Given ten years and many iterations, that approach has allowed AWS to rapidly expand into the world’s most comprehensive, widely adopted cloud service. As with our retail business, AWS is made up of many small teams with single-threaded owners, enabling rapid innovation. The team rolls out new functionality almost daily across seventy services, and that new functionality just “shows up” for customers—there’s no upgrading.

  Many companies describe themselves as customer-focused, but few walk the walk. Most big technology companies are competitor focused. They see what others are doing, and then work to fast follow. In contrast, 90 to 95 percent of what we build in AWS is driven by what customers tell us they want. A good example is our new database engine, Amazon Aurora. Customers have been frustrated by the proprietary nature, high cost, and licensing terms of traditional, commercial-grade database providers. And while many companies have started moving toward more open engines like MySQL and Postgres, they often struggle to get the performance they need. Customers asked us if we could eliminate that inconvenient trade-off, and that’s why we built Aurora. It has commercial-grade durability and availability, is fully compatible with MySQL, has up to five times better performance than the typical MySQL implementation, but is one-tenth the price of the traditional, commercial-grade database engines. This has struck a resonant chord with customers, and Aurora is the fastest-growing service in the history of AWS. Nearly this same story could be told about Redshift, our managed data warehouse service, which is the second fastest growing service in AWS history—both small and large companies are moving their data warehouses to Redshift.

  Our approach to pricing is also driven by our customer-centric culture—we’ve dropped prices fifty-one times, in many cases before there was any competitive pressure to do so. In addition to price reductions, we’ve also continued to launch new lower cost services like Aurora, Redshift, QuickSight (our new Business Intelligence service), EC2 Container Service (our new compute container service), and Lambda (our pioneering server-less computing capability), while extending our services to offer a range of highly cost-effective options for running just about every type of application or IT use case imaginable. We even roll out and continuously improve services like Trusted Advisor, which alerts customers when they can save money—resulting in hundreds of millions of dollars in savings for our customers. I’m pretty sure we’re the only IT vendor telling customers how to stop spending money with us.

  Whether you are a start-up founded yesterday or a business that has been around for 140 years, the cloud is providing all of us with unbelievable opportunities to reinvent our businesses, add new customer experiences, redeploy capital to fuel growth, increase security, and do all of this so much faster than before. MLB Advanced Media is an example of an AWS customer that is constantly reinventing the customer experience. MLB’s Statcast tracking technology is a new feature for baseball fans that measures the position of each player, the baserunners, and the ball as they move during every play on the field, giving viewers on any screen access to empirical data that answers age-old questions like “What could have happened if …” while also bringing new questions to life. Turning baseball into rocket science, Statcast uses a missile radar system to measure every pitched ball’s movements more than two thousand times per second, streams and collects data in real-time through Amazon Kinesis (our service for processing real-time streaming data), stores the data on Amazon S3, and then performs analytics in Amazon EC2. The suite of services will generate nearly 7 TB of raw statistical data per game and up to 17 PB per season, shedding quantitative light on age-old, but never verified, baseball pearls of wisdom like “never slide into first.”

  About seven years ago, Netflix announced that they were going to move all their applications to the cloud. Netflix chose AWS because it provided them with the greatest scale and the broadest set of services and features. Netflix recently completed their cloud migration, and stories like theirs are becoming increasingly common as companies like Infor, Intuit, and Time Inc. have made plans to move all of their applications to AWS.

  AWS is already good enough today to attract more than one million customers, and the service is only going to get better from here. As the team continues their rapid pace of innovation, we’ll offer more and more capabilities to let builders build unfettered, it will get easier and easier to collect, store and analyze data, we’ll continue to add more geographic locations, and we’ll continue to see growth in mobile and “connected” device applications. Over time, it’s likely that most companies will choose not to run their own data centers, opting for the cloud instead.

  Invention Machine

  We want to be a large company that’s also an invention machine. We want to combine the extraordinary customer-serving capabilities that are enabled by size with the speed of movement, nimbleness, and risk-acceptance mentality normally associated with entrepreneurial start-ups.

  Can we do it? I’m optimistic. We have a good start on it, and I think our culture puts us in a position to achieve the goal. But I don’t think it’ll be easy. There are some subtle traps that even high-performing large organizations can fall into as a matter of course, and we’ll have to learn as an institution how to guard against them. One common pitfall for large organizations—one that hurts speed and inventiveness—is “one-size-fits-all” decision making.

  Some decisions are consequential and irreversible or nearly irreversible—one-way doors—and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation. If you walk through and don’t like what you see on the other side, you can’t get back to where you were before. We can call these Type 1 decisions. But most decisions aren’t like that—they are changeable, reversible—they’re two-way doors. If you’ve made a suboptimal Type 2 decision, you don’t have to live with the consequences for that long. You can reopen the door and go back through. Type 2 decisions can and should be made quickly by high judgment individuals or small groups.

  As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision-making process on most decisions, including many Type 2 decisions. The end result of this is slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.* We’ll have to figure out how to fight that tendency.

  And one-size-fits-all thinking will turn out to be only one of the pitfalls. We’ll work hard to avoid it—and any other large organization maladies we can identify.

  Sustainability and Social Invention

  Our growth has happened fast. Twenty years ago, I was driving boxes to the post office in my Chevy Blazer and dreaming of a forklift. In absolute numbers (as opposed to percentages), the past few years have been especially significant. We’ve grown from 30,000 employees in 2010 to more than 230,000 now. We’re a bit like parents who look around one day and realize their kids are grown—you blink, and it happens.

  One thing that’s exciting about our current scale is that we can put our inventive culture to work on moving the needle on sustainability and social issues.

  Two years ago we set a long-term goal to use 100 percent renewable energy across our global AWS infrastructure. We’ve since announced four significant wind and solar farms that will deliver 1.6 million megawatt hours per year of additional renewable energy into the electric grids that supply AWS data centers. Amazon Wind Farm Fowler Ridge has already come online. We reached 25 percent sustainable energy use across AWS last year, are on track to reach 40 percent this year, and are working on goals that will cover all of Amazon’s facilities around the world, including our fulfillment centers.

  We’ll keep expanding our efforts in areas like packaging, where our culture of invention led to a big winner—the Frustration-Free Packaging program. Seven years ago we introduced the initiative with nineteen products. Today, there are more than four hundred thousand globally. In 2015, the program eliminated tens of millions of pounds of excess packaging material. Frustration-Free Packaging is a customer delighter because the packages are easier to open. It’s good for the planet because it creates less waste. And it’s good for shareholders because, with tighter packaging, we ship less “air” and save on transportation costs.

  We also continue to pioneer new programs for employees—like Career Choice, Leave Share, and Ramp Back. Career Choice prepays 95 percent of tuition for courses that teach in-demand skills, regardless of whether those skills are relevant to a career at Amazon. We’ll pay for nursing certifications, airplane mechanic courses, and many others. We’re building classrooms with glass walls right in our fulfillment centers as a way to encourage employees to participate in the program and to make it easy. We see the impact through stories like Sharie Warmack—a single mother of eight who worked in one of our Phoenix fulfillment centers. Career Choice paid for Sharie to get licensed to drive an eighteen-wheeler. Sharie worked hard, passed her tests, and she’s now a long-haul driver for Schneider Trucking—and loving it. This coming year, we’re launching a program to teach other interested companies the benefits of Career Choice and how to implement it.

  Leave Share and Ramp Back are programs that give new parents flexibility with their growing families. Leave Share lets employees share their Amazon paid leave with their spouse or domestic partner if their spouse’s employer doesn’t offer paid leave. Ramp Back gives birth mothers additional control over the pace at which they return to work. Just as with our health care plan, these benefits are egalitarian—they’re the same for our fulfillment center and customer service employees as they are for our most senior executives.

  Renewable energy, Frustration-Free Packaging, Career Choice, Leave Share, and Ramp Back are examples of a culture that embraces invention and long-term thinking. It’s very energizing to think that our scale provides opportunities to create impact in these areas.

  I can tell you it’s a great joy for me to get to work every day with a team of such smart, imaginative, and passionate people.

  It’s still Day 1.

  *The opposite situation is less interesting and there is undoubtedly some survivorship bias. Any companies that habitually use the light-weight Type 2 decision-making process to make Type 1 decisions go extinct before they get large.

  Fending Off Day 2

  2016

  “JEFF, WHAT DOES Day 2 look like?”

  That’s a question I just got at our most recent all-hands meeting. I’ve been reminding people that it’s Day 1 for a couple of decades. I work in an Amazon building named Day 1, and when I moved buildings, I took the name with me. I spend time thinking about this topic.

  “Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death. And that is why it is always Day 1.”

  To be sure, this kind of decline would happen in extreme slow motion. An established company might harvest Day 2 for decades, but the final result would still come.

  I’m interested in the question “How do you fend off Day 2?” What are the techniques and tactics? How do you keep the vitality of Day 1, even inside a large organization?

  Such a question can’t have a simple answer. There will be many elements, multiple paths, and many traps. I don’t know the whole answer, but I may know bits of it. Here’s a starter pack of essentials for Day 1 defense: customer obsession, a skeptical view of proxies, the eager adoption of external trends, and high-velocity decision making.

  True Customer Obsession

  There are many ways to center a business. You can be competitor focused, you can be product focused, you can be technology focused, you can be business model focused, and there are more. But in my view, obsessive customer focus is by far the most protective of Day 1 vitality.

  Why? There are many advantages to a customer-centric approach, but here’s the big one: customers are always beautifully, wonderfully dissatisfied, even when they report being happy and business is great. Even when they don’t yet know it, customers want something better, and your desire to delight customers will drive you to invent on their behalf. No customer ever asked Amazon to create the Prime membership program, but it sure turns out they wanted it, and I could give you many such examples.

  Staying in Day 1 requires you to experiment patiently, accept failures, plant seeds, protect saplings, and double down when you see customer delight. A customer-obsessed culture best creates the conditions where all of that can happen.

  Resist Proxies

  As companies get larger and more complex, there’s a tendency to manage to proxies. This comes in many shapes and sizes, and it’s dangerous, subtle, and very Day 2.

  A common example is process as proxy. Good process serves you so you can serve customers. But if you’re not watchful, the process can become the thing. This can happen very easily in large organizations. The process becomes the proxy for the result you want. You stop looking at outcomes and just make sure you’re doing the process right. Gulp. It’s not that rare to hear a junior leader defend a bad outcome with something like, “Well, we followed the process.” A more experienced leader will use it as an opportunity to investigate and improve the process. The process is not the thing. It’s always worth asking, do we own the process or does the process own us? In a Day 2 company, you might find it’s the second.

  Another example: market research and customer surveys can become proxies for customers—something that’s especially dangerous when you’re inventing and designing products. “Fifty-five percent of beta testers report being satisfied with this feature. That is up from 47 percent in the first survey.” That’s hard to interpret and could unintentionally mislead.

  Good inventors and designers deeply understand their customer. They spend tremendous energy developing that intuition. They study and understand many anecdotes rather than only the averages you’ll find on surveys. They live with the design.

  I’m not against beta testing or surveys. But you, the product or service owner, must understand the customer, have a vision, and love the offering. Then, beta testing and research can help you find your blind spots. A remarkable customer experience starts with heart, intuition, curiosity, play, guts, taste. You won’t find any of it in a survey.

  Embrace External Trends

  The outside world can push you into Day 2 if you won’t or can’t embrace powerful trends quickly. If you fight them, you’re probably fighting the future. Embrace them and you have a tailwind.

  These big trends are not that hard to spot (they get talked and written about a lot), but they can be strangely hard for large organizations to embrace. We’re in the middle of an obvious one right now: machine learning and artificial intelligence.

  Over the past decades, computers have broadly automated tasks that programmers could describe with clear rules and algorithms. Modern machine learning techniques now allow us to do the same for tasks where describing the precise rules is much harder.

  At Amazon, we’ve been engaged in the practical application of machine learning for many years now. Some of this work is highly visible: our autonomous Prime Air delivery drones; the Amazon Go convenience store that uses machine vision to eliminate checkout lines; and Alexa,* our cloud-based AI assistant. (We still struggle to keep Echo in stock, despite our best efforts. A high-quality problem, but a problem. We’re working on it.)

  But much of what we do with machine learning happens beneath the surface. Machine learning drives our algorithms for demand forecasting, product search ranking, product and deals recommendations, merchandising placements, fraud detection, translations, and much more. Though less visible, much of the impact of machine learning will be of this type—quietly but meaningfully improving core operations.

  Inside AWS, we’re excited to lower the costs and barriers to machine learning and AI so organizations of all sizes can take advantage of these advanced techniques.

  Using our prepackaged versions of popular deep learning frameworks running on P2 compute instances (optimized for this workload), customers are already developing powerful systems ranging everywhere from early disease detection to increasing crop yields. And we’ve also made Amazon’s higher-level services available in a convenient form. Amazon Lex (what’s inside Alexa), Amazon Polly, and Amazon Rekognition remove the heavy lifting from natural language understanding, speech generation, and image analysis. They can be accessed with simple API calls—no machine learning expertise required. Watch this space. Much more to come.

  High-Velocity Decision Making

  Day 2 companies make high-quality decisions, but they make high-quality decisions slowly. To keep the energy and dynamism of Day 1, you have to somehow make high-quality, high-velocity decisions. Easy for start-ups and very challenging for large organizations. The senior team at Amazon is determined to keep our decision-making velocity high. Speed matters in business—plus a high-velocity decision-making environment is more fun too. We don’t know all the answers, but here are some thoughts.

  First, never use a one-size-fits-all decision-making process. Many decisions are reversible, two-way doors. Those decisions can use a light-weight process. For those, so what if you’re wrong? I wrote about this in more detail in last year’s letter.

  Second, most decisions should probably be made with somewhere around 70 percent of the information you wish you had. If you wait for 90 percent, in most cases, you’re probably being slow. Plus, either way, you need to be good at quickly recognizing and correcting bad decisions. If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.