How to Do Customer Success at Scale with Amazon Web Services

Well, good morning. They tell me to break a leg, so I did. Well, we are going to have some fun today because we're going to be talking about how you do customer success at scale. I've been with Amazon Web Services now for about 2 years and I've learned some really interesting things. About how you can take innovation and really set a set of mechanisms together to really make some things happen. But as you know, when you innovate, one of the things that happens sometimes is that you're not always 100% successful, right? So, interesting, I have two daughters and we have a couple of traditions. Anybody here have traditions in their family with their kids? Awesome. So one of the things that we do is every Sunday morning I make pancakes for my two daughters. And they bring me interesting pictures and they say, "Mom, can you make me these pancakes?" So they brought me this picture from Pinterest and they said, "Mom, can you make me this pancake?" I looked at it, it looks pretty easy, so I said, "Of course I can make that pancake." So— Lesson learned here. First, I was very customer obsessed. I took input from my daughters, my customers, and I tried to produce what they were looking for. Now, they actually consumed the pancake, so it was a good MVP, but the user interface was just a little bit off, as it is a lot of times when you're starting out. And that's really the way innovation is, and it really starts starts with customer obsession. And if you think about customer obsession, the things that really drives customer obsession, I believe, is the cloud. And I know this conference is all about the cloud and leveraging the cloud in lots of different ways for lots of different value. But for me, one of the strongest value points of the cloud is that it enables that customer obsession. It enables you to get faster time to market. I was just reading an interesting report and it showed that startups 7 years ago would spend 80% of their venture funding on IT, getting the right hardware, the right software. Now it's only 20% because of the cloud allowing startups and entrepreneurs to really focus on what's most important, your subject matter expertise, not rebuilding that infrastructure. So the cloud really brings together that time to market, that agility that you need And also that customer experience. I was just reading the latest Gartner report and it showed that customer experience from all the research is really the last form of true competitive advantage. So the cloud enables that as well. So as we're thinking about customer obsession, I know you're probably sitting there thinking, I've heard this before, right? Customer-focused, it's customer-driven, it's customer-obsessed. So if you look at my background, I, I've worked at several different companies and I've also been a startup CEO. And so I've been through lots of different programs on customer-driven, customer-obsessed. But this, what we have at Amazon, is truly differentiated. And it is one of the main reasons that I decided to leave my startup and to go to work at Amazon was to really learn how they innovate, how we innovate, and how we monetize that innovation while bringing such tremendous value to our customers, because it is very different. And what I'd like to do is show you kind of some of those differences, and I'd like to start with just a little video. So if you guys could run the video for me. Alexa, reheat pasta. Reheating pasta. It's cool, right? Yeah, I didn't know you guys put Alexa in a microwave. Yeah, we're putting her in a lot of stuff now. But trust me, there are a lot of fails. Like what? Like— Alexa, play my podcast. When you heard that, did that surprise you? Bark seems exciting. We need an Alexa dog collar for dogs. Ordering dog food. Ordering dog food. You can bark all you want, I'm not paying for any more dog food. Ordering gravy. Ordering sausages. Hey, you better cancel that order. Alexa, hot tub. A sucked. Alexa, play music. Okay. The incident. Wait, that was you guys? I don't know, was it? [MUSIC] [FOREIGN LANGUAGE] Powering up, powering down. She says she's doing it, but I don't see anything. Do you? Holy— I'm not talking to you. So I wanted to show you that commercial. I used to use another example of a company that was very technology-focused, not customer-obsessed, but that company actually asked me not to share that anymore, so I've taken that down. But what I think is interesting about this is I think some of those examples, they show that to innovate and to be customer-obsessed, you have to to fail. But many of those examples, like the dog collar on the dog, where you're having artificial intelligence with a, with a dog, which is probably two of the most adorable things on the planet, doesn't mean that you're gonna have a customer-obsessed solution. It means you might have a technology-focused solution. And I know that a lot of companies really love to create a product from technology and then apply that to a problem versus starting with a problem. Now let me show you an example that I really love. It's not a tech example, but I really think it shows great product management capability and great customer obsession. So I don't eat jerky, but everybody in my family does, and they tell me this is the best. I don't know if anybody here eats jerky, but they say this is the best jerky that's out there. So the product management team got together and they were talking to customers, and one of the things that they found with this jerky was that would get stuck in people's teeth. So being customer obsessed, they started developing new formulas for the jerky, looking at things, customers didn't like it as well. And so they decided to be creative and look beyond, but still focus on the customer problem. So if you can see it up there in that circle, I tried to circle it in purple, that's dental floss. So they've actually included dental floss now in the packaging. This went viral, customers were going crazy for it. Not just because it helped to solve a problem, but it showed how customer-obsessed this company was. And so when you're thinking about customer obsession, having that customer success, what you're really looking for is you're looking to, um, have empathy for your customer. And there's a number of ways that you can do that. Uh, one is that you can hire people who look like your customer. So Amazon Web Services, we are builders, we innovate, I have engineering teams, we create things. And so a lot of our customers are very similar to us and therefore our empathy comes from us being just like our customer. But I also manage a team that puts.NET applications, Windows applications on AWS. Most of our developers are not.NET developers. And so there we have to figure out different ways different ways to obsess over our customers and really learn about what they want and what they need. So you've got to figure out how you're gonna have that empathy for your customer. And I know one of the things that's really hot here in Silicon Valley, my startup actually used it, is innovation labs that focus on empathy, or an empathy lab. One I went to was for healthcare, and we used VR, virtual reality, to simulate what it was like for a doctor or researcher to have a broken leg, to have glaucoma, to really be able to feel and empathize with their customer. So the hardest thing that I found from my work over the past decade in, in the industry was how do you have mechanisms to really capture that customer obsession, to innovate, to get to that customer success? And so what I wanted to do is to share some of those that we leverage at Amazon. The first one that we use is something called working backwards from the customer. So everything we do starts with the customer, not with the technology. In fact, the first service that I built with my team, I talked to personally 141 customers. Now that was just me, that wasn't my whole team as well. So really understanding what they want, really understanding what the pain points were, really understanding what the desired outcome was so that I could then innovate and come up with a great solution. A lot of times, especially here in the Valley, we start with the technology. Artificial intelligence, so cool, what can I do with artificial intelligence? Versus what does the customer want or need and how do I begin that journey from the customer working backwards. And the way that we start that mechanism from thinking about the customer working backwards is we start by writing a press release. Now this may seem interesting and a little bizarre. I know it was for me when I came in because we usually started building code and then we would present that to sales and marketing and they would come up with a press release. That we were trying to solve. But the way we do it at Amazon is we write the press release before we write a single line of code, before we have a single design document. We write the press release around that. Now why is that so important? Well, if you think about a great press release, it's usually one page. You know, Mark Twain used to say, "I would have written a shorter book if I'd had more time." So it really causes you to think very deeply but also very succinctly about what you're trying to build. You start out with a headline, so you've gotta summarize what you're trying to do, and your subtitle, and then you have to explain what you're launching, what you're going to release in a paragraph at the very beginning. And that paragraph, you know, is only a few sentences, so you can't have a 10-page design doc. You really have to hone in on what you're trying to launch. Our next paragraph hones in on the customer problem or opportunity that we're trying to go after. So again, very targeted, crisp, succinct, not every problem, not world peace, world hunger, but specifically what are we trying to do in this particular example. And then we have to include customer quotes, which shows that we've already talked to the customers and that they would be willing to stand on stage with us or to actually be in a real press release with us around this particular announcement. It's a lot harder than it seems, so I would encourage you to try it. Just try writing a press release, and your press release could be around a new product, maybe a new business model that you're working on. It could be around an operational excellence area. So it doesn't have to be just on a product. It could be on any innovation that you're looking for. The second thing that we do as a mechanism is we write a set of frequently asked questions. Now I thought, this is gonna be so easy 'cause I'll make up my own questions and I can answer those questions. But wouldn't you know it, in Amazonian style, Amazon already has the questions written that you have to answer. So some of those questions are basics. What are you launching? But some of them are, if your customer was having lunch, what would they say about this new service or this new offering or this new business model? Then one of the hardest ones for me is what would they say over dinner complaining about, wow, I wish this had this, or I wish the user interface looked like that. What would they say? Then we actually debate in a room. We get together in rooms and we talk and we say, now given that our customers might not like that about what we're releasing, can we get the next version out quick enough? Or should we hold, should we wait on the release until we can make sure that we meet that particular need? Other interesting things that we do in the frequently asked questions is we actually have a set of wireframes in there for our user design. What does the customer experience look like? And this is really important to us because we wanna make sure we get that customer experience right. We obsess over what is the experience our customer will have. One of the services we just released at our re:Invent conference, our CEO actually went through the wireframes with us to say, is this design simple enough? Is it good enough? Who did you talk to? Who walked through it? Are you sure that this is the right design? And we do all of this before we even write a single line of code. Now we do something at Amazon called Digital Innovation Workshops, and this is where we teach our customers how to do this customer obsessed customer success model. And most recently I had a customer that came in and he was looking for some of the nuances about the questions that we ask. So I created this slide based on some of the questions that he asked me to really explain and some of the nuances behind those. So the first one is, who is the customer? So when I was talking to my customer and I asked them who the customer is, do you know what he said? Telco. Now, telco to me is an industry. It's not a customer. A customer is a person. It's a person that you're going to sell to. So whereas this customer of mine was trying to target telco, we go down and we say, we're targeting the DBA, the database administrator for this database at this type of customer in this industry. What is he or she looking for? So we're very specific about who the target is. It's not generic, it's not hand wavy, it's not done through focus groups, it's through really knowing our customer. The second nuance that we do is we focus in on the most important customer problem or opportunity. So there's thousands of them out there, but we really hone in on what's the that number one problem and we ensure that we not just meet that need, but that we exceed that expectation. And so that means you have to know what that number one problem or opportunity is. And that's also hard because typically customers will talk to you about numerous things. The third one is you have to know the most important benefit that you're trying to get for the customer. And then And a really interesting one is how do you know that your customers need something? It's not just because they tell you or because you heard it. And here's where having diverse teams really makes a difference. Diversity both in terms of men and women, different countries, but also different thought processes. We had a team once that was formed of all graduates of a certain MBA school. And you know what, it was interesting because they were coming at the problem from one perspective based on their training. So without that diversity of thought, they really weren't able to relate to what their customers really needed. So you need to look at that diversity and ask the question, how do you know that this is what the customer needs? Who did you talk to? Who interpreted the data? How did, how did that happen? And then the fifth one, which is I think really important, is what's the customer experience gonna look like? And we truly obsess obsess over this. We spend times in rooms debating this. We write 6-page narratives about what this is, what this looks like, what is the experience gonna be. We take a lot of time and we consider that we are inventing on behalf of our customer, and that's a serious job, so we take that very seriously. I told you we teach this concept to our customers, and one customer that I thought was a great example is Just Eat. I don't know, any of you know them? They're a £5 billion startup out of the UK. And they still have very much startup-like tendencies. What do they do? They actually deliver food, so they do takeout food. And you would think, wow, how can you really innovate around that? But their CTO, Dave, is amazing. So he took our working backwards kind of mechanisms, just like we had a presentation today, And he started really trying to understand how is he gonna get to his customers. So he did Shark Tanks with them in the UK to try to find out what were his customers thinking about, what did they want. Did hackathons, did a blog to try to get information from customers on what they loved and what they didn't love. And they came up with this really interesting idea where you can now say to Alexa, Alexa, have Just Eat order my pizza. Not only will she order the pizza, but she'll know what pizza you like because she's been trained to know your preferences. Or you might say, "Alexa, order sushi," and sushi will take 60 minutes, but they know that Sandy won't wait more than 45 minutes, so they might suggest something else that could be delivered in that amount of time. This all came up from Just Eat working backwards from the customer Not saying, "How could we use Alexa?" or "How could we use artificial intelligence?" but "What is our customer's problem? How can we solve that?" and then going forward to the solution. They actually wrote a, what we call a PR FAQ, that press release and frequently asked questions, to really design and determine how they would best solve this particular problem or solution. Now in addition to writing that press release and that frequently asked question, the other thing that we do at Amazon that I think is very unique from any other company that I've worked at is we do two pizza teams. Now what is a two pizza team? Well, we say that any problem should be able to be broken apart into smaller pieces. And in fact, we believe the largest size team that anyone should have is one that could eat 2 pizzas in a sitting. Now I try to calculate who could eat less food so I could have a bigger team. But the concept is that because you're gonna break apart the problem into something that's manageable and something that can fit with a small team, that team is empowered to make decisions. They don't have to go through any bureaucracy. They can move very quickly. They can innovate because they are the startup that's gonna do this particular service. You know, we see t-shirts around, you know, 4 or 6 people teams at Amazon because they're, they're the ones that are developing it just like a startup does. So as you start to grow, just like Just Eat, I told you they consider themselves a 5 billion pound startup. We at Amazon Web Services consider ourselves a startup. How do you continue to have that startup innovation and that mentality? I'll give you some interesting numbers. So when you are a startup, typically you'll spend about 80% of your time with your customer. 'Cause you have to, right? You usually have 2 or 3 customers, you're doing MVPs, you're really learning from them. As you get to be a small to medium business, that number goes to 50/50. You're spending about 50% 50% of your time with customers and the rest of the time you're spending on internal things, meetings, development, workshops. But when you get to be a big, big, big, big company, you end up spending 20% of your time with customers and 80% of your time is spent elsewhere, which is exactly opposite of customer obsession. So part of what we try to do is to ensure that we're always in that startup mode, always resource constrained so that we can make those right decisions on behalf of our customer, prioritize that correctly. Now the other big one for us, and this was probably my hardest to learn, is something we call high velocity decision making or two-way doors and one-way doors. So I'd worked at very large companies before, And we spent a lot of time making decisions. We'd review things up the chain, down the chain, sideways in the chain, get data from 50 different places. Then I had also done my own startup, worked in startups, and there you're making decisions, you don't have enough data, you're moving very, very quickly, so you're making decisions on the best stuff that you have and you're having to do it very fast. Well, the interesting thing about Amazon is we have both kind of decisions happening at the same time. And the way we differentiate around are you going for speed or you're really going for very thoughtful cognitive risk-taking is one-way doors or two-way doors. So a two-way door is if you can take that decision and if you made the wrong decision, which you'll, you'll make wrong decisions, that you can easily reverse it. Changing the color of a website, or, you know, changing the name of a product, or even certain products that you release. You could release them, they may not be great, and then you can retract them from the market. But a two— a one-way door, one-way door is a very thoughtful discussion because once you go through that door, having to go back out is really tough for your customer, in particular your customer. And so what we really do there is we're very thoughtful, we spend a lot of time on the decision, we get lots of different people from different groups together to help us make that decision so we're doing that in the right frame of reference. Now what's so hard about that is in my day I'm two-way door, two-way door, two-way door, two-way door, and then one-way door. So I've got to slow down, I've got to think, I've got to collaborate. In that frequently asked questions document that we talked about, one of the big questions is, is this a one-way door or is this a two-way door? And if it's a one-way door, we really slow down the process so we can really think about it. If it's a two-way door, we work rapidly and innovate very, very quickly. Does that make sense? 'Cause this was one of the toughest ones, but I think one of the most valuable ones that I found working at Amazon. So let me give you an example of this. This is Redfin. I don't know how many of you guys know Redfin. They're based in Seattle. You guys know them? Okay. So they've taken the real estate market and they put it online. They consider themselves a SaaS vendor for real estate, for selling and buying real estate. Now as they were working backwards from their customer, they realized that their sellers, their buyers, and the real estate agents themselves all needed to know an estimated value of the property or the house that they were buying. And so they came up with this first. And so initially their CTO, Bridget Frey, if you don't know her, you should get to know her, she's incredible. She said, we need to go fast. We need to develop this feature fast because the first to market will be really important. And then they paused to say, is this a one-way door or a two-way door? What do you guys think? Is this one-way or two-way door? Nobody knows? One-way. One-way door. This is a one-way door because think about it, if you come out there and you say your property is worth $400,000 and then tomorrow you say, oh, your property is worth $1 million and then 2 days later your property is worth $100,000, You've lost all trust and confidence from all 3 of your biggest customers, your buyers, your sellers, and your real estate agents. So they actually slowed down and went out, they hired some really interesting real estate agents. They started doing direct feedback with technologists around the buyers and the sellers. They tried to figure out how banks did assessments They combined all that information together and they said, you know what, in order to do this correctly, we need to use artificial intelligence. So then they went out, they had some skills and staff, they went out and hired those skills. Took them a long time to build this particular function. And if you use this today for your house, this is considered the most accurate assessment of your house's value on the market. And in fact, in some places, banks will actually use this as an assessment of value versus doing their own assessment. Again, you gotta think through, is this a one-way door or a two-way door? They are rapid to develop on any two-way doors, but if a one-way door happens, they slow down, make the right decision, and here they actually made the right decision. They were trading off being first to market versus being the most accurate, and accuracy won the game here. So when you put all of that together, all of this together, what you really get is what we call the Amazon flywheel. There have been books written about this, lots of articles. I just saw a new article that was in Forbes around this flywheel. And the whole concept here on the flywheel is that if you select for your industry the right areas, the right segments, right focus areas, and you feed those elements of the flywheel, all of them, not just one of them, then what you'll do is you'll create this, this continuous cycle of growth for your particular company. So for instance, for Amazon, what Jeff Bezos and the team said is, you know what, if we have lower prices, then that will drive more traffic to our site, and therefore more sellers will come, which will enable us to deliver lower prices, which will get more traffic, which will attract more sellers, which will allow us to lower prices. So if you think about that, that continuous flywheel continues to drive that growth. But it doesn't just work for Amazon, it can work for other areas too. So let's take, um, a musician. A musician as an entrepreneur, they could produce an album, and by producing an album, they could get tours, right, to go and sing. Those tours make them more popular, which would drive endorsements. Endorsements will drive more albums, which will continue the cycle. And as long as they're feeding all parts of that flywheel, then they'll continue to grow. And as we walk through with customers, we have many examples of this, and we work on this with customers of how to create that flywheel of value for their industry, for their technology, and for what they're focused on. And typically it centers around some of the same things, but it could center around the value, the selection, the convenience that you drive in the marketplace. One of the customers that we have that I think does this the best is one of the largest SaaS vendors, maybe the largest, which is Infor. Infor is really, I think, very interesting because they were working backwards from their customers and figuring out what is their flywheel, look like. And part of that hinged on producing not just horizontal solutions, but also vertical solutions to feed their flywheel. And so as they've gone out and grown, they still consider themselves a startup, but as they feed that flywheel, and as that flywheel is moving, it continues to feed their growth. And as they add new industry verticals, as they're adding more selection of their product and their portfolio, as they're supporting different options, then they'll continue to grow and feed that flywheel. And you can do this for really any business that you have in the marketplace. So the way I would like to, uh, to end this is we have a lot of companies who study this for long periods of time. And they'll look at the flywheel and they'll look at some of the mechanisms like the press releases and the PR FAQs and they'll say, Well, how do I get started? What I'm gonna say to you is don't overanalyze this. Figure out a couple of these experiments that you wanna run, right? 'Cause this is all about experiments, going back to my pancakes. You know what, it might not be the prettiest thing, but if people eat it and consume it as an MVP, you've got something and you can continue to innovate on that and create your own flywheel. A couple of numbers, CEOs today, 93% of them say say that customer obsession and innovation is the only way that they can stay competitive going forward. But 88% of them do not believe that their teams know how to do this in a consistent fashion, know how to understand their customer and know how to innovate. So I think practicing these things makes you more valuable, not just as a company, but as an individual as well. Listen to your customers. Wherever they are. And I'll tell you one last story, which I think is really interesting. When I came to Amazon, we would be in rooms and meetings, and people would say to me, "Sandy, what would Low Flying Hawk think about this idea?" I'm like, "Wow, I don't know. I'll go find out." So I went, I'm like, "Who the hell is Low Flying Hawk?" Well, it turns out that Low Flying Hawk is on one of our technical blogs, blog, and he or she writes comments on that blog and they give their opinion. And it turns out that they had a lot of really, really good insight and really good opinions. And so I asked Low Flying Hawk, what would you think about this particular thing? And he or she gave me the answer. So I started thinking, wow, I wonder if Amazon knows how much money Low Flying Hawk spends. So I went and did this research and I was pretty proud. I was like, wow, I know how much Low Flying Hawk spends. And it was interesting because when I came back, they were like, we don't really care. This person's a customer, they have great ideas, and we're always gonna reinvent on behalf of the customer regardless if they spend $1 a month, $10 a month, or a million a month with us. We're gonna innovate on behalf of that overall customer. So what I would challenge you to today is to get started, look at some of these mechanisms and techniques, I know I won't have time to answer all the questions that you have, but I, I've included my Twitter handle here, and you can reach me via a lot of different areas on social media. But if I can help you at all getting started creating your flywheel or looking at some of these mechanisms like that PR FAQ, I would love to do that. Thank you and happy innovating.