Sustained business growth with high-velocity decision-making

Dan Slater: Well, welcome everybody. Thank you so much for being here. Um just to introduce myself, I'm Dan Slater. I oversee our Culture of Innovation programming as part of our Innovation and Transformation program team at AWS.

Kayle Siegel: Hi, everybody. I am Kayle Siegel. I'm head of the Americas region for the same group Innovation and Transformation Programs. What I tell my parents I do is I'm a professional brains stormer with Amazon's toolbox. So I kind of go into external companies and use AMI asan mechanisms and methodologies to help them build out new digital products on the cloud and then build them on AWS. So, so excited to see such a great crowd. I can barely see you guys but looks like a good crew.

Dan: Today, we're going to go over some of the ways that we think about sustaining business growth, especially in this day and age where that's on everybody's mind, both because of headwinds that you find external to your business, but also the need to really build sustained business growth. It's always important. I don't think it's lost on anybody in this room that's speed matters in business. In fact, andy jasse at re invent two years ago, brought this up as one of the core things that companies should really focus on in that speed matters inordinately in today's day and age and you really need to have a culture that builds up that urgency and be doing it every day. It's like a muscle that you need to be exercising. Um but it's not easy with the ever increasing pace of business, with the ever increasing needs of your own internal business. You really have to deliberately make speed a choice every single day and given that, that's kind of the day and age that we live in. I would, we would love to kind of throw this open to you and ask you a couple of open ended questions about how you think about sustaining business growth in the speed at which you're driving that within your own organization.

So maybe the first question we'd like to ask and we'll keep it very scrappy and low fidelity. If the answer is yes, just raise your hand. And so the first question we have for you is do you believe that your team, your organization needs to increase the speed to market in this coming 12 months. The answer is yes. Raise your hand looks pretty unanimous. Nice. If anyone's here saying my boss said we should slow down this year, you're probably in the wrong place.

Next question. Uh my organization excels at decision making. Don't worry, i can't read your badges. So a little less hands up right now, which is kind of, I think normal and we'll talk about that in a second. But the last question we have is do you think that you or your organization spends the right amount of time deliberating the decisions that you need strategic, operational and otherwise, do you think you're spending the right amount of time? Raise your hand if you think that's, yes. See a few hands.

All right. Well, for those of you who didn't have their hands up in the last two questions that we asked, don't worry, you're not alone. We struggle with that in our own business every day, making good decisions at speed and scale is critical, but it's really tough. Um according to a mckenzie survey, a global survey that they did a few years ago, they really found that 20% of people surveyed feel that they really excel at decision making. So not a lot. But at the same time, these people, and I'm sure this is true for you spend over a third of their day, 37% of their time making decisions. So if you weigh those, like spending a lot of time, how do you become better at making fast and high quality decisions? And in this data set, it was a global data set, 61% of those polled said that they consider that the time that they're spending, making decisions is inefficient and there's a real cost to that as well. Mckenzie had put a cost for this and looked at it and came up with about 530,000 lost working days for the average fortune 500 company lost to inefficient decision making, which translates to about $250 million in wasted labor costs in a given year. And the c suites that answered this poll also recognized that they weren't making decisions efficiently, either. 64% said that they, they could be delegating more of their decisions that they were getting called into areas where their team, their organization should have been making these decisions faster.

So we're going to dig into a little bit more about how you achieve, not just the speed but the quality in decision making. Some of the things that we've at aws have found that has worked for us and we do realize that every organization is different. So we're not going to share things with the and that lens that what we do is the best rather hopefully by sharing how we try to balance this every day, that could spark some ideas about how this operates in your own organization and where you can apply things to help you make better, faster quality and faster and higher quality decisions.

The first thing we'll really touch upon is around the mental models and mechanisms that allow people and teams within a corporate culture to do that, to move very quickly. And also to make those high quality decisions. And a mental model is something that we all carry every day. It's the filters for information that we as human beings use to get to a quick decision. We process an inordinate amount of information all of the time. And mental models help serve as a shortcut to get us to a fast answer. And oftentimes you'll find that the mental model that you bring yourself or your organization brings are very efficient and get you to a decision. But also you tend to fall back on that method of decision making because it's worked, it's natural to do that. But that also creates some blind spots. If you're only making decisions in the same, involving the same parties of people, you start to lose out on the diversity of ideas that come out or you may just be blind to other more efficient ways. So this is more of the theoretical. Um maybe kayleigh you can help us ground us in what this looks like in an actual corporate culture.

Kayle: Yeah, absolutely. So we kind of take this simple equation here. Mental models like Dan was talking about plus the behavior that we all, you know, um uh lead every day equals kind of the, the artifacts, the things that are kind of the tangible results, um the processes and the other documentation and the decision based on that output. I tend to think about it a little bit where when these pieces are aligned as leaders within your company. You don't have to be really that micromanager participating in every single line decision. I kind of joke, I'm only 5 ft tall. So I'm like, I'd think a great micro manager in that way. But in actuality, you know, you really want to make sure that you are empowering your teams with the best possible attributes to be able to move those things forward.

So in thinking about it like this, these specific mental models and behaviors, you know, it's, it's great to think about the ideal state. Everybody on your team is rowing in the same direction. Uh there's, you know, enough space between the way that everyone is operating. So they're not banging into each other with their elbows. But in reality, to get the most out of your team members, um you want to be able to flow like this, but this is not reality. This is probably slightly closer to reality. I'm, i think i'm the orange one. I don't know what color would you like a general direction but still orthogonal, right? Like trying to move fast.

So, so as we think about just the visual representations of how teams operate in the wild. The idea kind of comes to, well, how can we get everyone kinda on the same page and uh have a unifying view towards that common goal without actually going down into the weeds on every decision and telling everybody exactly what they should do and how they should do it. You're sacrificing the benefit of having a team in that way.

So, fundamentally, how do we make sure we're delivering on the right things and considering that we would like to orient the team in this particular direction. So this first approach um is kind of beginning to put up a specific filter and team members are encouraged to bring their ideas forward, but there's this really small hole that's properly aligned with one prior one prior arrow. And what do you know only that one goes through? Um and what you get in a model like this is actually a lot of disenchanted, disenfranchised employees and then that one star who made it swimmingly um through the obstacle course.

So in thinking about it in that way, then how do we have a more scalable approach that is inclusive and is able to um capture the ideas of of many and then be able to orient in a particular direction that makes the most sense. But what we found fundamentally is a mental model that is inclusive and is a policy that then is built by the specific team members so that there can be the the confusion at the beginning, the opening of the aperture in the center that allows for great discourse, the building of policies kind of together and then that orientation can persist um as things move forward to kind of encourage everyone to think about uh their particular uh filtering systems and how to begin to um not tell everybody exactly what to do and how to do it, but to be as inclusive as possible. Passing down those um best practices, the specific mental models, I can think about one for my team, which is stay as busy as possible on the best possible opportunities. And that allows the team members to make those decisions once they have the definitions behind them.

So in understanding that you have your teams all rowing in the same direction. One of the great ways to illustrate that is amazon's uh retail growth flywheel. And well, it looks like a pretty unique napkin sketch type drawing. I like to think about it this way. Never has a customer ever said, I'd like to pay more for something. Never has a customer ever said, I'd like there to be lower selection and the last one never has a customer ever said, I'd like this to be less convenient. I'd like it to arrive slower. So in this specific flywheel mechanism, as we begin to talk about it, these are the areas that amazon and jeff bezos, according to the law of the napkin sketch illustrated to say these are the durable needs and these actual areas are the places where we can allow our teams kind of full access and freedom to be able to move forward in the areas that would help lower prices increase selection, improve convenience. And so beginning to think about how the particular flywheel works and what are the areas that are then those durable needs for your customers? The things you can make big bets on to be able to push that decision down to the edges. Any thoughts on the fly wheel?

Dan: Well, no, i think it's a really great organizational principle that every business at amazon has because as you focus on continuously improving the customer experience, the things that are durable in at least the retail model here for amazon's retail model are built into that, right? You are building more traffic as you are focused on a great experience, you are bringing more sellers onto your platform which helps bring more convenience, often, sometimes price and broadens your selection. But as you do, so you notice that there's another loop here that as you do, so we lower our cost structure, we achieve economies of scale and as we do that, we can sink the savings into that right back into another one of these long term durable needs, which is price. So the energy and effort that you put around doing this is going to not just make that fly wheel spin faster, but it's going to remain durable. Because as kayleigh as you, you know, well articulated, these are things that are never going to change for customers in a retail environment. Absolutely

So when we think about that growth at the center, um how might you actually build that specific flywheel? So like Dan mentioned, focusing on the long term goals or aspirations, and I could say we're gonna go through a few customers specific flywheels, but think about the specific industries that you are in, you know, uh any bank that would be like like to be less secure, probably not, you know, any farmer who would like to harvest slower or have a dirtier sample or something like that. You know, you can begin to say understanding those longer term goals or aspirations, focusing on the durable needs of the customer and the business with a specific nod to the horizon of where your company might be going.

You know, I think there's a lot of great things in the idea of like selection, but it doesn't say selection of books. So being having it be broad enough uh to be inclusive of where, you know, your company is going, but maybe where you don't know your company is going uh linking those drivers together. So really understanding the different attributes that are above the flywheel that drive what is most important to you and your teams at the center. And over time, that idea becomes that specifically that well oiled machine and it translates into predictable and profitable business outcomes.

So, like I mentioned, um Amazon.com has a retail flywheel that we went through uh price selection and convenience, hard to kind of argue with that. But I had a great experience recently in working with a company, uh a Brazilian um home goods uh store, essentially the IKEA of Brazil if you were to say it simply, and this was their fly well that they worked on, but it's also great to be able to see all types of customers, not just ones that have uh pretty similar business models, but like I mentioned, you could see this happening at a health care company, an automotive company, financial services, et cetera.

I'm going to pass it over to Dan to talk through this next section as we think about the customer value.

Yeah, because I think one of the things as Kayleigh mentioned, we work with companies all the time to really think through how you can think through a durable flywheel. I would argue that one of the things that are most durable in any of those flywheels is that section on the customer experience. And if you can extrapolate all of the different flywheels that we've built in our businesses, Amazon and AWS those that we've helped other customers think through that is the common denominator, this customer value flywheel. They are all laser focused on bringing quick and incredible value to their customers. And they do this by really thinking innovatively and experimenting and getting new features and new offerings out to their customers based off of their customers' needs. By doing that, they will look at the option and hopefully increase it. But as they do so you will know from your customers what they like about your service. If you've instrumented the right metrics and the right voice of your customer, you will get really helpful feedback and maybe even clues as to what new use cases you should be thinking about to not just build new features but potentially new models around or additional fly wheels. By doing that, you're going to be constantly experimenting and iterating and driving that customer experience internally for your own organization by putting this rigor around constant experimentation and iteration inspection. What tends to happen is you're empowering your teams to stay close to their customers and move fast. And by doing that helps reduce the time to value for any given product line.

So this really any flywheel could be thought of as a general customer value fly wheel. This is all well and good. Like how do you actually put this into practice boots on the ground in your business? Well, one way we think about it at Amazon and AWS is translating that customer value flywheel into a specific mechanism. A mechanism is something that has a very specific meaning within Amazon and AWS. It really is a complete process, a virtuous cycle that much a flywheel is you operate it, it will spin faster and it does exactly what I had talked about, you will build a tool, you will release an offering and you will want to drive adoption. You want to instrument the right inspection of how customers are using it. What are they saying about it? What works, what doesn't and constantly iterate and that helps turn your inputs into the outputs for your business that you want to see. It helps translate intentions into ideas. And you can think of a mechanism like it could be even a business meeting where you have a very structured way of looking at the data, looking at asking those open ended questions around, how do we increase the speed? How do we increase the growth based off of what our customers need? And by iterating and constantly experimenting, you'll be able to get to that value proposition for them that much faster.

A little bit about inputs and outputs because I'm sure like you, you're all very data focused as we are, as we think about inputs and outputs to the business. Those need inspection and even updating and even some time experimenting around um the outputs of of a any mechanism. Are those things that are probably things that you're used to in order to drive the top line growth of your business, things like sales or revenue or subscription numbers. And those are very important to ensure that they are the demonstrable aspects that show your business is growing in the way that you want to. But you also need to examine the inputs of your business. Some of the things that help drive sometimes the bottom line. Some of those things that within your organization, you might have firmer hands on rudders around firmer troll over some of the inputs like productivity or time or cost, certainly not full control. But these may be things that are a little bit closer to your business and you need to think about as your business grows as your customer needs change. Are your outputs still relevant? Are your inputs still relevant?

A really quick example, from my own experience, I helped launch the Kindle device and our worldwide business. And over time, our outputs were very similar to what you see here. They were things like how many units sold, books, sold, the revenue that we were getting the subscription sales over time as we were inspecting our business, we had a mechanism where monthly we would sit down to really not just look at if we were hitting our numbers, but do these inputs and outputs still make sense to how we wanted to think through customer value. And at one point, we realized that, you know, we're not thinking big enough just by focusing on some of these outputs, even some of the inputs of our business. We realized that in order to truly get to growth and new areas of the business, it wasn't just about selling book units, it was about getting people to read more because maybe you are all readers. But if you are, you know, your attention, there's a lot of competition for it from not just other forms of media, music, streaming video, but even from news, your email, your busy day to day lives, if we could find a way of thinking through, how do we get people just reading more that really transformed and pivoted how we thought about the inputs of our business.

So we started to instrument things that did measure the time that people were reading. And that total changed our road map in terms of the features and discovery and personal, the personalization that we put behind some of the features so that we were stay closer to what we really thought. Our think big moment for this business should be, which wasn't just selling units. So just a little background about how you do need to think through not just the outputs of your business, but the inputs and inspect them frequently.

And I mentioned building a voice of customer. I did want to also mention that one of the crucial things that you need to do in your business is not just look at your performance data, but really build instrumentation around the voice of customer. This isn't just around sentiment and users, but how are you as a business as a team organization? How do you meet with your customers? How are you making sure that you're on top of what they need and how that changes? How do you stay on top of the things that your customers often don't tell you they need by building this voice of customer into the and using the same rigor that you do around that instrumentation as you do around your business instrumentation. That is a great start. And that's what we try to do at Amazon and AWS. We bring in not just our business data but our customer from our data into our business meeting and try to make it transparent and look for areas where they start to diverge and by divergence. I mean, you may be looking in a business meeting of your metrics clicking up and to the right like you need them to do. But you may from your voice of customer bring in an anecdote from a customer that says there's something about your service that's fundamentally broken for me. And it's very easy in those moments to look at that anecdote and look at the data and dismiss the anecdote as an outlier. But when you have both, when you're inspecting both, it starts helping you to question what's going on here in this divergence. And at Amazon and AWS where we see that we actually want to focus more on the anecdote and dive in and maybe we do prove out that it is an outlier. But oftentimes we do find that there's a use case that we've missed or there's a fundamental issue in our business that are aggregate data isn't catching. And by looking at the anecdote and peeling away those layers and inspecting it, we're able to solve that issue permanently before it becomes a networked one across our business. So again, like Kayla, I'll turn it to you to really talk a little bit more about like how, how do we do this with customers? Um what does this look like in actuality?

Yeah, absolutely. So I love the blend of data and anecdotes and really being a customer of your own business doing ethnographic research, spending time, literally or figuratively in the field uh as a naive expert or new user of whatever this new thing might be really is invaluable. Obviously, you want to expand that beyond, you know, just the focus group of one, but it is a really important thing to begin to look around corners on behalf of your customer. So as we think about making better, faster decisions, I wanted to share a little bit about some of the work we did with a customer in the southeast of the US Agco, they're an industrial farm equipment manufacturer. They make a bunch of different great things like smart tractors, connected silos, connected coups, et cetera. So in the projects that I got to work on with them, I learned more about uh uh smart harvesting combines and bushels per minute than I ever thought I would. But it's one of the benefits of getting to get basically a master's degree after every project.

So when I was working with Adco, one of the things we focused on was the creation of this autonomous robot, think of it as like a cooler sized robot that could operate autonomous, autonomously with like a mechanical weeding device, like almost like scissors attached to it. So pretty cool, pretty revolutionary. There's a labor shortage, people don't want to do these particular jobs. They could be highly automated, but this could be any product at the center that really is game changing in a business. And in order to bring these new capabilities to life, autonomy specific mechanisms, batteries, uh it really needed an underlayment for that autonomous platform. And when we begin to think about the durable needs of the smart farming company of the future, I don't think anybody is saying we're gonna go for less autonomy, right? So understanding that this is an investment that is quite significant for the first right thing, this autonomous robot. But what it does is set the conditions for success for a number of different products and services that would actually harness the power of that underlying platform.

So we're really beginning to think, you know, this this platform itself could be a customer data c 360 type of platform with the first right thing in the center. But setting the conditions for how many other things you know, you might like to build and the great thing is sure it's really complicated to build the first thing, but hopefully you'll have then fertile ground for great new ideas and the platform will still be right for the ideas that you don't yet know about. Again, we're not going to get less autonomous in farming. So you can say, oh, we're going to make a big bet and i don't know what that vehicle of the future might be like. So to kind of say it simply in building out a platform like this and being able to test the products and services pretty quickly bringing value propositions to life kind of faster than ever. It helps to accelerate your time to know as in know, I know this is a great idea. I'm sorry about that or alternately no and no as in no kill it, fast, kill it often and really make sure that you're not wasting time on those incorrect decisions, kind of boondoggle projects that really should have been shelved a long time ago and having the data behind that and kind of the real world evidence based on the creation of a net new thing tends to help allay the concerns and um mitigate the slow death of products because you do really have a great product or service either out in the market and, and they'll tell you if it's a good idea or not.

So I would encourage everyone to think about how you can accelerate your time to know that it's a good thing or no, and you should stop work uh sooner rather than later. I'm going to pass it back to Dan actually to think about how we can make this happen with the small autonomous teams that could operate things like a platform.

I really like that mental model though. The no and no, like, you know, getting to realization of how you can scale quick when things are working or pivot when things aren't earlier in the development cycle is crucial and sometimes very hard to do organizationally. And the way we think through how to support this type of innovation, this constant experimentation and iteration from an organizational standpoint is through our Amazon's way of creating two pizza teams. If you haven't heard of that structure, what are two pizza is fundamentally is that no team at Amazon or AWS should be big enough that two large American style pizzas that should be enough to feed the entire team. That's the mental model behind how large this team should be, think about 10 or less people. And that just makes a lot of sense. There's a lot of common sense because it starts to drive down the lines of communication, right? As teams grow and swell decision making may start to bog down the process behind decision making might start to bog down itself. So keeping teams small really helps with that quick autonomous action.

But one aspect that I think is crucial, even more important than just the size of teams is the ownership and autonomy that you can lend to a team by asking a team to own their business and by owning it. I mean, they're not just responsible for gathering the customer data coming up with the great feature or new offering and launching it. They need to own that feature or product end to end through its entire life cycle. And we don't ask these teams to solve multiple different customer problems or run multiple different services because that alone starts to bog down decision making it crowds road maps. So what we want two pizza teams to be are single threaded focused on one service and the one customer segment that uses them by focusing their attention. Not only does that help with speed, it puts that organization, that team really in the driver's seat of being responsible, having the responsibility of knowing their customer very, very intimately and being able to then move fast on their behalf and surprise and delight them every day.

So that's really overall what a how we think through keeping teams small and as a service grows, you may find that you need more people. And in those aspects, what we try to do is go through a mitosis if you will - split like a cell - and keep maybe a one large two pizza team now on two different areas, again, focus maybe each on a subset of the offerings, maybe it is focused on operations versus, you know, new features. Maybe it's geographical distribution, distribution and focus. It all depends on the business.

But by doing, going through that mitosis that helps retain that single threaded focus, not just for your business but on the customers that use it. And so what we ask these teams to do the operating model that we want them to follow is what at Amazon, if you've ever talked to an Amazonian, you probably heard us mention our leadership principles. These are a set of 16 principles that really teach anyone at Amazon, whether you've been there for a day or been there for 20 years, how to operate, how to weigh different things in the decision so that you can get to a high quality, high speed decision faster.

One of those tenets is bias for action. Remember what I said at the beginning or I should say I give the credit to Andy Jasse - speed matters in business. We have a leadership principle around that called bias for action and the mental model we have behind that is that you will have tons of ideas. Your customers will be asking for tons of things from you. How do you decide where or when to move fast? We have a really helpful mental model around focusing on that and that's that many of the decisions that you're making can be reversible. They don't need extensive study and you can value calculated risk taking.

So Kayleigh, why don't you talk through? Like, what is the kind of mental model behind it that allows us to do that?

Yeah. So I, I love this one. It's a one way or two way door. Excuse me, I think it went forward one. There we go. Um so as we think about it, you know, I in life or work, there's a number of different examples. You could give up a one way or two way door. And the basic math behind this is if you're 70% confident, you should just do it, you should just make it happen. You know, more extensive study is not going to give you more proof points. If you, if you have that, it's not really an experiment, but beginning to think um about how uh these two way door decisions can be pushed to the edges of your organization. And that one way doors are things that need more heavier consideration that should have greater study that maybe you'd be more comfortable in that 75 to 90% confidence rate uh whether you have of budget allocations. And so and so can sign off at a certain amount. But in those one way and two way doors, as you begin to articulate what that means to you either as an individual or as your company, it allows many, many people to be able to move forward with alacrity to say that indeed, they are confident in this level of decision to be able to bring it to life.

So, we have a little interactive exercise. Can I go through here? Um, let's see. Should we, should i quiz Dan or quiz you guys? Let's do it. All right. Building a proof of concept prototype. Do you think that's a one way or two way door? Anybody, if you think it's a one way door hold up like this gentleman is doing with a one way, one finger or two fingers. Is it a one way door decision that takes a lot of study and time or a two way door? I'm seeing, we're seeing that there are a lot of twos, right? Anybody think it's a one way door decision? Nope. What do you know two way door? Absolutely. Like the power of experimentation is absolutely a two way door. It's great to be able to put those uh capabilities in the hands of all members of two pizza teams to be able to build measure and learn as they move forward with just about anything.

I think the word prototype really maybe clued people off. But my question is, are there any one way to door decisions as you think about launching a prototype? What happens if the prototype doesn't work and you need to roll it back or what happens to the customer set that is using it? Maybe it isn't at scale, but you still have to think about what their off ramp might be. And having that articulated and thought through and inspecting that as your prototype is working is instrumental, right?

So there could be around the peripheral a few one way door but purely like a proof of concept prototype way all the way, run fast, for sure, for sure. Next question here is uh reducing prices of an ec2 instance. So this was an interesting one. I i liked how you raised your hands with the peace sign or the number one. So we can kind of see, do you think reducing prices is a one way or two way door? So I see a ones i see two. Ok. Ok. Interesting kind of a kind of a split decision here. Um but yeah, reducing prices of an ec2 instance really is that one way door? And when you think about it, I think since it's in in 2006, the number of price reductions, what was it? 115 i think in the notes now. 120. Ok. There you go. 120 ish price reductions, you can tell we're pretty serious about maintaining that trend.

And when you begin to think about, you know, what does it do um for for customer comfort or to be able to understand where is that particular services place in business? How does it reflect the ec2 brand itself? Like this is, this is a decision that, you know, from a business and cultural perspective, you would not, it would do significant harm to go back on your work.

So beginning to think of about the two way doors that the one way doors that exist within your business is really important as well. And just ec two stands for elastic compute cloud. It's our computing, you know, cloud services. But I do want to just give credit to those who did have two fingers. Like you had mentioned, we have reduced prices over 120 times since we've released. Could be that customers. It isn't so much a contract that you need to think through so heavily. If you've started that cadence and that expectation, then hopefully you've instrumented your business to continue that. And so customers expect it right. Is that a one way door? Is that a two way door? Maybe it becomes a little bit easier to continue that if you've plumbed that into your business model, think about prime as well like prime. We actually raised the prices of prime not so long ago, but we thought of that as actually surely there are one way door decisions around every price. It becomes a contract with your customers. But we were confident that this was more of a two way door decision with prime because we built so much value behind it. It wasn't just around shipping services. It added the benefits of streaming video and use it and delivering groceries to your door as well. So um, again, there could be some two way door, uh, thoughts around how you're thinking through price primarily though. Like, yeah, be very careful as you're thinking through it. At least we are.

Absolutely. And the last one, releasing a new aws service, any thoughts in this area? Two or 12 or one. Ok. Seems kind of split nice this side, see more ones, this side, almost, see more too. Well, this is such a great answer because everybody's right. I love these types of quizzes. Um so releasing a new aws service, we have many that have lasted, you know, since the inception of aws itself and we have many that were spun up operationalized and decommissioned as well. And so this really is a one way and a two way door in that way, you have to have great consideration and of, you know, who have you brought on to use the service. Do you have a plan for them if you intend to retire it maybe before, you know, the dinosaurs or something like that? But it really is, um there are aspects of a decision like that that have characteristics of both one way and two way doors.

And while there isn't actually an empirical science behind each one of these pieces, we do find it really important to recognize where are there, um both bidirectional decisions that could happen and how, how to manage those who to alert, who to bring into the fold so that you can get, you know, as close as possible to what you believe is the right thing to do. Awesome.

So, another great way that we begin to think about how we guide and speed up decision across our businesses is um to create and kind of live by tenants for a specific team. So uh maybe a team that's made up of a few of those two pizza teams would form a working group to craft tenants to think about these kind of guiding principles. Um that would help kind of like a constitution for a team. I guess you could think about it in that way, guiding principles that would help with the decision making, uh orchestrate tiebreaking and um be able to drive decision again without that aspect of, of micro management. Uh we all work on highly distributed teams. Um there isn't always gonna be somebody awake to tell you exactly what to do.

So in the um creation of uh tenants, um we like to think about them that they could drive agreement around questions that are very hard to verify, um keeping the focus on driving value to the customer and then caveat it with unless you know, better ones, we actually use just an acronym for that. Um unless you know, a better way. Um and beginning to think about how we can um ensure that we work really hard, we craft these things to be as exacting as possible. And they may change. In fact, they probably definitely will. Um but you know, but through the effort to create them at the beginning, sometimes that feels slightly untenable, but they should evolve um through use. And that, that's why tenants as a mechanism is so instructive. Because if you're giving an organization a set of principles that help them make decisions really quickly, you need to inspect the tenants themselves over time because as you mentioned, like things will change in your business with your customers needs.

Absolutely. So we have a couple of uh great ways to think about writing tenants for your specific team.

"And then some, some examples. Uh our newly formed organizational structure just went through this exercise recently. Um but I'll talk through a few of my favorites and then we'll go through some, some examples.

So first, it's kind of to be very team specific. So rather than aligning with a larger organization, like one of the things I think about for our leadership principles is it's great every Amazonian globally has them uh understood in the same vernacular and knows how to use them, which is great. That's a nice shortcut to be able to use that across the company, but it's not super specific to your day to day. So being very specific to the team, I think is quite important for being able to then empower your team members both to use them and have the tenants actually um be super valuable um oriented towards the long term.

So they should be persistent and focused on the durable value of the team. There's that word again, durable, uh innovation was the word of 2022 but I'm gonna say durable is 2020 three's word. Um but oriented towards the long term, like unless you're planning to, you know, cease existence really, really quickly, um it really should be with a view towards that horizon.

They should be um able to be applied every day. Um so really in helping to make these decisions, it's um it's, it's really important that they can be part of the lived experience of the employees so that they're not just i, i like to think about not just on a shelf, like not just on the wall with next to the picture of the corporate eagle or something like that. It's really important that they're in the dna of, of the teams.

Um having one main idea per tenant and uh the last couple of ones be memorable. So kind of that a little bit of snappiness to it is, is great. Uh the moniker behind it to make it easy to roll off the tongue is quite important.

Um cancel, but don't prescribe action. This isn't your sop document, this isn't your uh you know, your specific files on how to do everything. But um being able to counsel your teams through that.

And then the last one, i love that this is no more than seven. Um so recently in our exercise in the creation of tenants, um our team landed on one that I really liked, which was um lead with authentic Amazonian mechanisms and kind of complement with in industry best practices.

So by putting that in a tenant and assembling a global team from all around the world who has experience in design thinking and lean and agile and da da da Amazon has very specific ways of doing things. Our value proposition to our external customers was we do the Amazon way. We know that there's many great ways to innovate products and services. But our job is to share the Amazonian approach.

And so with that, that tenant automatically says thank you for your experience, but we're not the lean start up company and that's fine. You're able to complement with the experience that you bring to the table. But we start with Amazonian mechanisms and I love that because it frankly allays some of the concerns without needing to be super dogmatic.

What, what about you? What are some tenants from Kindle days? Yeah. Um one of them really leaned into the fact of like how we were making decisions and how we were sustaining business growth and how we were making decisions that impacted different customers.

I'm sure in your business, you have many different types of customers and the decisions that you make in your business may impact different customer segments. A little bit differently is that the case? Um does that sound familiar us and Kindle when we were making some business decisions that impacted certainly readers and sometimes publishers who we relied on for ip also, it impacted authors in terms of payouts or other ways that we thought about making the business sustainable for them and other content creators.

And we kept having this issue where we had to get back. Like what did we decide on last time? Why did we decide to move in this that impacted one customer base a little bit differently? So we created a tenant and that was incredibly helpful for us and that tenant was where we came to a decision that impacted customers differently.

We would index towards reader first and foremost, followed by the author, followed by the publisher, maybe this is common sense because in the value stream between the reader and anyone else, like no one exists if there reader isn't there. But that was incredibly helpful for anybody on the team. Even people who just joined the team to be able to make a quick decision, they could weigh how this impacted different customers.

But having that tenant was a very clear line in the sand for the team of where they can absolutely move fast if there's something that was going to benefit the reader and maybe impact others a little bit differently. We had a tenant, strong tenant that led that it also helped internally because certainly we had a lot of conversation with our vendor management team that handled publishers.

Like, why are you making these decisions? Like, don't you know what you're doing to our business? Well, we had a tenant that was in all of our docs. They were very aware of it. They could see it. It also helps evangelize what is most important for any given teams, you know, forward looking business to other teams and helps bring alignment much, much quicker.

Yep. Absolutely. So, bringing alignment quicker is a good segue to the next slide. Um which is another slice of how we get things done fast, which is fast and friendly escalation.

So when you think about it, um really like there, we're in a fully distributed environment. It's really important to be able to move quickly as possible and to be able to bring issues to the forefront um in an expedited way. So that frankly nothing that you can control is slowing down your teams.

There's be plenty of things that are outside of the wheelhouse of things that you can actually affect, but we're really focused on and it's actually, you know, it doesn't mean that it's easy but being able to use the actual term escalation, make sure that it's written down in an email, but also to move forward with a number of different other mechanisms that will then help support the escalation of when something goes wrong.

Whether that is a specific correction of error or blameless postmortem. But to be able to understand that it's actually a great thing to be able to say this was incorrect. This indeed hurt my business. Maybe it might be an escalation coming from the publishing team. But uh it's really an important thing to uh live and breathe it, uh model the way. But also then to talk about it openly that, you know, in the onboarding or um the creation of teams that uh escalation isn't a bad thing. It isn't a dirty word. You're not in trouble. This is really about how do we fix this, learn from it and move forward.

So when we think about that escalation as a mechanism, it's important to consider uh how do you have competing priorities? How do we tell somebody? Ok, you're gonna back down and here's why um when there's too many players or too many cooks in the kitchen, um that's often one are, you know, are you doing my job? Are you stepping on my toes? So that often happens and actually competing priorities aren't necessarily a bad thing, but it is, you know, it is cause for some friction.

Um anytime there's scope strain, so really um feature creep is the term i like to use. Um but indeed, like when those pieces are coming down and you are the responsible party, it's an important thing to be able to raise the hand and say this is gonna hurt the work product. This is gonna affect customers. I need to put a stop to it.

And then the last one is unclear ownership. And when you think of the, uh, two pizza teams and single threaded owners that kind of run and operate them, it's really important that we say, uh, you are the owner of this thing. The buck does stop there and it's important to be able to, uh, drive decision by saying, you know, no, i am. This is the end all meal.

So to kind of wrap it up and we also have a few minutes. We're happy to take any live questions and repeat them back or take a minute to chat outside. But in sustaining this high velocity decision making culture and ensuring that you can make the best and fastest decisions uh for yourself or for your teams.

Um we would recommend creating flywheels around your customers long term needs. Think about that term durable. What are the pieces that are going to remain constant within your organization for the next five or 10 years? I know many of us are tasked with projects that have, you know, 32nd deadlines, not five years or 10 years on the horizon, but really taking a longer term view.

And the second thing we'd encourage you is build a voice of customer. Again, i know everybody here is laser focused on their customer. It's the one differentiating piece that you can bring to the table as an organization, but how you build that, how you are diving into both the data and the anecdote and really empowering teams to get as close to the customer as possible. So you can respond quickly. You can experiment constantly around the needs of your customer and sustain the business growth that you have by not just iterating your core business, but finding new areas to grow and sustain and build new lines of growth. Absolutely instrumental.

And then the last one is kinda develop and constantly inspect those customer centric mental models and mechanisms. And the three that we share today are one way and two way doors. That's something like it could literally just be a sticker and you look up. Oh, is that a one way or two way door? It's pretty, you know, quick question to ask yourself and then be able to either, you know, own or delegate uh the creation and instantiation of tenants and making sure that they are yummy and useful to your teams.

And then the last part is, you know, don't be shy, think about how you can bring in a culture that would celebrate learning through mistakes and moving through escalations in that way.

And before we take maybe one or two questions, i just want to remind you if this was helpful in any way, we have a whole slate of innovation and culturally focused areas of how to not just build the processes and the tools but the organizing the people behind that, that's the holy trinity, right? People process and tools. So please attend any of our great innovation sessions throughout um from monday through. Yeah, the end of re invent on thursday. Absolutely.

And no presentation post covid would be complete without the scanning of a qr code. Um so if you'd love to uh check out any of the uh work product from mine and dan's groups um or get in touch, uh these links will be able to get you there.

Thank you all so much for your time. Um please complete the session survey. Um and if you have other questions, we're more than happy to meet you outside to answer everything. Thanks everyone. Thanks so much."