贝索斯致股东信(2014)

qimoe 发布于 2 个月前

致我们的股东:

一个梦幻般的业务有至少4个特点:用户喜爱;能扩大规模;高回报;持久——能连续几十年的那种。如果你找到了这样的机会,别往右滑,直接结婚!(译者注:此处用的典故是国外热门的相亲应用Tinder)

呃,我想说从这个角度来看亚马逊可不是“一夫一妻制”哈。在20年冒着风险的团队努力和各方支持下,我们很高兴能和下面三个“伴侣”“结婚”:Marketplace,Prime和AWS。这三个业务都是当时大胆的下注,很多敏感的人都觉得成不了。但是现在已经很显然我们拥有这些业务有多幸运。同样显然的是,这三个业务可不是随便三个业务,我们的工作就是不断养育并增强他们。

我们会遵循我们一直以来的方法:用户至上而不是竞争对手为中心;打心底的热爱创新;追求卓越运营;以及主动思考长期。好好干加上一点运气,Marketplace、Prime和AWS将持续在未来很多年服务用户并为我们创造收益。

Marketplace

Marketplace在早年并不简单。最早我们推出了Amazon Auctions。一共有七个人使用了这项服务,算上我爸妈和兄弟姐妹们。然后Auctions变成了zShops,基本上就是一个固定价格版本的Auctions。一样,无人问津。然后我们把zShops变成了Marketplace,在内部,我们把Marketplace称为SDP(Single Detail Page,单一详情页),用意是和第三方商家共享我们唯一的“不动产”——商品详情页,引入和自家零售的竞争。这样对我们的用户更方便。第一年的时候有5%的商品是通过这个渠道销往第三方商家,如今是40%,在我们平台上售卖的商家全球超过200万。在2014年,用户们在这些商家购买了超过20亿件商品。

这个混合模式推动了亚马逊的飞轮。最初消费者是被我们平台上大量价格公道合理的商品和用户体验所吸引。当时我们引入第三方商家和我们竞争,我们对消费者能产生更强的吸引力,继而吸引更多第三方商家的进入。这也进一步在我们扩张的同时节约了成本,使得我们可以持续低价并对符合标准的订单免运费。在美国推出这个项目之后,我们迅速在别的地方也推广开了这个模式。结果就是在全球的亚马逊网站上都有了Marketplace。

我们尽力为卖家排忧解难,帮助他们取得商业上的成功。通过我们的Selling Coach项目,我们自动生成一系列机器学习的“助推”建议(一周大约7千万条),试图预警商家可能的库存不足或者提醒他们价格可能过高。这些助推让我们的卖家获得了几十亿的增收。

为了进一步全球化Marketplace,我们在全球各地都帮助卖家们,即使在那些我们没有进入的市场我们也会帮助商家把商品卖到亚马逊上。我们去年接纳了超过100个国家的商家,并且帮助他们和185个国家的用户联系。

大约1/5的第三方商家的商品卖到了国外,去年这样的跨境销售翻了倍。在欧洲,卖家可以只开一个账户来管理不同语言的业务,并且让他们的商品在五个欧洲国家的网站里同步销售。最近,我们开始加强跨境物流的能力,让从亚洲来的到欧洲和北美的商品享受更便宜的运费。

Marketplace在我们快速增长的印度市场占据了核心位置,因为我们在印度的所有商品都是由第三方商家提供的。Amazon.in现在提供比任何当地其他电商网站更多的商品,超过2000万商品和21000商家。使用Easy Ship服务,我们提供从选取商品到送货上门一条龙服务。基于这项服务,印度的团队开发了Kirana Now服务,做到了街区零售铺的商品两小时之内送达,在方便用户的同时还大幅提升了这些零售铺的销量。

可能对第三方卖家最重要的是FBA,这一条我们放到Prime那一节来说。

Amazon Prime

十年前,我们开启了Amazon Prime,最早是针对“所有你能吃的(all-you-can-eat)”设计的快速免费运货服务。有人反复告诉我们这一举措风险很高,在某种程度上确实是这样。在第一年,我们损失了几百万美金的运费收入,没有什么数学计算能证明我们的业务能值回票价。我们继续干下去的决定是基于在我们推出Free Super Saver Shopping服务时得到的正面结果,也是基于一种直觉:用户可能很快就能觉察到这会给他们带来史无前例的购物体验。除此之外,分析指出,如果我们能因此而扩大规模,我们的运送成本最终能大幅降低。

我们的自营零售商品是Prime的基础。在组建零售团队来创建一个个品类的同时,我们也建造了大型系统来自动化库存补给、库存摆放和物品定价的过程。精准的送达承诺让我们必须使用全新的方式来做物流管理,而这一整套系统的革新是一个伟大的成就。自2005年我们开始Prime服务,我们的物流中心从当时的13个增至如今的109处。我们正处于第八代物流中心,用自有软件系统来管理收货、存储、挑拣和出货全过程。在2012年收购Kiva,变成Amazon Robotics之后,我们目前雇佣了15000个机器人来协助装货和取货,这史无前例的提高了效率并降低了成本。我们自有的零售库存管理系统始终是推动Prime发展的最好引擎,也是我们吸引第三方商家和流量的关键部分。

尽管快速送货是Prime的最大优势,但我们也不断为Prime注入新的能量。其中最重要的莫过于数字服务和电子设备。

在2011年我们在Prime增加了Prime Instant Video服务,目前已经支持几万电影和剧集,可以在美国无限观看。并且我们也在把这个项目扩展到英国和德国。我们在这些内容上花了大价钱,所以我们必须好好算算它的影响。我们问自己:这值吗?这有推动Prime发展吗?我们看到了用户从这里开始Prime免费试用,随机转化成了付费会员,重复订购,购买更多商品。我们很高兴目前为止这个项目带我们走到了现在这样的阶段。

尽管绝大多数PIV上的支出都是在购买版权,我们也斥重金打造自建内容。我们的团队开门红,剧集Transparent成为了第一个得金球奖的网剧,Tumble Leaf获得了最受学龄前儿童欢迎的动画片奖。除了这些奖项之外,还有其他的奖项也呼之欲出。自建内容的一大好处是,它最先被Prime用户享用,独家领先全市场!加上我们剧集的高质量,这个独家权益吸引了很多剧迷。我们也很喜欢这种一个会员全都能干的服务方式,并且在各个方面都在扩展这种独特商业模式。我坚信我们是第一家通过拿了金球奖而提高了工具箱和纸尿布销量的公司!

Amazon设计制造了硬件设备,从Kindle到Fire TV再到Echo,也为PIV和Prime Music等业务赋能,在很大程度上促进Amazon生态的发展。我们还有更多的计划,硬件组有很多令人惊喜的计划,敬请期待。

Prime在最核心的快速免费运送的道路上也没有停下发展的脚步。最新推出的Amazon Now让Prime会员能够享受几千件商品的2小时送达和加7.99美元的一小时送达。我们收到了很多如下评论:“在过去的四周我和我老公在Amazon now下单了无数次,真是太便宜太快太方便了!”我们在Manhattan、Brooklyn、Miami、Baltimore、Dallas、Atlanta和Austin都开启了这项服务,而且还有更多城市在计划中。

现在我想说说Fulfillment by Amazon(FBA)。FBA非常重要,因为它把Marketplace和Prime粘合在了一起。因为有了FBA,Marketplace和Prime不再是两件事 。事实上现在我已经没办法把这两个分开来谈了。他们的经济效益和用户体验都深度结合在了一起。

FBA是一个针对Marketplace卖家的服务。当卖家决定使用FBA,他们将他们的商品存储在我们的物流中心之中,我们负责所有的运输、 客服和退货。如果用户同时购买了FBA的商品和Amazon自营商品,我们可以在一个包裹里运输——这极大增加了效率。更重要的是,当卖家加入FBA时,它们的商品就进入了Prime范畴。

能够紧紧抓住这个看起来很显然的事实并没有大家觉得的那么简单。但是试试看总是好的。如果你问卖家想要什么,正确(或者显然的)答案是:他们想卖更多。所以,在他们参与FBA时结果是什么?卖更多。

从Prime用户的角度来看,每有一个商家加入FBA,这就意味着Prime会员可以享受更多精选商品,会员的价值就得到提升。如此很强力的推动了我们的飞轮。FBA完成了我们的闭环:Marketplace为Prime注入能量,Prime也为Marketplace注入能量,两相促进,形成强有力的增强回路。

在一个2014年针对美国商家的调查中,71%的FBA商家表示他们的单位销售在加入了FBA之后提高了20%。在节日期间,全球FBA商品销量比去年增加50%,囊括了40%的第三方销售商品。Prime付费会员数量去年全美增长了50%,全球增长了53%。FBA对商家和消费者是双赢。

Amazon Web Services

始于九年前的一个大胆的想法,AWS已经变得很大,并迎来快速增长。最早使用的是初创公司,比如Pinterest、Dropbox和Airbnb,他们现在还是AWS的用户。

从那之后,大公司也上了贼船,他们的加入和初创公司是一样的理由——快速和灵活。IT成本的降低很吸引人,有的时候节约的绝对金额是惊人的。但是便宜从来没好货,大公司非常依赖IT技术。所以“我们提供一个和你现在一样好的更便宜的服务”不会吸引大多用户。用户真正需要的是“更快更好”,如果恰巧还能更便宜,那就太棒了。但是价格便宜永远只是调料,而不是牛排。

IT特别高杠杆。你不希望你的竞争对手的IT部门比你更灵活。每个公司都有一长串技术业务想尽快实现。痛苦的现实是,在经历了很多权衡取舍之后,大多数项目都没有办法真正开展。即便是那些开展的,也延期交付或者做的功能都不全。如果IT部门能够想出一个办法快速交付高质量的技术产品来为公司赋能,他们就会为自己的企业做出重大贡献。

这也就是AWS快速发展的原因。IT部门发现当他们采用AWS的服务时,能完成更多任务了。他们花更少的时间在低附加值的事情上,比如管理数据中心、网络、操作系统补丁、容量规划、数据库扩容等等。同样重要的是,他们能使用我们强大的API和工具来戏剧性的简化打造可扩容的安全可靠高性能系统的过程。这些API和工具还在不断的被我们的团队优化,不需要用户操心。

如今AWS已经有超过100万活跃用户,其中包括大大小小各种公司和个人,在你能想到的各种行业。2019第四季度AWS用户相较于去年同期增长了90%。包括GE、Major League Baseball、Tata Motors和Quantas等公司都在AWS上开发手机应用,从众包个性化医疗应用到卡车管理App。其他的用户,比如NTT DOCOMO,金融时报和Securities and Exchange Commission使用AWS来分析他们的数据。还有另外一些客户,比如Condé Nast, Kellogg’s和News Corp已经把它们的核心应用甚至整个数据中心都移植到了AWS上。

我们这一路也不断加速创新,从2012年160个新功能到2013年的280,再到去年的516个。有很多都可圈可点:从WorkDocs和WorkMail到AWS Lamda和EC2 Container Service再到AWS Marketplace。出于长度所限,我在这里只着重介绍一条:Amazon Aurora。我们希望Aurora能够为一个非常重要又常常出问题的功能重新树立标准——关系数据库。Aurora是一个和MySQL兼容的数据库引擎,既能够为你提供像商业数据库系统的速度和可用性,又能为你提供开源数据库的简洁和低价。Aurora的性能是一般MySQL的5倍,价格是它的十分之一。关系型数据库困扰了各种组织和开发者很久了,我们对Aurora的面世感到无比兴奋。

我相信AWS是那种可以在未来很多年持续服务用户并获取盈利的梦幻业务。为什么我那么自信?首先,机会很大,围绕在全球在服务区、网络、数据中心、软件基础设施、数据库等的投入周围,遍地黄金。如同我当时对Amazon零售业的判断,我认为AWS的市场潜力无限大。

其次,它现在的市场领先地位(大幅领先)是一个很强的优势。我们努力,非常努力,让AWS变得尽可能简单易用。尽管如此,它还是一个具有丰富功能的相对复杂的工具集,学习曲线比较长。一旦你学会了,而它又那么好用,你大概就不想再在另一个平台上重新学一遍了。当然我们永远不会停下改进的脚步,但是如果我们持续为用户提供亮眼的服务,他们有充分的理由持续使用我们的服务。

除此之外,还是因为我们的领先地位,我们现在有数千AWS“大使”在为我们四处奔走传播。软件工程师在换工作的时候会说:“我在前雇主那用的就是AWS,咱们也应该考虑一下,我觉得会更好的!”能看到程序员们跳槽时把AWS写在简历上,这是个好兆头。

最后,我对于AWS的资本回报很乐观。这是一个团队做出的研究因为AWS确实是一个重资本项目。好消息是研究的结果很乐观。大体上说,AWS相比于另一种方式就没有那么“重资本”了——DIY数据中心,而后者使用率极低,几乎不会超过20%。把设施提供给多个用户共同使用使得AWS的使用率很高,这也就意味着高资本效率。更近一步,再一次我们的领先地位帮了忙:大规模经济又从另一个层面提升了资本效率。我们会持续关注并调整AWS让他的资本回报更高。

AWS还很年轻,还在增长和进化中。我们坚信只要以用户为中心,我们还会继续领先。

Career Choice

在结束之前,我想要跟股东们再谈一件令我们感到兴奋又骄傲的事情。三年前我们推出了一项创新的员工福利Career Choice计划,这个计划旨在为员工自己选择的培训项目报销95%的学费,比如飞行器机械或者护理学,不管是和工作有关还是无关。目的很简单:提供更多选择。

我们知道,对于我们在物流中心的一些员工来说,Amazon可能是一个理想的职业。对于其他人来说,Amazon可能只是一个跳板,跳向其他可能需要专业技能的工作。如果合适的培训能帮到他们,我们愿意帮助。目前为止我们已经帮助超过在8个国家的2000员工加入了这个项目。鉴于很多人参与,我们现在在工作地点开设了培训室,这样很多培训都可以在物流中心内完成,更方便我们的员工。

现在有8个物流中心开设了15门课,目的是培训技术内容。这些教室都是透明玻璃打造,这样可以吸引更多人也鼓励很多人。我们我们相信Career Choice是一个吸引优秀人才加入我们物流中心和客服部门的创新方案。这些职位能成为他们走向新的需要技能的就业市场的中转站,就像第一个参与这个项目的员工,她现在成为了一名社区护士。

我还希望你能加入参观我们的物流中心的队伍,现在已经有超过24000人报名了,这样你就能知道在你惦记了购买按钮之后发生了哪些神奇的事情。除了在美国,我们在英国Rugeley和德国的Graben都开了供参观的工厂,而且我们还在继续扩大。

—华丽分割线—

Merketplace,Prime和Amazon Web Services是三个牛逼的点子。我们很幸运拥有它们,同时也下定决心发展它们,让它们对用户更友好。你要相信我们还会找到下一个好机会。我们已经有好几个备选了,正如我们在20年前所说,我们会持续做出大胆的决定。我们永远不会停止通过创新来服务我们的用户。

一如既往,1997年的第一封信在最后。不忘初心,始终Day 1

Jeffrey P. Bezos

(Vinchent翻译)


英文原文

To our shareowners:

A dreamy business offering has at least four characteristics. Customers love it, it can grow to very large size, it has strong returns on capital, and it’s durable in time – with the potential to endure for decades. When you find one of these, don’t just swipe right, get married.

Well, I’m pleased to report that Amazon hasn’t been monogamous in this regard. After two decades of risk taking and teamwork, and with generous helpings of good fortune all along the way, we are now happily wed to what I believe are three such life partners: Marketplace, Prime, and AWS. Each of these offerings was a bold bet at first, and sensible people worried (often!) that they could not work. But at this point, it’s become pretty clear how special they are and how lucky we are to have them. It’s also clear that there are no sinecures in business. We know it’s our job to always nourish and fortify them.

We’ll approach the job with our usual tools: customer obsession rather than competitor focus, heartfelt passion for invention, commitment to operational excellence, and a willingness to think long-term. With good execution and a bit of continuing good luck, Marketplace, Prime, and AWS can be serving customers and earning financial returns for many years to come.

Marketplace

Marketplace’s early days were not easy. First, we launched Amazon Auctions. I think seven people came, if you count my parents and siblings. Auctions transformed into zShops, which was basically a fixed price version of Auctions. Again, no customers. But then we morphed zShops into Marketplace. Internally, Marketplace was known as SDP for Single Detail Page. The idea was to take our most valuable retail real estate – our product detail pages – and let third-party sellers compete against our own retail category managers. It was more convenient for customers, and within a year, it accounted for 5% of units. Today, more than 40% of our units are sold by more than two million third-party sellers worldwide. Customers ordered more than two billion units from sellers in 2014.

The success of this hybrid model accelerated the Amazon flywheel. Customers were initially drawn by our fast-growing selection of Amazon-sold products at great prices with a great customer experience. By then allowing third parties to offer products side-by-side, we became more attractive to customers, which drew even more sellers. This also added to our economies of scale, which we passed along by lowering prices and eliminating shipping fees for qualifying orders. Having introduced these programs in the U.S., we rolled them out as quickly as we could to our other geographies. The result was a marketplace that became seamlessly integrated with all of our global websites.

We work hard to reduce the workload for sellers and increase the success of their businesses. Through our Selling Coach program, we generate a steady stream of automated machine-learned “nudges” (more than 70 million in a typical week) – alerting sellers about opportunities to avoid going out-of-stock, add selection that’s selling, and sharpen their prices to be more competitive. These nudges translate to billions in increased sales to sellers.

To further globalize Marketplace, we’re now helping sellers in each of our geographies – and in countries where we don’t have a presence – reach out to our customers in countries outside their home geographies. We hosted merchants from more than 100 different countries last year, and helped them connect with customers in 185 nations.

Almost one-fifth of our overall third-party sales now occur outside the sellers’ home countries, and our merchants’ cross-border sales nearly doubled last year. In the EU, sellers can open a single account, manage their

business in multiple languages, and make products available across our five EU websites. More recently, we’ve started consolidating cross-border shipments for sellers and helping them obtain ocean shipping from Asia to Europe and North America at preferential, bulk rates.

Marketplace is the heart of our fast-growing operations in India, since all of our selection in India is offered by third-party sellers. Amazon.in now offers more selection than any other e-commerce site in India – with more than 20 million products offered from over 21,000 sellers. With our Easy Ship service, we pick up products from a seller and handle delivery all the way to the end customer. Building upon Easy Ship, the India team recently

piloted Kirana Now, a service that delivers everyday essentials from local kirana (mom and pop) stores to customers in two to four hours, adding convenience for our customers and increasing sales for the stores participating in the service.

Perhaps most important for sellers, we’ve created Fulfillment by Amazon. But I’ll save that for after we discuss Prime.

Amazon Prime

Ten years ago, we launched Amazon Prime, originally designed as an all-you-can-eat free and fast shipping program. We were told repeatedly that it was a risky move, and in some ways it was. In its first year, we gave up many millions of dollars in shipping revenue, and there was no simple math to show that it would be worth it. Our decision to go ahead was built on the positive results we’d seen earlier when we introduced Free Super Saver Shipping, and an intuition that customers would quickly grasp that they were being offered the best deal in the history of shopping. In addition, analysis told us that, if we achieved scale, we would be able to significantly lower the cost of fast shipping.

Our owned-inventory retail business was the foundation of Prime. In addition to creating retail teams to build each of our category-specific online “stores,” we have created large-scale systems to automate much of inventory replenishment, inventory placement, and product pricing. The precise delivery-date promise of Prime required operating our fulfillment centers in a new way, and pulling all of this together is one of the great accomplishments of our global operations team. Our worldwide network of fulfillment centers has expanded from 13 in 2005, when we launched Prime, to 109 this year. We are now on our eighth generation of fulfillment center design, employing proprietary software to manage receipt, stowing, picking, and shipment. Amazon Robotics, which began with our acquisition of Kiva in 2012, has now deployed more than 15,000 robots to support the stowing and retrieval of products at a higher density and lower cost than ever before. Our owned inventory retail business remains our best customer-acquisition vehicle for Prime and a critical part of building out categories that attract traffic and third-party sellers.

Though fast delivery remains a core Prime benefit, we are finding new ways to pump energy into Prime. Two of the most important are digital and devices.

In 2011 we added Prime Instant Video as a benefit, now with tens of thousands of movies and TV episodes available for unlimited streaming in the U.S., and we’ve started expanding the program into the U.K. and Germany as well. We’re investing a significant amount on this content, and it’s important that we monitor its impact. We ask ourselves, is it worth it? Is it driving Prime? Among other things, we watch Prime free trial starts, conversion to paid membership, renewal rates, and product purchase rates by members entering through this channel. We like what we see so far and plan to keep investing here.

While most of our PIV spend is on licensed content, we’re also starting to develop original content. The team is off to a strong start. Our show Transparent became the first from a streaming service to win a Golden Globe for best series and Tumble Leaf won the Annie for best animated series for preschoolers. In addition to the critical acclaim, the numbers are promising. An advantage of our original programming is that its first run is on Prime – it hasn’t already appeared anywhere else. Together with the quality of the shows, that first run status appears to be one of the factors leading to the attractive numbers. We also like the fixed cost nature of original programming. We get to spread that fixed cost across our large membership base. Finally, our business model for original content is unique. I’m pretty sure we’re the first company to have figured out how to make winning a Golden Globe pay off in increased sales of power tools and baby wipes!

Amazon designed and manufactured devices – from Kindle to Fire TV to Echo – also pump energy into Prime services such as Prime Instant Video and Prime Music, and generally drive higher engagement with every element of the Amazon ecosystem. And there’s more to come – our device team has a strong and exciting roadmap ahead.

Prime isn’t done improving on its original fast and free shipping promise either. The recently launched Prime Now offers Prime members free two-hour delivery on tens of thousands of items or one-hour delivery for a $7.99 fee. Lots of early reviews read like this one, “In the past six weeks my husband and I have made an embarrassing number of orders through Amazon Prime Now. It’s cheap, easy, and insanely fast.” We’ve launched in Manhattan, Brooklyn, Miami, Baltimore, Dallas, Atlanta, and Austin, and more cities are coming soon.

Now, I’d like to talk about Fulfillment by Amazon. FBA is so important because it is glue that inextricably links Marketplace and Prime. Thanks to FBA, Marketplace and Prime are no longer two things. In fact, at this point, I can’t really think about them separately. Their economics and customer experiences are now happily and deeply intertwined.

FBA is a service for Marketplace sellers. When a seller decides to use FBA, they stow their inventory in our fulfillment centers. We take on all logistics, customer service, and product returns. If a customer orders an FBA item and an Amazon owned-inventory item, we can ship both items to the customer in one box – a huge efficiency gain. But even more important, when a seller joins FBA, their items can become Prime eligible.

Maintaining a firm grasp of the obvious is more difficult than one would think it should be. But it’s useful to try. If you ask, what do sellers want? The correct (and obvious) answer is: they want more sales. So, what happens when sellers join FBA and their items become Prime eligible? They get more sales.

Notice also what happens from a Prime member’s point of view. Every time a seller joins FBA, Prime members get more Prime eligible selection. The value of membership goes up. This is powerful for our flywheel. FBA completes the circle: Marketplace pumps energy into Prime, and Prime pumps energy into Marketplace.

In a 2014 survey of U.S. sellers, 71% of FBA merchants reported more than a 20% increase in unit sales after joining FBA. In the holiday period, worldwide FBA units shipped grew 50% over the prior year and represented more than 40% of paid third-party units. Paid Prime memberships grew more than 50% in the U.S. last year and 53% worldwide. FBA is a win for customers and a win for sellers.

Amazon Web Services

A radical idea when it was launched nine years ago, Amazon Web Services is now big and growing fast. Startups were the early adopters. On-demand, pay-as-you-go cloud storage and compute resources dramatically increased the speed of starting a new business. Companies like Pinterest, Dropbox, and Airbnb all used AWS services and remain customers today.

Since then, large enterprises have been coming on board as well, and they’re choosing to use AWS for the same primary reason the startups did: speed and agility. Having lower IT cost is attractive, and sometimes the absolute cost savings can be enormous. But cost savings alone could never overcome deficiencies in performance or functionality. Enterprises are dependent on IT – it’s mission critical. So, the proposition, “I can save you a significant amount on your annual IT bill and my service is almost as good as what you have now,” won’t get too many customers. What customers really want in this arena is “better and faster,” and if “better and faster” can come with a side dish of cost savings, terrific. But the cost savings is the gravy, not the steak.

IT is so high leverage. You don’t want to imagine a competitor whose IT department is more nimble than yours. Every company has a list of technology projects that the business would like to see implemented as soon as possible. The painful reality is that tough triage decisions are always made, and many projects never get done. Even those that get resourced are often delivered late or with incomplete functionality. If an IT department can figure out how to deliver a larger number of business-enabling technology projects faster, they’ll be creating significant and real value for their organization.

These are the main reasons AWS is growing so quickly. IT departments are recognizing that when they adopt AWS, they get more done. They spend less time on low value-add activities like managing datacenters, networking, operating system patches, capacity planning, database scaling, and so on and so on. Just as important, they get access to powerful APIs and tools that dramatically simplify building scalable, secure, robust, high-performance systems. And those APIs and tools are continuously and seamlessly upgraded behind the scenes, without customer effort.

Today, AWS has more than a million active customers as companies and organizations of all sizes use AWS in every imaginable business segment. AWS usage grew by approximately 90% in the fourth quarter of 2014 versus the prior year. Companies like GE, Major League Baseball, Tata Motors, and Qantas are building new applications on AWS – these range from apps for crowdsourcing and personalized healthcare to mobile apps for managing fleets of trucks. Other customers, like NTT DOCOMO, the Financial Times, and the Securities and Exchange Commission are using AWS to analyze and take action on vast amounts of data. And many customers like Conde´ Nast, Kellogg’s, and News Corp are migrating legacy critical applications and, in some cases, entire datacenters to AWS.

We’ve increased our pace of innovation as we’ve gone along – from nearly 160 new features and services in 2012, to 280 in 2013, and 516 last year. There are many that would be interesting to talk about – from WorkDocs and WorkMail to AWS Lambda and the EC2 Container Service to the AWS Marketplace – but for purposes of brevity, I’m going to limit myself to one: our recently introduced Amazon Aurora. We hope Aurora will offer customers a new normal for a very important (but also very problematic) technology that is a critical underpinning of many applications: the relational database. Aurora is a MySQL-compatible database engine that offers the speed and availability of high-end commercial databases with the simplicity and cost effectiveness of open source databases. Aurora’s performance is up to 5x better than typical MySQL databases, at one-tenth the cost of commercial database packages. Relational databases is an arena that’s been a pain point for organizations and developers for a long time, and we’re very excited about Aurora.

I believe AWS is one of those dreamy business offerings that can be serving customers and earning financial returns for many years into the future. Why am I optimistic? For one thing, the size of the opportunity is big, ultimately encompassing global spend on servers, networking, datacenters, infrastructure software, databases, data warehouses, and more. Similar to the way I think about Amazon retail, for all practical purposes, I believe AWS is market-size unconstrained.

Second, its current leadership position (which is significant) is a strong ongoing advantage. We work hard – very hard – to make AWS as easy to use as possible. Even so, it’s still a necessarily complex set of tools with rich functionality and a non-trivial learning curve. Once you’ve become proficient at building complex systems with AWS, you do not want to have to learn a new set of tools and APIs assuming the set you already understand works for you. This is in no way something we can rest on, but if we continue to serve our customers in a truly outstanding way, they will have a rational preference to stick with us.

In addition, also because of our leadership position, we now have thousands of what are effectively AWS ambassadors roaming the world. Software developers changing jobs, moving from one company to another, become our best sales people: “We used AWS where I used to work, and we should consider it here. I think we’d get more done.” It’s a good sign that proficiency with AWS and its services is already something software developers are adding to their resumes.

Finally, I’m optimistic that AWS will have strong returns on capital. This is one we as a team examine because AWS is capital intensive. The good news is we like what we see when we do these analyses. Structurally, AWS is far less capital intensive than the mode it’s replacing – do-it-yourself datacenters – which have low utilization rates, almost always below 20%. Pooling of workloads across customers gives AWS much higher utilization rates, and correspondingly higher capital efficiency. Further, once again our leadership position helps: scale economies can provide us a relative advantage on capital efficiency. We’ll continue to watch and shape the business for good returns on capital.

AWS is young, and it is still growing and evolving. We think we can continue to lead if we continue to execute with our customers’ needs foremost in mind.

Career Choice

Before closing, I want to take a moment to update shareowners on something we’re excited about and proud of. Three years ago we launched an innovative employee benefit – the Career Choice program, where we pre-pay 95% of tuition for employees to take courses for in-demand fields, such as airplane mechanic or nursing, regardless of whether the skills are relevant to a career at Amazon. The idea was simple: enable choice.

We know that, for some of our fulfillment and customer service center employees, Amazon will be a career. For others, Amazon might be a stepping stone on the way to a job somewhere else – a job that may require new skills. If the right training can make the difference, we want to help, and so far we have been able to help over 2,000 employees who have participated in the program in eight different countries. There’s been so much interest that we are now building onsite classrooms so college and technical classes can be taught inside our fulfillment centers, making it even easier for associates to achieve these goals.

There are now eight FCs offering 15 classes taught onsite in our purpose-built classrooms with high-end technology features, and designed with glass walls to inspire others to participate and generate encouragement from peers. We believe Career Choice is an innovative way to draw great talent to serve customers in our fulfillment and customer service centers. These jobs can become gateways to great careers with Amazon as we expand around the world or enable employees the opportunity to follow their passion in other in-demand technical fields, like our very first Career Choice graduate did when she started a new career as a nurse in her community.

I would also like to invite you to come join the more than 24,000 people who have signed up so far to see the magic that happens after you click buy on Amazon.com by touring one of our fulfillment centers. In addition to U.S. tours, we are now offering tours at sites around the world, including Rugeley in the U.K. and Graben in Germany and continuing to expand. You can sign up for a tour at www.amazon.com/fctours.

***

Marketplace, Prime, and Amazon Web Services are three big ideas. We’re lucky to have them, and we’re determined to improve and nurture them – make them even better for customers. You can also count on us to work hard to find a fourth. We’ve already got a number of candidates in work, and as we promised some twenty years ago, we’ll continue to make bold bets. With the opportunities unfolding in front of us to serve customers better through invention, we assure you we won’t stop trying.

As always, I attach a copy of our original 1997 letter. Our approach remains the same, because it’s still Day 1.

Jeffrey P. Bezos

Founder and Chief Executive Officer

Amazon.com, Inc.