#171: Working Backwards with Amazon's Colin Bryar and Bill Carr
DC: Before we get to the show, did you know you can get more insights just like the ones you're listening to right here on Seeking Wisdom, delivered right to your inbox? Sign up to get my weekly newsletter. It's called The One Thing at drift. com/ dc. All right. We're back with another episode of Seeking Wisdom. This is going to be an uncomfortable episode because I am a fan boy of the two people that we have as guests here. They're going to feel like they're rock stars and I'm a young teenage fan boy right here. And so I want to welcome Colin and Bill, authors of Working Backwards. And if you listen to this podcast you know I've been singing the praises of this book. I've been waiting for this book for a very long time. And so welcome to the show guys.
Colin Bryar: Well, thank you for inviting us. crosstalk
Bill Carr: Thanks for having us.
DC: We're excited. So before I introduce you formally to the show with your backgrounds, is there any coincidence with the timing of this book and Bezos' retirement?
Bill Carr: Oh, absolutely. We're discussing with him about when he should change his role for the last couple of years. And he was a little reluctant on the date, but eventually he agreed that it needed to correspond with our book release.
DC: To celebrate your book. I just thought it was funny because for so many years I have been trying to scour and piece together different legends of how Amazon worked internally and nothing that had come out, and it was very hard to find information. And then all of a sudden, Bezos announce his move to executive chair. Your book comes out, Wilke is on lots of podcasts. I'm like what? What's going on? Like everything's coming out all of a sudden, obviously coincidence, but I thought it was funny. So I'm just going to formerly introduce both of you so the audience knows. First is Colin Bryar. And so Colin joined Amazon, and I'm reading this, which I usually don't, but Amazon in 1998. And four years after its founding spent 12 years as part of Amazon senior leadership team and was famously known as Bezos' shadow aka Jeff shadow, which we'll get into, and ran many different things, including IMDb, which I was a huge fan of IMDb for many, many years, still am. And currently, co- wrote Working Backwards with Bill and they run a coaching firm, coaching and advisory firm, which they'll talk more about called Working Backwards LLC. And thank you, Colin, for joining us. And Bill joined just a year after Colin, 1999, five years after founding of Amazon. Spent 15 years there on the senior leadership team and spent lot of time, really focused on all things non- physical. So digital stuff, everything from Kindle to movies, to more. And so I'm excited to talk to you both, and I've worked front and backwards on this book. I've given it to most team members at Drift, and I've been a big recommender of this book. So thank you both for joining us.
Colin Bryar: Thank you. And thanks for recommending the book.
DC: Oh, I love it. So how long ago did you two start working on this idea for a book, and why did you think you needed to write a book?
Bill Carr: Well, the origin story goes back to, believe it or not, more than three years ago, February of 2017. And at that time I was spending time as an exec in residence with a venture capital firm called Maveron. And I was attending their annual offsite meeting with their portfolio CEOs, as well as some public CEOs who are in attendance. And there was a round table discussion about a variety of topics, but all of a sudden the topic of Amazon came up in this discussion. And one of the most prominent CEOs in this group, who's well- known figure I will not out him for the sake of this discussion because he said, " Amazon, I don't understand how Amazon does it. How have they been able to be so successful in everything from cloud computing, e- commerce, to digital media and devices, to third- party logistics? We're still trying to get our core business right. I don't understand how they do it." And I was incredibly struck by that, a well- known prominent very successful CEO would say such a thing. And was also struck by the fact of course, that I knew the answer to this question. And I relayed this conversation to Colin shortly thereafter, we were actually on a vacation together where our families are friends, our kids are the same age. And apparently we were in the kitchen. I was cleaning up the kitchen and I mentioned this. I said, " We know the answer to this question. It can be distilled to a bunch of repeatable processes and no one knows this. There's no information out there in the wild. You can't find this on a website." And I said, " We should write a book to answer this question." I think I said, " I should write a book." And then Colin said, " Yeah, I agree. We should write a book." And he totally horned into my project from that moment on.
DC: I liked the way you roll Colin. The part I can never understand about Amazon, because for so long, I've been trying to figure out how it actually worked internally, because everything seems so logical from what I heard from the outside, from two- pizza teams to all the legend was just like, why none of this information ever made it out. So many people worked at Amazon, but I never could find anything written down in the wild.
Colin Bryar: I think putting that together comprehensively as something that hadn't been done. Bill and I, we had seen blog posts or people talk about one or more processes, and it was really just their view of how they used it at Amazon. And that's why we realized there was something there. If you put everything together, it really describes some advances in management science that Amazon has created. And if you looked at holistically, it is something that smaller organizations and large organizations can do. And then the other thing is that both Bill and I were lucky enough to have been in the room and participating in making actually a ton of mistakes with these processes as they were being developed. We described them as fully formed in the book, but none of them started out that way. There was a journey to get there. And when I was working with Jeff Bezos as technical advisor, we were creating these processes to solve very specific problems that we were experiencing at that time. And often, for the Kindle example, for example, Bill is on one side of the room trying to figure out how to write a PR FAQ document. And I was on the other side with Jeff, with his TA coming in and reviewing that and then moving over to AWS. So we just had a unique insight and that was another reason we felt excited to get the story out holistically. We wanted to really get the story right, because there is something special that Amazon has created. And we just had a unique bird's- eye view. But also we were some of the early adopters on the bleeding edge who were involved in creating these processes.
DC: I love that you brought up the PR FAQ because I cannot tell you how much of my life energy I have spent trying to reverse engineer understand how that worked within Amazon. And maybe you can talk more about what the PR FAQ is for people, but I could never find a real example of how it works. Obviously I understand it in concept, but how did it actually work? How did it evolve? How did you stay true to it and not give in, which is the easiest human tendency to go backwards to a presentation format or something else. How did you not compromise? How did it actually work in practice?
Bill Carr: Well, I guess to answer the last part of your question first, is one of the hallmarks of Amazon is discipline. And that really starts with Jeff, which is there are many times I remember he would say when we were discussing a specific business problem or operational issue, and people would try to maybe take a route to say well about why it may not be possible to achieve it. And he would remind them, well, what you're describing really just requires the discipline for us to focus on it and ensure that we do achieve it. So that word got used a lot. And he employed it where you see too often in companies, there's many companies that flipped from management practices and processes inaudible and flip from one to the other. They're just bouncing around seeking new ones. And they're very much influenced by what may be the latest business book or business concept. And no doubt Jeff and Amazon were actually quite influenced by a number of outside business concepts. But the difference was that really beginning in the 2003, 2004 timeframe, we became more disciplined and so to them, we were very focused on, what is really the end goal that we're focused on here with this process? And we started to move away from or reject a lot of the processes that are commonly used or traditionally used, and really had our own mind to say, well, what is it we're trying to achieve? And in the case of the PR FAQ, we're trying to achieve as like, how do we make sure that when we're developing a new product, we stay relentlessly focused on the customer? Because in fact, if you look at how we started out and how most companies did it at that time, those processes didn't involve much focus on the customer, it involved a lot of coming up with different, we were thinking about entering the digital media business, we would say, okay, so what's the size? What does Gartner say that digital music will be in five years? What market share should we have that? What will the P&L look like, and how should we negotiate with the record companies, et cetera, et cetera, et cetera. Or maybe we would perform a SWOT analysis and look at our strengths and weaknesses, the opportunities, threats, and none of this would the customer actually come up. And so where we really became, how these processes came about was Jeff being tenacious about a specific end goal. In this case, how do I stay focused on the customer when I'm thinking about developing a new product and we iterated over time, the press release turns out to be the ultimate way to do that because you cannot help, but focus exclusively on why this product would appeal to a customer or not when you're writing a press release because that's the audience. And it also proved to be an incredibly simple tool. It's not heavyweight. Anyone can write a one page press release, anyone. Writing a really good one that actually really captures a true customer problem and a great featured the solves that problem is hard, but anyone can do it. Whereas certain other methods of designing products like building a prototype, making a design requires specialized skills and are time consuming. So long story short, it was really just about having that relentless focus on what's the goal. And then as Jeff would say, " Stubborn on the vision, flexible on the details," iterating, iterating until we get to the right long- term vision.
Colin Bryar: I just add that the long- term is really important there combined with the discipline, and willing to try and to invent new things. Even if everyone else uses slides, we realized, well, Darren does actually a better decision- making tool writing memos, because for a number of reasons, it's a harder path. It's harder to write a six page memo than it is to throw some slides together. That's one example of choosing the harder path. It may take you a while to get started, but you'll eventually be able to make better decisions across a wider variety of businesses. Another quick example is with metrics, we've come across companies who say, well, our executives want high level summaries of how the businesses running. At Amazon we inverted that. And we said for the same investment and the executive team's time, how can we put as much information in, in a digestible format by looking at controllable inputs for that one hour of weekly business review or executive time. And it's harder to figure that out, and it's harder to have the discipline to look at that every week, but those are just the PR FAQ narratives and metrics are three examples where Amazon chose the harder path in the short term, but it unlocked a ton of value and it actually speeds you up once you start operating in that manner.
DC: You said an interesting word to me that I'd love to double click on, which is controllable input. So controllable, what do you mean by that?
Colin Bryar: So once you have your company goals, be the free cash flow or share price, you don't have control over in any short- term time horizon. So what you need to do is you need to back up from that and say, what are the set of activities that we have control over that if we do those things right, we'll generate the desired direction and the output metrics that I named before. Think of a company as like a process. You know what the output of the process has to be, but the way you get there, as you focus on the other side of the inputs and the knobs and dials that you have control to finally tune this machine. And so for Amazon, the big categories at CM is on flywheel, and we're talking about the retail business. It's really price, which is all about how can you lower your cost structure to be able to afford lower prices selection? How can you on a metronome add more and more categories and products to your catalog that are available for sale, and then convenience, which a lot of it is easy to find products, but then also really click to deliver when a customer decides to buy something, how long does it take to show up on their door? And so Amazon realized that if you focus and get those things right, your customers will reward you with their trust and their hard- earned dollars and your business will grow, and the share price will appreciate, but you can't focus on the share price. You need to have a detailed set of metrics. I outlined three broad categories, but if you were to look at all of the individual metrics that you have to get right in order to do that, that's a long list. And you look at those every day and every week. You have to have faith that once you do those controllable inputs in the right way, that it will yield the right outcome in your business.
DC: You both mentioned the word discipline, and I love to talk more about that. Because that's the part I could never figure out as an outsider. And the idea of building around the customer is something I've been fascinated with ever since the early 2000s when I read Made in America by Sam Walton, and I became obsessed with it, which I bought on Amazon, by the way, very, very early 2000s. Became obsessed with that. Or some of the PR FAQ or some of the controllable inputs that you're talking about, they're all things that are simple to understand. They're logical, they're rational, but they're not easy to actually implement. The discipline part is almost the impossible part. It's almost like how do you lose weight, stop eating and move more. Yeah, I understand that, but no one can do it. What made it possible within Amazon to have that long- term vision, have that discipline, especially in the early years when, I mean, you guys were there in 1998, 99, and then you started run run- up and then the world fell apart, including Amazon stock in the early 2000s. How did you maintain discipline during those times?
Bill Carr: Well, first of all, I would say I wouldn't submit, and Colin may disagree with me, but I would say in many ways, Amazon was not a terribly disciplined company in those early years. It was like most startups where we were growing in a million different directions and we really didn't have a lot of great processes and controls. And for example, the year I showed up in 1999, we exited the year with excessive amounts of inventory relative to the sales that we had. We had, I can't remember what it was, but there was some of the cardboard corrugate, or one of the elements we needed to ship products. We had a mountain of it in one fulfillment center, way more than we need, because people didn't know how much to order. We were just trying to hold on to, for our dear lives, that runaway train of demand. So it really wasn't that. So to be clear, this is an important point because it's not as if Amazon was Athena that was born out of Zeus's head fully formed. We went through all of the awkward growth stages that any growth company goes through, where we were successful is that I'd say we transitioned in the 2000s, the early 2000s, and probably most notably starting around 2003 to start to build out and establish some more run processes. And in some cases early on we'd actually started to implement processes from other companies. Colin remind me of the exact date, but it was around about 2001. When we started, what's called the NPI process of the company to try to plan out how, now we have so many different divisions and groups, we have international, how are we going to plan out the work across this now complex company? And we adopted a process from GE or a simplified version of it. And we ended up projecting that process and rejecting that method of doing really deep communication collaboration, and went in a different direction with initially two- pizza teams, then later single- threaded separable teams. So we started to go down this path and we're actively seeking, we had a problem. The problem was the company was big. It was growing fast. We were in many countries, many businesses. So how are we going to effectively manage this, but how do we maintain our customer focus and how do we maintain our agility and ability to continue to innovate? And so those are really the parts where, because we wanted to maintain those things. Some of the off the shelf management practices that are out there did not actually enable companies to maintain those two things. And so we sought to make changes.
Colin Bryar: Part of it is fostering a culture that allows that to happen. At Amazon, we said off and a lot, " You don't let defects travel downstream." So if you notice something that goes wrong, you need to make sure it gets fixed. Because if an error happened, it's usually a process error and it will happen again. And it probably will happen again. And you double in size, it's going to be the magnitude only increases over time. So actually today is the cheapest time in the history of the company to fix that defect. It's hard when I do all these great things, but you realize, well, here's something that's not impacting a ton of customers, but I need to go fix it. And we found, usually, when you go back and fix those, you find a lot of other causes or landmines that just hadn't happened yet. And you end up fixing more than you think. And Jeff has talked about having pride in operational excellence because a lot of the work that you do as an individual or as a team, no one else is going to see. The discipline has to come from within. And so when we interview people at Amazon, we find out, does that person have the... they want to get those little details. It doesn't matter if someone's standing over your shoulder, or if you're going to get a bonus at the end of the quarter, it's just, I need to get the details right, because that matters. And you all trace it back to the customer experience. If you ship something and it's late for the customer, you don't know what the impact is. It could not make a difference for some people, but it could be something very important that really has a material impact on the customer's day, week, or month or life in them. You have to put yourself in the customer's mind, you have to want to get the details right from within, and then you have to have specific processes at the company to enforce and be able to measure and enforce that discipline.
DC: I was just going to ask you about that, the enforcement part, because something that you mentioned which I've been fascinated with and implemented many times is this idea of the single- threaded team. And every time that I have, I've experienced and I've seen the success of it, but the minute that, it's almost like a garden, the minute that you turn away from it, we'd start growing and they want to, especially on the technical side and engineering side, they tend to want to come back together into a single unit. No matter how much you fight. And so you have to constantly be pruning and making sure that you keep this discipline. How did you enforce something like single- threaded teams?
Bill Carr: So this is a good example of where a lot of people over time have looked at these concepts as a trade- off, either you can have autonomy or you can have control. And Jeff was frequently, frequently I heard him say we should not have the tyranny of the or, it is possible to have low cost and quality. It is possible to have autonomy and control. And so there were a couple of mechanisms that were designed to do this. And so the first one in the early days for two- pizza teams, it was this idea of a fitness function where we recreate this set of metrics and create this hash of those metrics and be able to track what they were doing that way, but that didn't work. But it was an attempt at how do I maintain control of an autonomous body as a CEO and a senior leader, but what it evolved to was instead, this is why the whole, why we keep saying, it's not one thing, it's this whole machine. So one part of the machine was the written document process. So then later what happens is that autonomous teams would write their plan, their business plan, what's their operating plan. What am I going to do for the next 12 months? Here are the key initiatives that I'm going to work on. Here are the specific metrics and goals that I'm going to chase after. And then Jeff and the executive team would harvest from each one of those plans, a subset of those goals and make that something that they would personally track. The second part is then there's check- ins, there's accountability and this layer is up and down. So not every team, of course, Jeff checks in with, but for my team, in digital video, I could have a monthly business review with different subsets of my team. They would be autonomous and separate, but it didn't mean that I couldn't check in with them and they couldn't write up a review of what progress did they make and how are they doing, and I can't give guidance and feedback. And then another one of the mechanisms really was the PR FAQ process. So if you want to set a team off to go build something completely new, if you give them sort of just high level guidance and set them off and then check in with them in six months, the odds that they're really doing and exactly what you think that they should be doing are probably pretty low. So instead the mechanism was to spend a lot of time upfront defining what's the new product you're going to go build using this PR FAQ process and getting that right. And once you get it right, then you can say, great. Now the team could go off and run hard to go build this.
Colin Bryar: Yeah. I mean, that's a great answer. The only thing I would add is that there's no perfect org chart and org structure. So whatever your org structure is, you want to optimize for the strengths of that. But then if you want to share what people learn across the org, you have to just have some processes. If you want data scientists who are spread out through the company, to be able to share advances in machine learning or data science, you just have to put that in place. But it can be independent of the org charts. So sometimes people focus a little too much on the org chart. It's really the autonomy and control and capabilities of those single threaded leaders that mattered the most.
DC: You mentioned a couple of times, Bill, you mentioned the word mechanisms, can we double- click on the mechanisms, another area that I became obsessed with mechanisms or at Bridgewater Associates, Ray Dalio column machines, what is this idea of mechanisms at Amazon?
Colin Bryar: Well, so yeah, there's a saying at Amazon, " Good intentions don't work, mechanisms do." So when something goes wrong, one of the things that we asked at Amazon is, " Well, what caused that?" And you can go through the the five why process that we get from Toyota, but it's not, the answer we'll try harder or we're going to just make sure it doesn't happen again, that's an example of good intentions and those don't work mainly because people are already trying pretty darn hard and they had good intentions going into it. So you're not actually asking for any change in behavior. And usually what happens is you, well, first of all, you need to ask, well, why did this error or defect happen? And what do we need to change? What process or do we need to change in the company in order to prevent that from happening again? And so a good example is we had voice of the customer, voice of the seller, voice of the developer. A group would bring problems that are, one or more of our subsets or customers had. And they're very, very painful to read about how you let down those customers' lives that day. And they're not afraid to tell you, " Listen, this is what happened. My credit card was locked and I couldn't go buy medicine for my kids." And you listen to that and you think, I can't believe this happened. But then the next step is you say, okay, well, what specifically do we need to fix or change? It could be software. It could be a process, or it could be someone's not appropriately trained to handle the job. And so that's where it's not solved until you've actually implemented a specific mechanism to do so.
DC: What's a difference to you between a mechanism and a process in the way that most companies think about a process.
Bill Carr: I don't think that there's any difference. I think mechanism is just a different word for process. So the point is any time as we work with companies in the early stage, and they have less experience. To me, any time you hear a CEO or a leader complaining about XYZ problem keeps happening, or there's this problem. And the root cause of that is not probably some person or the org structure, it's simply that they don't have a process. They haven't gone upstream to think about, well, why is that really happening in the first place and what process or mechanism do I really need to put in place to prevent that from happening again?
Colin Bryar: And so it can be as simple as there's a very simple mechanism that was put in place to prevent a very large AWS outage from occurring. Again, a software administrator added another zero when they were taking some of the network routers offline into the percentage and it did some damage. So they've said, how can you prevent a single human from making this mistake? So they fixed that process. There are two people that look at it and they had an automated check A, are you sure you really want to take down 40% of the capacity in this data center and you need an approval to do so, but all the way up to where a CEO, a lot of times you get frustrated, why are people making these decisions? It's so clear to me that they should make a decision in a different way. And really where the process broke down is that CEO did not accurately give a set of guidelines and principles to the rest of the people in the company to make tough decisions with not a whole lot of information when they have to move fast and the CEO's not in the room. So that mechanism is having a well- defined codified set of leadership principles, and then stitching them into how you hire, how you measure things. So it can be as simple as before you hit the keystroke, get another person to look at it, or you the CEO may know exactly what needs to happen, but you actually haven't told everyone how to think and how to act when you're not standing over their shoulder or in the room. And so you need to put in a mechanism to make sure that everyone in the company operates from that same set of first principles.
DC: Did you have to fight this tension between the single thread in this, distributed model and ever having the feeling or hearing from people that we have too many processes, we have too many mechanisms, like it's hard to maintain some agility in the company?
Colin Bryar: All the time. And this is where as Bill mentioned, being stubborn on the vision and flexible on the details, sometimes you do have too much process. So Bill mentioned the NPI, New Project Initiative process we brought in from GE and modified it. We tried that because we wanted to prioritize a bunch of global and cross- functional projects, but we realized it was too heavyweight and we were listening to people and there is feedback. So we came up with something simpler, which evolved into two- pizza teams and single- threaded leaders. So yeah, no company is perfect and you always have to be curious and also willing to admit that you're wrong, that you made a mistake. It's actually quite liberating, if you can think about that, because then you can fix something and actually get better. If you say, no, this is the process. And you get too focused on the process versus what is the outcome we're trying to achieve. That's where your focus is the process itself, which isn't the outcome.
Bill Carr: Yeah, the way you can figure that out is if you were to parachute in and talk to middle managers about a certain process. And if they describe ways in which they basically are trying to game it to win it.
DC: That's exactly it.
Bill Carr: If they describe, oh, here are my tricks to win this process, then you know the processes is too much or wrong and needs to be edited.
DC: Oh my goodness. I can't tell you how many times all of us have had that happen, to come in, and they're just like this wild gamification that's happening. That's like, what is happening? This is craziness. So you have Amazon, you have what you learn there. I understand the leadership principles. And I understand that ethos. How do you localize a lot of these mechanisms for a new employee to come into a group, there's a whole set that has to cascade below the leadership principles. Where did those live? How did those get codified?
Bill Carr: So new people coming into Amazon, the more experienced they were, frankly, the more challenging it was for them to learn how to be successful and operate with an Amazon. Because the way we operated was so different than most companies, probably less, so if you were a software development engineer, but if you are a traditional" business person," it felt very jarring and different. When you started to have to write documents and you realized that a lot of decision- making in business- oriented companies is actually is about winning arguments and opinion and at Amazon, it was much more about truth seeking, sharing facts broadly and through a process of writing and discovery arriving at truth. And that is very different than a lot of companies and cultures. So what I found is that the longer I was at Amazon, the longer I stayed at Amazon and the more senior people I brought into the org, the more time I was spending coaching, teaching about our principles, about our processes and how to enable these people to become successful. That's partly why this is what we do now, because I realized that this is how I'm spending 80% of my time inside Amazon. And wouldn't it be cool if I can now help the next generation of business leaders a variety of companies outside of Amazon.
DC: Now that you've gone through this experience and you started working backwards, and now you're helping other companies. I would imagine by definition, all of those companies are smaller than Amazon. How do you, maybe they're not, but I would imagine they are, how do you translate some of the things that worked or the things that worked at Amazon that you think are core, how do you incorporate those into an existing company that may be smaller, maybe have a different set of constraints? How does that process work today for you?
Colin Bryar: Some of it depends on the stage the company is at. So we do work with some smaller companies, all the way up through public companies too. Quite honestly, we can't add that much as if you're still trying to chase market fit. But once you have something and you now want to add another zero to number of customers, employees, revenue, and essentially the biggest job of your career at this point. And those are multidimensional problems that we get involved with. So, first of, for those earlier stage companies, if you don't have leadership principles, if you haven't really codified who you are and how you make decisions, that is the first starting point. And then the second is, as you hire people, if you're going to go from 20 people to 100, or 100 to 500. If you don't have a deliberate process for vetting people to say, will this person reinforce what our culture is or where they detract from it, they're probably going to detract from it. So when you go from a 100 to 500, you'll look back year later, and you'll say, this company is just not what it used to be. It's not what we're about. And really the cause of that was that you didn't have that deliberate. You didn't have the very clear definition and then a deliberate process to go attract people who reinforce those who you want to be, because each company is unique. And so that's the first part for smaller ones, but for larger organizations, it's really, where are their pain points. These processes we described were very specific solutions to very specific problems that companies have. So we don't go in and try to make every company like Amazon. We try to understand what's going well at the company, so we don't mess with it. And you don't tinker with it. If that's going well, don't break it and don't mess with it, but then figure out what are the key problems. And quite honestly, some of them are, they sound simple, but they're hard. As a new executive, how should I be spending my day as a CEO? I just got a big series defunding and I am overwhelmed with meetings and I don't know where to focus my time. And that's probably one of the most important decisions a CEO can make, is how they want to allocate their time. So anywhere from decisions like that to how can I get in a good operating cadence so I know what my customer's experience was last week? And was it better than the week before? And are we on path to improve the customer experience? We're working on the right controllable input metrics to generate what we want to accomplish as a business.
DC: How much of the time for you taking on a new client now working backwards is spent vetting the leader, the CEO, whoever it is that you're working with, to make sure that they can maintain a longterm vision, that they can have this level of discipline, that they won't just revert back to whatever state or whatever practices or whatever things that weren't working in the past. To me, it seems like that would have to be a critical part, for inaudible able to have an effect or to implement change in an organization. Is that something you spend a lot of time thinking about?
Bill Carr: I think that's a great question. And one of the ways that we approach this is that we, in many cases, we start off with a company where we focus on working with them for say a window of three months to get started so that we both get to know each other. And that way we can figure out if there's a great long- term from their point of view, are we going to be helpful to them for the longterm and vice versa? Do we think that we can be helpful to them? Or is there a good fit and a good match?
DC: It seems like it would be a hard one. I think one of the things I always repeat is simple, not easy. Most things are simple, but not easy. Like on the face of it seems simple to implement some of the things that we're talking about because they're logical, they're rational, you can understand them, but it's human behavior to kind of revert or to go back to consensus thinking or go back to whatever practice that you have.
Bill Carr: Yeah. It's no different like a marathon, it's pretty straightforward. I can give anyone a training plan for how to do it. And the first week is probably pretty easy for everyone. The hard part is week six or five when it hurts. And there's like, you're tired and you have to go do that run today and can you make yourself do it? And that's, to me is the definition of discipline is that you can find that mental discipline to do it.
Colin Bryar: Yeah. And it appears in all the processes. I just interviewed someone, I need to write a very thorough set of feedback to my peer. So I can evaluate whether this person is a great candidate for the company, but I've got another three meetings coming up. How do I find the time to write this feedback? It's conceptually easy, but you just have to devote the time and the focus on that, because it's important. I just picked that as one random area, but that's an example of just the discipline and the pride in your operational excellence. I should get the details right. If you have to keep pounding people with that, then that's probably not the process for your company. There are different ways to build great companies. And we think this is worthy of study and for the companies that do want to adopt this, we're happy to work with them, but if they don't, that's fine too if they want to go another route. The only thing I would encourage is pick a method. That's probably the most important thing. So even if we don't end up working with the longterm, we encourage them to, well, here's something that you can prevent pick on how you want to do this and focus on what that solution is.
DC: With the companies you work with now and given your experience, how do you think about hiring now? It seems like so much of it had to be baked into, for it to scale, into the hiring process. I don't know exactly what techniques you use or how you kept that. I mean, I know of the concepts of, and I've read a million things about bar raisers and things like that, but how do you think about advising companies now around their hiring practice? One thing is obviously forcing detailed reviews. Like you mentioned Colin, but how else do you think about it when you advise them?
Bill Carr: Well, this is one of the parts that I think is in fact, one of the most straightforward and easy ones to apply. And there's really two parts to it. Part one is actually establishing real leadership principles for your company that are well- crafted, valid that really reflect where you are as a company, in addition to a bit of where you want to go and that they really reflect and define what does role model leadership look like inside your company, and what are the lenses through which you should make decisions in your company? So one of the things that people would probably early in their career really don't get is how different companies are, and that a single individual that can be wildly successful at a company like Google could be spectacularly unsuccessful to a company like Starbucks, or maybe even Amazon. It depends so much, and that finding the job search process is an important two- way thing. It's just as important for the candidate to find the company where they will be successful as it is for the employer. And too many people focus on just functional skills. So that's sort of the minimum, and those are actually not that difficult to ascertain for the most part. The more challenging part to ascertain is the way someone will do work and how they operate and how they lead. Does that fit with the principles and structure of our company versus other company, because each one is so different. So that's one where we spend time with companies first seeking to define and articulate their leadership principles. And then step two is to change the hiring process, that at the very least, the one change you can make your hiring process is to start to conduct interviews where the criteria are objective, and everyone is studying the same criteria, which are, how does this candidate meet our bar for our leadership principles? Would they be a role model leader here based on the way they've led and acted in the past. I used the word leader, by the way. Everyone is a leader. An individual contributor coming out of college their first job is a leader as is the CEO and everyone in between. And so the trick in an interview is to find not just a functional match, but for lack of better term, a personality match in terms of the personality of that leader in terms of how they work.
DC: This is an area I'm so fascinated on. I felt like for all of my hiring in the past, that really, there were only two dimensions that I would evaluate on and that most people would. There's the one that you mentioned, which is the quantitative stuff. Then that's easy. The hard skills, and it's very easy to test for hard skills, especially the closer you get to engineering or those kinds of disciplines. And then there's the other one that kind of qualitative of like the qualitative kind of stuff. But then there was this like hidden third dimension. In the way that you described, how close are they and how do they make decisions? How close are they to the leadership principles of this company versus the context that they may have operated in and been successful in other companies? And I thought this third part was actually the most critical part. That it's very easy to look at quantitative stuff. It's very easy to get romanticize on the qualitative stuff about a person and how they fit within a team and are they" a cultural fit" and all those kinds of things, but there was this context part that was actually the most important part. And so it sounds like for you it's first starts with the leadership principles. Defining those, making sure you use the word, you emphasize the word real, which is an important thing in leadership principles. Are they real? And then that's how you're backfeeding, that's how you're testing whether someone would be a great leader within this specific context versus the others they operated in. Is that correct?
Colin Bryar: Yeah. And you have to know how to test going into the interview. You know here are the two leadership principles. I am responsible for vetting the candidate. And I know going into the interview, what are the sets of questions I'm going to ask and have they experienced to know when to probe deeper if I didn't get an answer that didn't yield the right information that I need. So yeah, when people say a good culture fit, I usually ask, what do you mean? Because that is such, it could be a vacuous statement or there could be a lot of information you can unpack behind that, but a culture fit in and of itself doesn't give any information to the rest of the interviewing panel. You could say, this person is customer obsessed in this way, or long- term focus or that inventive because they created these three things outside of their day job. So I know that they have an inventive, curious spirit. And that rolls up into, are they a good culture fit? But you have to go down to that next level of detail and make a deliberate assessment in order to come up with that conclusion.
Bill Carr: Yeah, like a common mistake I've encountered when people are trying to test for cultural fit is they'll ask a question about like, " Why do you want to work at this company," which is not the worst question in the world to ask. By the way, it should only be asked once by the recruiter or the hiring manager, well in advance of actually bringing them in- house and it should never be asked by every single person in the loop in house. It's a dumb waste of time. Why would you have six people all ask the same question. And number two, they'll have one canned answer which can be based off of their clever study of the website or talking to a few people that then will sound, " Oh, they really understand our company and our culture," but the fact that they've been able to rehearse it, doesn't tell you if they're actually a good culture fit. The only way can figure out if they're a good culture fit is to then use behavioral interviewing techniques to get examples from their past work to find out well, how have they worked in the past? And how does that relate to our leadership principles? The other area where companies make mistakes is that I see, they haven't really actually codified or agreed on what their culture is. So if you have six different people interview the candidate, they'll have six different ideas of what their company's culture is in the first place. So not only do they not have agreement on what the culture is, but then they don't actually have the right tools to figure out whether the person is a match for their culture.
DC: And that started by auditing and just being maniacal on the interview process itself, and making sure that you had a system to ensure that or a mechanism to use your terms, that you couldn't repeat questions?
Bill Carr: To be completely literal, the interview process at Amazon, if you come in and you have six interviewers, there are 14 leadership principles. Each interviewer will be assigned two or three of the leadership principles, and they will interview the candidate about those principles and virtually nothing else because that's their only job is to determine, does this candidate meet our bar for those principles? And by the way, they've gone through specific training that helps them A, understand, well, there are many mechanisms to help people understand what the Amazon leadership principles in great detail, A, so what do they mean? And B, how do I actually interview someone to find out whether they are a fit for that or not? And then the debrief meeting that happens, which is they have to take notes and document the interview, which is not typical most companies. They have to attend a debrief meeting every time, which is also not typical. And in the debrief meeting, the objective criteria that are discussed are okay, we're going to decide about this candidate based on how they stack up on all 14 of the leadership principles. And now we have objective data because each one of you had specific data to go collect. We're going to put it all together, look at it and make a decision based on those data.
DC: And is there a mechanism in that meeting to prevent consensus thinking, of course you have the objective data, people to talk themselves into wanting to hire this because they heard Colin's great. You know example of how they embodied principle seven, how do you prevent that from happening? Is that where the concept of the bar raisers came?
Colin Bryar: That is the bar raisers job is to run the debrief meeting and to help the hiring manager make the right decision. And if you abstract this process, what we tried to do with bar raisers was take the person, the individual person out of the process and make sure that if you put trained people in, it still is going to work, and it's going to get better the more you do that, because there are some people quite honestly, who are just inherently great interviewers and with a small company, it's easy to say, okay these three people are going to interview the candidates coming in because they're great at assessing talent. Well, there are 168 hours in the week. And if you want to hire 100 people, you can't rely on those three people. So you have to take the person out of that process and make sure that you have the appropriate training checks and balances. And then feedback loops that go in there. And a great feedback loop in the bar raiser process is that debrief meeting, because you get to not only assess the candidate, but you figure out how did you get this information out of the candidate? I was trying, and I couldn't find that. What questions did you ask? A lot of that happens at the debrief meeting. Some companies when they make the hiring decision, they'll just read, it'll be the hiring manager and someone from HR or the recruiter, just reading the feedback. If there is written feedback or looking at the votes, and then the hiring manager decides do I want to hire this person. This debrief meeting is an important feedback loop. Everyone's a better interviewer coming out of each and every debrief, because I learned something or I teach something in pretty much every debrief. So the more that you go through this process, the better it gets. It's scalable because you just keep the ratio of bar raisers to the number of interviews and candidates in lock step so you can add more bar raisers. And the other thing is we tried this in one department at Amazon, it was the product development group. This is in 1999, there are 20 initial bar raisers and it worked so well that other groups just started to adopt it. And that's the sign of another good processes. It's not a stick approach. It's a carrot approach because quite honestly, the hiring managers realized this is going to help me build my team faster and attract higher quality people. And so it just started to spread through the company. And then we eventually mandated for all the salaried positions. It is required, but that was a noncontroversial decision at that point, because we said, yeah, it's great. So we should just instrument our recruiting pipeline and hiring process to just implement bar raisers. So there are a couple of things, I think notable lessons above and beyond recruiting about the bar raiser process itself and how and why Amazon adopt some of these scalable repeatable processes that have this nice fractal quality. The bar raisers works, or 15 person company. And it works. Amazon has 1. 3 million employees as of last week. It's probably a lot higher. So it works for that company size too. And it's the same process.
DC: Is that an important tool that you bring to companies that you work with, this idea of the bar raiser or something equivalent?
Bill Carr: It is if they want it. We are not salespeople.
DC: You're not spreading gospel.
Bill Carr: We are, if they have read our gospel and they say, "I want that."
DC: And I want to adopt you gospel.
Bill Carr: Then you say, " Great, I want to adopt your gospel," then great.
Colin Bryar: This is getting too weird for me.
DC: I have a jar of Kool- Aid here.
Bill Carr: But we're not missionaries.
Colin Bryar: I'm joking.
DC: Yeah. That's amazing. Well, it's been fascinating talking to both of you. I could talk to you forever, but you two would be bored talking to me all day. Where can people follow you online? I'm fascinated by your work, fascinated by your book. I will gift this book to many, many people this year and in subsequent years, thank you for this great book.
Colin Bryar: Well, and thank you for having us on the podcast. People can find more about what we do, it's just workingbackwards. com all one word. And so that's our website where we talk a little bit more about the concepts in the book, what we do, and we're putting more and more content up to supplement what's in the book that's coming up, but we're still working on that too.
Bill Carr: I'm going to go fill out that form. I'm going to be the first one there before anyone listens to this podcast, and make sure you work with drift before anybody else.
DC: Awesome, thank you both. And if you liked this, this episode was groundbreaking. The universe is only six star podcasts, but this was an eight star worthy episode. Thank you, Bill, and thank you Colin, for joining us today. Cheers. Thanks. Thanks guys. Let me know what you thought of this episode by texting me at 1- 212- 380- 1036. Again, 1- 212- 380- 1036. Now, if you're looking for more leadership insights, sign up for my weekly newsletter, The One Thing at drift. com/ dc. Every week, I'll share a habit tool or mental model that's helping me reach my goals. Hope to see you there. Text me, hit me up.
There's a saying at Amazon: "Good intentions don't work. Mechanisms do." For well over a decade, Colin Bryar and Bill Carr worked in tandem with Jeff Bezos, building out the mechanisms and leadership principles that shaped the company into what it is today. How'd they do it? They started with a relentless focus on the customer – and worked backwards from there.
Colin and Bill were in the trenches for the creation of some of Amazon's most innovative products and services including Kindle, Amazon Prime, Amazon Studios, and Amazon Web Services. And on today's episode of Seeking Wisdom, they sit down with DC and talk about what Amazon was really like in the early days, the development of critical processes like the PR FAQ, single-threaded teams, and their best practices for hiring the right talent.