all the job creation is really in service of trying to provide the right, great customer experience. I think that when people sometimes look at Amazon and they, and they look at the company getting larger, and, being a bigger revenue business, and sometimes people will ask questions:

"well, you know, do you think people have to use you?" or "is it just because of your scale that you're able to pursue these other businesses?" I understand the question, but I always think it's a little bit of an odd question, because I think if you look at the world we live in today, particularly in the access you have to so many products. Customers have the choice. And simply because you want to offer a solution in a particular area, if it's not great, if it isn't what customers want, it doesn't do their job the way they need it to in a better way, and they can do it elsewhere. They're going to choose something else. And so I think what's really unusual about Amazon is not just that we are unusually customer focused and unusually focused on the longterm. But that, we've been able over time, through a lot of invention and a lot of experimentation--and failure, by the way, we've failed a lot of things with trial--but we have been able to build really great customer experiences in a few different business areas and dimensions that customers have loved and have continued to push us to keep evolving and we have done so at the time. That's why we've been able to hire as many people as we have. We built great customer experiences that people really liked and were able to grow.

Derek van Bever: That's such a good reminder. You know, one of the things that worries us as a teaching team, is that we're concerned that our students, you know, many of whom aspire to be startup founders, kind of think of customers as an abstraction. Like, "if we build it, they will come," or, you know, if we price it a certain way, we can induce them to buy it. And we're like, "Ah!" you know, I forget who it was. Who said "the job of a businesses is to make a customer", but boy, customers have more choices today than ever. And you have to be really, really sensitive to the progress they're trying to make and how they think about, you know, what attracts them to you. What, what repels them perhaps from your offering and you, you've got to be really outside and focused. And we try to reinforce that with our students.

I was doing a little bit of research before the call in some conversations you've had with other folks about, you know, why competitors took so long to respond to AWS. And of course, this is what disruption theory would predict, right. Clay called this asymmetric motivation. But, how did this play in your experience? Were you surprised as AWS you know, continued this explosive growth path, were you surprised that competitors took so long to catch on and follow on?

Andy Jassy: Totally stunned.

And in my wildest dreams, I never anticipated having as much runway as we did. You know, we were trying to just get to launch without anybody else knowing and trying to beat us to the punch. And at the time, you know, back in 2000, this was, I'll call it between 2004 and 2006. We launched in March of 2006. Amazon was known as has having good technology, but really was a consumer company. And I think that it was imperative that we were first to market to be successful on building an enterprise technology business. Like we have made in AWS. So we were just trying to get to launch first. And, you know, we live in a relatively big city, but it's a small, big city in Seattle.

And, we all know a lot of people across the lake at Microsoft and, as I said, it's small city. We have a lot of friends in both places. And we just weren't sure we'd get to launch and then to have several years. Yeah I don't think we anticipated that at all. We always knew there'd be a lot of companies pursuing it because it's a great value proposition for customers.

I think there are a few reasons why maybe people waited a while. I don't know for sure. You know, I can just tell you what I think, but I've been asked a lot.

The first thing was that I think most of those big technology companies probably didn't respect Amazon as a technology entity the way they might have. Now, I think they just saw that we were kind of a book retailer. And how are they going to build technology that's going to be more general purpose to enable lots of companies, particularly larger companies?

So I think in the beginning there was maybe some underrating or dismissing of Amazon as a technology provider. Based on what people knew about us back then. I think this, the second thing is that, I think it's really hard when you have an existing business with an existing margin structure that's going reasonably well to disrupt that business yourself.

You don't want things to disrupt it. So I think companies have a tendency to wish away new technologies that are disruptive and better for customers. We often call that fighting gravity. You know, a good story from our early days that helped inform how we think about this at the company, which was:

In the early days of our retail business, we had this old inventory business where we bought products from distributors and publishers, stored them in our warehouses, then shipped them. So we did all the inventory shipping. And we had a very high quality customer experience. And then around the years, call it 1998 to 2000, you started seeing these marketplaces like eBay or ask.com where third parties were selling these products.

We had this huge animated debate inside Amazon, whether we should have third parties able to sell on Amazon. And there was a lot of resistance to it in part because the whole company was set up as an owned inventory company. And in part, because we just couldn't imagine that third parties would take care of customers the same way that we were.

But we ultimately, after several weeks, decided that we would pursue a marketplace capability. In part, because, you allow customers to have more variety of products, as well as prices. But also because we just concluded you can't fight gravity. If something is going to happen, because it's a great value proposition for customers, you can howl at the wind all you want, but it is going to happen. You are much better off trying to shape it then to chase it.

And so I think the same thing was true here in the AWS space where all these big enterprise technology companies poo-pooed the cloud and said, "oh, you know, I don't think anybody will use it in any meaningful way" and then startups started using it in.

And they said "well they'll never use it in production." Then startups started building their entire businesses on AWS.

And these companies said, "well, enterprise will never use it." Enterprises are building new projects.

And they said, "well, never building anything mission critical on it..." and then enterprises started building mission critical applications--they realized that that was going to happen.

In the early days, we found it frustrating at times that so many big companies were poo-pooing the cloud and journalists who write about it. But, we kept seeing the adoption grow. And the more those old guard companies, poo-pooed it, the more we loved it because it just meant they weren't building, you know? And, and so I think that, you know, it's always in Clay writes about this a lot.

It's always these new companies with a new approach that just feel like irritants, but you kind of dismiss them and assume they're not going to get traction and you often wait too long. And they always start at the bottom, you know, like we did, we got startups and we got, you know, they were building business.

And then we got enterprises to try their new websites and put development tests on us, you know? And then we kept building capabilities. As we got traction and we got feedback. And before you know it, there's something much more pervasive there. And I think those were really, at the end of the day, those are the big things.

And I think a couple players then decided they need to really get going and they just chose the wrong abstraction to build. They built too high in the stack as opposed to these building blocks like we built, that allowed developed to stitch them together however they saw fit. You know, at the end of the day, the companies after a few years that really got it in gear and pursued it in earnest--I think have done okay in this era. And those that didn't for the longest, have really struggled.

Derek van Bever: One of Clay's most endearing homespun images was he always talked about, you know, disruption always enters at the basement door. And so your idea of starting low in the stack, you know, just, it's the kind of thing that most incumbents look at and go, "Oh, you know, who would, who would buy that? That's just a toy!" But toys have a way of turning into jet fighters over time.

Andy this whole conversation has been chock full of advice for our listeners as to how to think about innovation, but I wonder, are there any role models or thinkers other than Clay that have really influenced you that you might want to commend to our listeners?

I know Wendell Weeks is on your board and we've always had a particular soft spot for Corning, but are there companies or thinkers that you would recommend?

Andy Jassy: Well, there are so many amazing companies and examples of innovation. You know, we talked a little bit earlier about Jeff, who you know -- I'm not sure there's a better innovator in our era. I think that if you look at what Apple's done, it's really quite remarkable. And I think over the years, Nike and what they've done in their space. I mean, there's, there's so many great examples. I think that if you look now, there are a lot of mainstream well-established enterprises that are really reinventing themselves in ways that a lot of people may not know and may not appreciate but I find really impressive.

If you look at a company like McDonald's. If you look at how they radically changed their point-of-sale systems and how they've been able to use it to get better analytics and to do personalization as people buy all sorts of things from them. If you look at, they bought this company called Dynamic Yield, that was an AWS customer, and built really great automatic speech recognition, and the way the McDonald's is starting to evolve with the experiences and drive-throughs, which is a surprisingly large percentage of the total orders they do every day. It's really impressive. You know, you look at some of these companies like if you look at Intuit, and, most small businesses use their products in some meaningful way--just how they're using machine learning in their products. You look at their small business tax products, they now have built machine learning algorithms on top of AWS that allow them not just to build and deploy the models 90% faster than they were able to do so before, but they they're now automatically able to detect and predict savings that customers can have. They're saving people on average, you know, $4,300 a year and it allows them to file the return in 15 minutes when it used to take many hours.

So just up and down. Look at Goldman Sachs, totally rebuilt and reimagined what a consumer banking platform app looks like what they're doing with Marcus. Capital One too. The way that they've totally reinvented the consumer banking platforms and the experiences they provide, the features of their products.

There's a lot of companies who are really substantially inventing. And I think the keys, if I were going to give any counsel would be: you have to first decide as a company, that you really want to invent. And I know that sounds like everybody would want to invent. But not every company really is committed to inventing.

You know, it's, you're going to be committed to inventing, you have to find ways to encourage your builders, to be thinking of new ideas. Because the ideas can't just come from two or three people, it just doesn't scale. So you have to build a culture where you hire the type of people who want to invent, want to take good customer experiences and figure out how to make them better. You got to give them the right tools. And it's part of what we are most proud of with respect to AWS is that it just allows organizations to invent several orders of magnitude faster than they could before. So you have to give them the right tools and then you have to willing to experiment. In that, I think most companies don't really like experimenting.

Derek van Bever: Couldn't agree more.

Andy Jassy: They really want a sure bet. When you're inventing there aren't sure bets. It wouldn't really be inventing. And so, it's hard to preside over failure. I've presided over many failures and and it hurts. I don't like it. I feel badly about it. We all want to succeed. But, to have the really big successes. You got to take a lot of shots and you got to take some big shots and you got to be willing to figure out how to build a culture that accepts those failures and takes the learning and moves on.

Derek van Bever: Andy let's end the conversation there. What a great charge to our listeners to have courage as you go out and invent new things. You know we talked about Clay's courage, his humility, his kindness, his availability, when people would come on campus with hard problems or wanting his counsel. And, I can't think of a better way to celebrate his life and impact than this conversation. So, Andy Jassy thank you for your graciousness and agreeing to spend this time with us today and Clay, wherever you are, and I think we know, there is no end to the lives that you've touched and the difference that you've made in all of our worlds.

So to you, Clay, and to all of you for listening. Thank you very much.

Thank you Andy.

Andy Jassy: Thank you Derek, thanks for having me.

Clay Christensen: Thank you for listening to us at Disruptive Voice. If you like our show and want to learn more, please visit us at our website or leave us a review on iTunes. Until next time, good luck everybody.