(For the exclusive use of E. Babalola, 2021.9 -

(
For the exclusive use of E. Babalola,
2021.
9
-
609
-
048
R
E
V
:
F
E
B
R
U

A
R
Y
3
,
2
0
1
2
)R O B E R T S . H U C K M A N G A R Y P . P I S A N O
L I Z K I N D
Amazon Web Services
“For years, Wall Street and Silicon Valley alike have rolled
their eyes at the legendary Bezos attention disorder. What’s the
secret pet project? Spaceships! Earth to Jeff: You’re a retailer.
Why swap pricey stuff in boxes for cheap clouds of bits?”1
— Spencer Reiss, Wired Magazine
When Andy Jassy joined online retailer Amazon.com, Inc.
(Amazon) after graduating from Harvard Business School in
1997, he never anticipated that he would end up running the
company’s developer business. Yet, in 2003, Jassy helped start
Amazon Web Services (AWS), Amazon’s storage, computing,
and Web-based technology services business. By 2008, Wall
Street analysts estimated that AWS generated about $46 to $92
million of revenue for Amazon annually2, and as of October
2008, served over 400,000 registered developers.
AWS was a unique business for Amazon, which was founded in

1994 as an online retailer of books but had since expanded to
include a wide array of products sold online—ranging from
electronics and music to housewares and toys. Since its
founding, Amazon had invested billions of dollars to build the
sophisticated Web infrastructure required to support its massive
retail business.
Amazon’s expansion beyond tangible consumer goods to selling
business services grew out of the company’s experience in
creating its own technical infrastructure. In July 2002, Amazon
decided to release product data in a developer-friendly format
that could be manipulated directly by Amazon’s third-party
affiliates—websites that advertised Amazon products and
received a portion of Amazon’s resulting sales. The positive
response to the release far exceeded the company’s
expectations, leading Amazon to consider creating a broader
developer-oriented business. Further, the company realized
that it was not alone in spending upwards of 70% of its time
building and maintaining the back-end technology “muck” that
did not differentiate Amazon from its competitors. Amazon
executives believed this paradox frustrated other companies in a
range of industries that might be interested in devoting more
attention to their products and customers and less time to back-
end infrastructure. In 2003, the company tapped Jassy, who had
been with Amazon for six years, to write the business plan
for and run the AWS division.
1 Spencer Reiss, “Cloud Computing. Available at Amazon.com
Today,” Wired Magazine, April 21, 2008, www.wired.com/
print/techbiz/it/magazine/16-05/mf_amazon, accessed July
2008.
2 Mylene Mangalindan, “Business Technology: Small Firms Tap
Amazon’s Juice—Web Services Unit Gains Popularity Renting
Storage, Server Capacity,” Wall Street Journal, January 15,
2008, p. B3, via Factiva, accessed July 2008.

Professors Robert S. Huckman and Gary P. Pisano and Senior
Researcher Liz Kind prepared this case. HBS cases are
developed solely as the basis for class discussion. Cases are
not intended to serve as endorsements, sources of primary data,
or illustrations of effective or ineffective management.
Copyright © 2008, 2012 President and Fellows of Harvard
College. To order copies or request permission to reproduce
materials, call 1-800-545- 7685, write Harvard Business School
Publishing, Boston, MA 02163, or go to
www.hbsp.harvard.edu/educators. This publication may not be
digitized, photocopied, or otherwise reproduced, posted, or
transmitted, without the permission of Harvard Business School.
(
This document is authorized for use only by Esther Babalola in
BUSO732 - Summer 2021 taught by JOSEPH
P.
BAILEY,
University of Maryland from Jul 2021 to Aug 2021.
)
(
2021.
)609-048Amazon Web Services

Amazon executives were convinced that the concept of turning
the company “inside out” and selling its storage, computing,
and other technology services to software developers would
grow into a significant business. At the same time, they
recognized AWS as an opportunity for Amazon to lead the
movement toward so-called “cloud computing”—the provision
of Internet-based information technology services—and the
ultimate creation of an Internet-based operating system for
computing.3 AWS initially targeted developers in startups and
small companies but expected its services eventually to be
attractive to large corporations as well.
By early October 2008, AWS had launched four major services.
Its two largest services—Elastic Compute Cloud (EC2) and
Simple Storage Service (S3)—provided scalable computing
power and allowed developers to rent any amount of storage
capacity they required on Amazon-owned servers. All the
AWS offerings were priced on a pay-as-you-go basis, allowing
developers to avoid the significant fixed costs that would
otherwise be required to manage their own data centers.
AWS had recently received a significant amount of positive
press regarding its leadership position in cloud computing.
Nevertheless, several outside observers believed that AWS
represented a distraction from Amazon’s core business and
that the services were unlikely to make a significant financial
contribution to the company’s bottom line. In addition, by
October of 2008, several high- profile companies such as
Google Inc. (Google), Microsoft Corporation (Microsoft), and
International Business Machines Corporation (IBM) were also
beginning to provide or announce intentions to offer Internet-
based technology services. Given the well-established brands,
broad customer bases, and existing sales infrastructures of these
companies, many experts considered them to be a significant
competitive threat. Despite these challenges, Jassy remained
confident that Amazon had the necessary assets and skill

sets to keep the company among the leaders in cloud computing
as the industry matured and grew.
Background on Amazon
Headquartered in Seattle, Washington, Amazon was founded by
Jeffrey Bezos with the aim of becoming “Earth’s biggest
bookstore.” The company launched its website in 1995 and grew
rapidly, reaching $15.7 million in sales in 1996 and $147.8
million in 1997. The company became known for its easy-to-
use, innovative site that incorporated personalization, customer
reviews, and “one-click” purchasing. Amazon went public in
1997 and raised $42 million.
In 1998, the company moved beyond books, initially adding
music, videos, toys, and electronics, all with the goal of turni ng
Amazon into the mass merchandiser of the online world. In
1999, Time Magazine named Bezos “Person of the Year.”
Around the same time, the company began allowing outside
merchants to operate virtual storefronts within Amazon’s
website for a fixed fee or sales commission. Amazon also began
providing e-commerce technology for retailers who wanted to
outsource their Web operations or did not have their own online
presence. For example, the company provided distribution and
website operations for several large independent merchants,
including
Target Stores Corp. and fashion retailer bebe Stores, Inc. By
October 2008, Amazon had grown to over
$17 billion in 12-month trailing revenue and had approximately
17,000 employees.
3 An Internet-based operating system represented a departure
from current operating systems that physically resided on the

machines using them. Instead, according to Gartner Research,
“massively scalable IT-related capabilities [were] provided as a
service, using Internet technologies to multiple external
[users].”
2
(
2021.
)Amazon Web Services609-048
Over the years, Amazon continued to broaden its consumer
product lines, adding clothing in 2002, video downloads in
2006, digital music in 2007, and a wireless portable reading
device called Kindle in 2008. By October 2008, Amazon offered
products in dozens of different retail categories, operated in
seven different countries, and had 1.4 million active seller
accounts and 10 million square feet of fulfillment center space.4
As noted in the company’s 2006 Form 10-K, Amazon “[sought]
to be Earth’s most consumer-centric company, where customers
can find and discover anything they might want to buy online,
and endeavor[ed] to offer customers the lowest possible
prices.”5
Since its inception, Amazon’s strategy was to grow fast, even at
the expense of profitability. Amazon invested heavily in
technology, both for website innovations and e-commerce
infrastructure. While in many ways effective, Amazon’s
approach drew substantial criticism from investors and research
analysts, many of whom voiced concern about the company’s
inconsistent margins. After many years of revenue growth,

Amazon reported its first full year of profitability in 2003. (See
Exhibit 1 for Amazon’s historical financial data for fiscal years
2002 to 2007.)
Amazon was known for its entrepreneurial culture, its use of
small, nimble teams, and its focus on long–term, rather than on
quarter-to-quarter, financial results. Bezos often spoke of
planting seeds to develop new businesses: “Planting seeds that
will grow into meaningful new businesses takes some
discipline, a bit of patience, and a nurturing culture In
our experience, if a new business enjoys
runaway success, it can only begin to be meaningful to the
overall company economics in something
like three to seven years.”6 Every year, the company’s annual
report included a copy of Bezos’ 1997 letter to shareholders
(Exhibit 2), emphasizing Amazon’s focus on creating value over
the long term, rather than on short-term profitability or Wall
Street reactions.
In assessing new business opportunities, Amazon considered a
number of criteria. According to Amazon’s 2006 annual report,
a new business had to be capable of (a) generating significant
returns, scaling substantially, (c) addressing an underserved
market, (d) being highly differentiated, and
(e) being an opportunity that Amazon was well-positioned to
provide.7 Amazon executives believed that cloud computing and
infrastructure Web services met each of these criteria.
Overview of AWSEvolution and History
Amazon started experimenting with Web services in the middle
of 2002 in conjunction with the company’s Associates group.
The Associates group consisted of hundreds of thousands of
businesses that advertised Amazon products on their websites.
Associates drove traffic to Amazon through specialized links
and earned commissions (typically 5% to 8%) for each sale they
referred back to Amazon. As Amazon sought to increase its
presence across Associate websites, the company decided to

release its product data—including prices, product titles, and
customer reviews—in an application program interface (API)
form, and let the Associates themselves determine how best to
present the information on their own sites to drive traffic and
sales.
4 Daniel H. Steinberg, “Web 2.0 Podcast: A Conversation with
Jeff Bezos,” O’Reilly Network, December 20, 2006,
http://www. oreillynet.com/pub/a/network/2006/12/20/web-20-
bezos.html, accessed April 2007.
5 Amazon.com, December 31, 2006 10-K (Seattle: Amazon.com,
2007), p. 3.
6 Amazon.com, 2006 Annual Report (Seattle: Amazon.com,
2007), p.3.
7 Ibid.
3
(
2021.
Jassy noted, “The API gave Web developers for these types of
sites raw material with which to work. It was a way for
developers to make the information come alive and build
interesting, customized presentations and displays on their sites.
Our goal was to have more success in driving sales back to
Amazon, which would also help the Associates make more
money.”

The results of the test program were dramatic. Jassy
commented:
We thought the idea had a chance of performing well, but we
really didn’t expect the response we received from developers.
However, there were three other things that probably surprised
us even more. The first was how excited developers were that
we were opening up part of our platform. They flocked to the
API in much larger numbers than we anticipated. Second, while
we knew that people would be creative with the data, we were
really surprised by their innovative building, especially on the
front end, to merchandise products. There were a lot of ideas
that we never would have gotten to ourselves nor would we
have had the resources to execute quickly on our own. The third
thing that surprised us was that people were building
commercial applications right from the start.
During the test release, the developers told us they would like
to see Amazon open up more of the platform in addition to
just some of the product data. At the same time, some other
properties—earlier entrants in our space—were starting to
experiment with exposing their assets via Web services. It
caused us to step back and wonder if something broader was
going on. If developers would build applications from scratch
using Web services, and if a broad array of Web services
existed (which we believed would be the case), then the Internet
would become the operating system. We asked ourselves, if
the Internet became the operating system, what would the
key elements be, which had already been built, and which
would Amazon be best-equipped to provide for the community?
At the time we were looking at it in 2003, none of the key
elements of the Internet operating system had been built.
When we thought about Amazon’s strengths as a technology
company that had simply applied its technology to the retail
space first, and what Amazon had done well over the last
decade, we realized we could provide a lot of the key building
blocks.
During a strategic planning meeting in 2003, the Amazon senior

management team decided to pursue the expansion of Web
services that would become AWS. Jassy elaborated:
We quickly recognized the broader developer business
opportunity and that, as a company, we had the unique skill set
to be a leader in the space. For example, most people
(excluding CIOs and CTOs) didn’t realize the extensive and
complex technology infrastructure required to operate Amazon.
You could not buy the software necessary to operate at
Amazon’s scale. You may have been able to buy pieces, but
they would have needed to be highly customized and carefully
strung together. Amazon built virtually every piece of software
necessary to run a Web business that could scale, on
demand, to virtually any level imaginable. Only a handful
of companies around the world could claim that level of
software competency.
To run Amazon.com, we had to build good services very deep in
the software stack— things like data storage, computing,
database functionality, and messaging. We learned a lot of
lessons, made a lot of improvements, and were forced to do
things at a significant scale and level of reliability. In many
ways, we had been working on the foundation of AWS since
Amazon’s inception but didn’t really know it.
In addition, Amazon had tremendous buying power and
leverage. The company spent over a decade building
relationships with the hardware and software companies that
provided
4
(
2021.

the necessary components of its technology infrastructure.
These companies also provided the hardware and software
necessary to run AWS. The relationships and buying leverage
translated into better terms for Amazon and, ultimately, AWS
customers.
Further, innovation, quality and speed of execution mattered a
lot in a race. Amazon had tremendous success operating,
organizing, and delivering on its first-mover advantage in the
retail business; as the first mover in this new developer
business, AWS hoped it could capitalize on some of the
same dynamics and similarly set the agenda.
Following the off-site that engendered AWS, Jassy and his
colleagues had to determine which services to build first. He
noted:
In determining where to start, we used three mechanisms. First,
we had a lot of feedback from internal teams about the things
they had to keep reinventing every time they constructed an
application. Second, we gathered about 25 business and
technology leaders from throughout the company and looked at
a variety of Internet applications that existed (or we thought
should exist) and then decomposed them. We evaluated which
Web services would be needed to make these applications
realities. Third, we got a lot of feedback from external
developers who’d been playing with the Associates’ API and
gave us suggestions about what else they wanted to see. The
feedback both from internal and external sources was
consistent, and we also got helpful feedback on how those
services needed to be delivered.
People said that no matter what we built, the services need[ed]
to meet a variety of requirements. They needed to be reliable so
that developers would feel confident that the services would be
highly available. The services also needed to be able to scale
seamlessly up or down as businesses grew or contracted.

Next, they needed to be low-latency. Just because we were
delivering services over the Internet didn’t mean they could be
slow. In fact, it was quite the opposite. They also needed to
be simple. People didn’t want it to take months to figure out
how to use the services. They wanted to read the
documentation and get going right away. Finally, the services
needed to be cost-effective.
In retrospect, the feedback was obvious. At the time, it
wasn’t as clear. The feedback we got from these various
constituents provided focus and guidance for the team on what
services to build and how to build them. It helped us develop
our core principles for the services.
Organizational Structure
Jassy assembled a team to develop foundational infrastructure
services. He noted, “We intentionally built the services feature -
poor and at a very low level, so that it would be easy to build on
top of them. There’s a risk in adding features in that they may
not end up being necessary.”
Jassy then organized small, autonomous teams for each service,
recruiting engineers for these teams both from within and
outside Amazon. Each team had a general manager and included
members with business backgrounds as well as those with
technical backgrounds. Jassy commented:
We’ve found that if you can get both the business and technical
personnel on the same team, the team tends to work more
collaboratively. It’s very useful for product managers to work
alongside engineers. They learn about each others’ problems,
there are fewer misunderstandings, and there is no “us versus
them.” We also decided to put several recent MBAs in
significant leadership product management jobs. This was a
function of my own belief, from my experience at Amazon, that
smart, motivated MBAs can surprise folks with
5

(
2021.
what they can do and how they can lead. In addition, there were
so many teams starting simultaneously within Amazon that there
were a number of great opportunities for people to rise to the
occasion.
The teams were responsible for defining, building, operating,
and running the service like a business with its own profit-
and-loss statement. AWS staff worked very closely with
Amazon’s infrastructure team and used several of Amazon’s
global data centers. According to Jassy, each service
typically took somewhere between several months to around two
years of development to be operational.
AWS Services
By early October 2008, AWS was offering 12 services, four of
which the company defined as “Infrastructure Web Services.”
These four were Amazon Simple Storage Service (Amazon S3),
Amazon Elastic Compute Cloud (Amazon EC2), Amazon Simple
Database (Amazon SimpleDB), and Amazon Simple Queue
Service (Amazon SQS). All four services primarily targeted
software developers working in companies of all sizes. The
computing and storage services, EC2 and S3, were intended to
meet the broad needs of developers while providing an
inexpensive, scalable alternative to buying or building in-house
data centers.
Simple Storage Service (S3)

In March 2006, AWS launched S3, which provided a basic
interface for storing and retrieving data from anywhere on the
Web. Developers could store and access unlimited amounts of
data, using the same technology created for Amazon’s internal
operations. To ensure secure storage, each object uploaded to
S3 was replicated immediately and multiple copies were stored
in multiple locations.
U.S. clients paid $0.12 to $0.15 per gigabyte per month for
storage; $0.10 per gigabyte of data uploaded; and $0.10 to $0.17
per gigabyte of data downloaded, depending on the size of the
download. By October 2008, S3 held roughly 29 billion unique
objects,8 up from 800 million in July 2006.9 According to
Jassy, at its peak to date, S3 processed 70,000 requests per
second.
A variety of companies used S3, ranging from entrepreneurial
concerns to large corporations. For example, SmugMug, Inc., a
startup offering online storage of customers’ photos, used S3 to
back up stored images. At the other end of the spectrum,
Microsoft used S3 to distribute copies of its Vista software.
Elastic Compute Cloud (EC2)
In August 2006, AWS released EC2, “a Web service that
provided resizable computing capacity in the cloud.”10 The
service, available on demand through the Internet, provided
developers the ability to run on Amazon’s proven computing
environment. EC2 was highly elastic, allowing for quick scaling
up and down, and designed to work in conjunction with S3.
Amazon charged EC2 users $0.10
8 This figure reflects pieces of data uploaded by users; because
S3 stored three copies of each object, S3 held a total of roughly
15 billion pieces of data on its servers as of April 2007.
9 Brad Stone, “Sold on eBay, Shipped by Amazon,” New York
Times, April 27, 2007, available from Factiva, accessed July
2007.
10 Amazon.com website, http://aws.amazon.com/ec2, accessed

October 2008.
6
(
2021.
to $0.80 per instance hour, depending on the size and
characteristics of the instance;11 $0.10 per gigabyte for data
uploaded; and $0.10 to $0.17 per gigabyte for data downloaded,
depending on the amount of data downloaded. Bezos
commented, “If you work out how many hours there are in a
month, it’s about $70 per month to have a server that’s
equivalent to a 1.7 gigahertz x86 box. But the really cool thing
is that it’s elastic, so you can either have one server a month
for $70 or you can have 700 servers an hour for $70.”12
Powerset Inc., a natural language search company that
Microsoft purchased, was using EC2 to build a consumer search
engine. Jassy noted, “Powerset is building a natural-language
index that crawls the Web and runs algorithms to figure out the
right natural language equivalents for different terms. It takes
huge amounts of computing resources, so it is very attractive for
Powerset to use EC2 and avoid building data centers.”
Executives at Powerset believed that, by using Amazon’s EC2,
they cut their first-year capital costs by more than half.13
Another startup, AideRSS, estimated it was saving over $12,000
per month by working with Amazon Web Services (Exhibit 3).
AideRSS provided a filtering tool for Internet news stories and
blog headlines. When the tool was launched in July 2007, traffic
levels greatly exceeded expectations, creating the need for rapid

scaling. AideRSS relied on EC2 to manage its ramp-up from 10
instances to 100 instances in its first 24 hours.
SimpleDB
In December 2007, AWS introduced SimpleDB, a Web service
for providing real-time lookup and querying of structured
data.14 SimpleDB was designed to work in conjunction with
S3 and EC2, adding further to AWS’s portfolio of services.
SimpleDB was not intended to replace sophisticated relational
database operations, but to provide an easy-to-use alternative
with the core functionality of a database. Clients paid $0.14
per Amazon SimpleDB machine hour consumed; $0.10 per
gigabyte of data transferred in; and $0.10 to $0.17 per gigabyte
of data transferred out, depending on the size of the transfer.
Structured data storage was $1.50 per gigabyte per month.
Simple Queue Service (SQS)
In 2004, Amazon built SQS, a scalable, hosted queue that stored
data messages as they traveled between applications. According
to Amazon, the service allowed developers to “move data
between distributed application components performing
different tasks, without losing messages or requiring each
component to always be available.”15 In other words, SQS
allowed components of an application to send messages to
other components and created a place to store these messages
for pick-up by the consuming components. Users paid $0.01 per
10,000 Amazon SQS requests sent; $0.10 per gigabyte for data
uploaded; and $0.10 to $0.17 per gigabyte for data downloaded,
depending on the amount of data transferred.
11 An instance was equivalent to one hour of processing time
on one computer.
12 Daniel H. Steinberg, “Web 2.0 Podcast: A Conversation with
Jeff Bezos,” http://www.businessweek.com/magazine/
content/06_46/b4009001.htm, accessed March 2011.
13 Robert Hof, “Jeff Bezos’ Risky Bet,” BusinessWeek,
November 13, 2006.

14 PC Magazine defines structured data as “data that resides in
fixed fields within a record or file. Relational databases and
spreadsheets are examples of structured data.” Source:
PCmag.com Encyclopedia,
http://www.pcmag.com/encyclopedia_term/0,2542,t=structured+
data&i=52162,00.asp, accessed September 9, 2008.
15Amazon.com Company website, “Amazon Simple Queue
Service (Amazon SQS),” http://aws.amazon.com/sqs, accessed
May 2007.
7
(
2021.
Other Major Services
Amazon Flexible Payment Service (Amazon FPS) In August
2007, AWS launched Amazon FPS, designed specifically for
developers to collect funds electronically from customers. The
service allowed developers access to the same payment
services used by Amazon.com’s retail business and the same
terms Amazon.com had negotiated with credit card and bank
processors. While the functionality of Amazon FPS was often
compared to PayPal, it differed in that it was the first
payment service designed from the ground up for software
developers. As a result, developers had unmatched flexibility
in defining the conditions under which transactions could take
place. For instance, with Amazon FPS, developers could
construct virtually any payment instruction they wanted to

attach to a transaction, and as long as all participants in the
transaction supplied matching payment instructions, the
transaction would succeed.
Amazon Premium Support In April 2008, AWS launched
AWS Premium Support to provide customers with one-on-
one technical assistance for S3, EC2, and SQS. The support
offering was available in two different plans. AWS Premium
Support (Silver) provided technical support, with response
times ranging from four business hours to two business days,
depending on the severity of the issue. Pricing for the Silver
Plan was the greater of $100 per month or $0.10 per dollar of
total monthly usage of Amazon S3, EC2, and SQS. AWS
Premium Support (Gold) provided 24- hour telephone
support with response times of one hour for urgent issues. The
Gold Plan was the greater of $400 per month or $0.10 to $0.20
per dollar, based on volume, of total monthly usage of Amazon
S3, EC2, and SQS.
Amazon Elastic Block Store (EBS) In August 2008, AWS
released Amazon EBS, which provided block-level storage
volumes for use with EC2. Before EBS, when EC2 instances
were terminated, the data within the instance was lost. EBS
allowed users to allocate storage volumes that persisted
independently from EC2 instances. In September 2008, AWS
announced a content delivery service to be released by the end
of 2008.16 In early October 2008, AWS announced it would
start offering Amazon EC2 for use with the Windows operating
system sometime during the fall of 2008.17 According to Jassy,
several additional services were in development.
Competition and Market Dynamics
Deutsche Bank estimated the digital infrastructure market AWS
targeted to be roughly $196 billion in 2006, growing to over
$225 billion by 2009.18 The investment-banking firm predicted
that while Amazon was unlikely to gain meaningful revenue
from its AWS division in the short term, the business had the

potential to generate an incremental $200 million in annual
revenue down the road, with the possibility of a 50%
incremental margin.19 Forbes magazine reporter Quentin Hardy
estimated that Amazon’s Web Services business could have a
gross margin of 45%.20
16 See http://www.amazon.com/gp/html-forms-controller/aws-
content-delivery-service for details.
17 EC2 previously had only been offered with the Linux
operating system.
18 Jeetil Patel and Herman Leung, “Amazon.com: Upgrading to
BUY on Potential Margin Turnaround/Web Svcs,” Deutsche
Bank Global Markets Research, April 15, 2007.
19 Ibid.
20 Quentin Hardy, “The Death of Hardware; Why Buy
Computers When You Can Rent Them from Amazon.com, EMC
or Yahoo?” Forbes, February 11, 2008, p. 36, available from
Factiva, accessed July 2008.
8
(
2021.
On the server and computing side, major players such as IBM,
Sun Microsystems, Inc. (Sun), and Hewlett-Packard Company
(HP) dominated the traditional information technology
infrastructure space. All three companies offered outsourcing
services and were developing initiatives to provide Web-based

infrastructure and software solutions. However, only Sun
offered a product similar to EC2, with variable pricing (e.g., per
hour) for processing power. In March 2006, Sun launched its
“Sun Grid” utility service, which offered computing power at $1
per processor per hour.21
The traditional storage market was highly competitive and
dominated by large corporations such as Network Appliance,
Inc., EMC Corporation, IBM, HP, Sun, and Dell Computer
Corporation. However, few of these large companies
provided pay-as-you-go online storage similar to Amazon’s
S3 offering. During the dot-com boom of the late 1990s, several
new storage service providers (SSPs)—companies offering
Web-based data storage and backup services—entered the
market, but many of them no longer existed. (See Exhibit 4 for
a description of the pricing structures for several SSPs
remaining in the market.)
According to Gartner Research, “In one of the most profound
ironies of the SSP market, the future of storage as a service is
brighter today than when most SSPs brought the concept to
market. Customers of large outsourcing and managed hosting
companies, such as AT&T, EDS and IBM, continue to warm to
the concept of paying for storage as a service. Many large
outsourcers believe that by 2012 more than two-thirds of the
storage they provision will be paid for on some sort of usage
basis.”22
At a high level, Amazon was at the front end of a broader move
by other Internet giants— including Google, eBay Inc., Yahoo!
Inc., and Microsoft—to host Web-based services for consumers
and businesses and remain omnipresent during the next wave of
the Web’s development. For example, in August 2006, Google
began testing Google Apps for Your Domain—an online office
software suite that allowed companies to offload e-mail systems
to Google while maintaining their existing e-mail addresses.
Google later added word processing, spreadsheet and calendar
services, and a Web page builder, all for a few dollars per
person per month. Google also offered other programmable

components including Web search, maps, chat, and advertising.
In April 2008, Google introduced App Engine, which allowed
developers to build and host Web applications on Google’s
scalable infrastructure. Though somewhat limited in its scope
compared to AWS, App Engine was expected to evolve and
compete with AWS.23 For example, App Engine’s capabilities
were restricted to online Web applications, rather than offline
batch computations.24 In addition, Google’s initial version of
App Engine only supported code written in the Python
programming language, “a scripting language used widely in
Google and elsewhere, but far from universal.”25 Google said it
would add other languages over time. Pricing was free for
applications
21 Martin LaMonica, “Amazon Servers, Starting at 10 Cents an
Hour; The Elastic Compute Cloud Service, From Amazon’s Web
Services Division, Pipes Processing Power Over the Internet,”
CNET News.com, August 26, 2006, available from Factiva,
accessed April 2007.
22 Adam W. Couture, “Alternative Delivery Models: Why Buy
Storage When You Can Rent It?,” Gartner Research ID Number
G00150752, August 30, 2007, p. 2, available from Gartner
Research, accessed August 2007.
23 Ray Valdes, “Google App Engine Goes Up Against Amazon
Web Services,” Gartner Research, April 11, 2008, p. 2, accessed
March 2011.
24 Ibid., p.3.
25 Stephen Ellis, “Google Comes from the Clouds to Tackle
Microsoft with Office Services,” Australian, April 15, 2008, p.
30, available from Factiva, accessed July 2008.
9
(
or the exclusive use of E. Babalola,

2021.
)F
609-048Amazon Web Services
that generated up to five million views per month and required
less than 0.5 gigabyte of storage. After that, Google charged
users $0.10 to $0.12 per CPU core-hour; $0.15 to $0.18 per
gigabyte-month of storage; $0.11 to $0.13 per gigabyte of
outgoing bandwidth; and $0.09 to $0.11 per gigabyte of
incoming bandwidth.26
In 2005, Salesforce.com began offering on-demand customer
relationship management (CRM) services and later that year
created AppExchange, an online marketplace where developers
could swap and sell applications. In 2007, Salesforce.com
announced the launch of force.com, a platform for developers to
create business applications on demand and without software. In
2005, Microsoft introduced two families of Web services. The
first, Windows Live, provided a suite of consumer software
services including an e-mail program, news headlines, blog and
audio feeds, and a Web page builder. The second, Office Live,
included a similar range of services and was aimed at small
businesses. Microsoft planned to offer some of the services for
free, supporting the costs through advertising revenue. In June
2008, Microsoft acquired MobiComp, an open-source cloud
computing company that specialized in the storage, backup, and
restoration of mobile data. Microsoft was widely expected to
launch a comprehensive range of cloud computing services
during its October 2008 Professional Developers Conference.
AWS management believed its services were well positioned in
the market, particularly with respect to ease of use, pricing,
scalability, performance, and the ability for developers to get up
and running quickly. Further, unlike other technology
companies pursuing cloud computing, Amazon did not have

a vested interest in how its customers used its services because
it did not provide other software products or services for which
there might be risk of abandonment or cannibalization.
Nevertheless, in addressing AWS offerings relative to those
of other companies, one Amazon manager noted, “We don’t
spend a lot of time looking at our competitors. A competitor -
focused strategy is a reasonable one, but it’s not Amazon’s
strategy.”
Jassy added:
Even though we expect more companies to provide services in
this space—it’s too good of a value proposition for more
companies not to want to pursue it—we are really pleased with
how rapidly the business has taken off. We’ve quickly gotten a
lot of traction with customers, and many of them are using
multiple AWS services. They have been very happy with the
functionality and performance of our services we’ve released.
At the same time, we’ve gained invaluable experience operating
these services at significant scale over the last two years. We’ve
learned what new features and services our customers would
like us to provide and are working hard to develop them.
Challenges for AWSPricing and Business Model
AWS management spent a considerable amount of time
determining how to price its services. Pricing for Amazon.com’s
retail business was relatively straightforward. By comparison,
the AWS team had to consider the amount of bandwidth that
would be consumed, how much storage would be used, and how
many requests would actually be processed. The company
decided to price very aggressively, as Amazon had traditionally
done with its businesses. The goal was to get many people
26 “Google’s Developer Strategy Rests on the Cloud,” CMP
TechWeb, May 28, 2008, available from Factiva, accessed July
2008.

10
(
2021.
using the services as quickly as possible. In addition, the more
scale AWS could build, the more it would drive down costs for
bandwidth and hardware.
Most other companies that provided substitutes for AWS’s
offerings tended to charge both an up- front fee and a
subscription fee with a guaranteed minimum to ensure a certain
amount of revenue from each customer. In comparison, AWS
did not impose an up-front fee and its pricing was exclusively
based on a pay-per-use pricing model. For example, some AWS
customers had monthly bills of less than $1. As of October
2008, Amazon executives remained optimistic and patient about
the financial performance of the AWS division. The early
results were promising and Amazon executives noted that the
company always believed it would take several years before
AWS was a meaningful, free-cash-flow-generating business for
Amazon.
Outside Reaction to AWS
Amazon management was pleased with the progress of AWS,
and in particular with its success in developing credibil ity
within the developer community. As of October 2008, AWS had
over 400,000 registered developers, up from 160,000 in 2006.
Amazon S3 objects grew from 5 billion in April 2007 to 29
billion in October 2008. In the spring of 2007, S3 won the

CODiE Award for best storage software. Jassy explained the
main reasons Amazon’s Web Services resonated with
developers:
The first thing any startup or company that’s building a
software application knows is that having reliable, scalable,
inexpensive storage or computi ng doesn’t differentiate their
business. Yet, they also know that if [they] don’t have highly
reliable, scalable, inexpensive storage or computing, they are
dead on arrival.
One of the things we tell customers is, “We make muck so you
don’t have to.” This resonates with developers because they
know what it means to actually have to negotiate bandwidth
contracts, build out highly available data transfer networks, buy
and install servers, scale and manage that hardware, operate
datacenters for 24/7/365 availability, and so on.
Another reason developers like AWS’s services is that they
allow businesses to get to market faster than they would
otherwise. They don’t have to hire additional people, and they
don’t have to figure out how to build a scalable architecture.
They don’t have to negotiate internally for more servers or to
test a new service or feature. It’s very easy and provides
instant capacity. Getting started on S3 and EC2 can be a matter
of minutes or just a few days, depending on the way it is going
to be used.
Finally, I think another reason developers like AWS has to do
with cost. There are no significant up-front capital expenditures
on servers and datacenters. You only pay for what you use. As a
startup, if you don’t have to take hard-earned venture capital or
angel-backed financing and spend it on servers, you can take
those funds and apply them to the pieces of your business idea
that can make or break your company. The same goes for
enterprises. They don’t really want to invest all that capex and
people resources in building out their infrastructure. They do it
because they haven’t previously had a viable alternative. In
general, and especially in this economic environment, it is very
appealing to budget for the features or people that differentiate

your applications and business, rather than on servers or
datacenter operations that do not.
Jassy’s group found that grassroots, viral marketing was
particularly effective in promoting AWS. Though AWS had
received considerable coverage from traditional business
publications, Jassy found
11
(
2021.
that blog and media coverage added credibility and dramatically
increased awareness. He noted, “We’ve been fortunate to
receive a lot of attention over the past few years in this space.
Developers are genuinely excited about what we’re doing and
they’ve tended to write about it. In this space, blogs and buzz
are really influential. When word gets out from the right
sources, it spreads rapidly.”
In September 2007, AWS introduced the “Amazon Web Services
Start-up Challenge,” a contest targeted at entrepreneurs and
developers. Jassy explained:
Earlier in 2007, we tested a start-up event, with the idea of
educating people about our services, sharing local startups’
experiences with AWS, and hosting a cocktail and networking
session at the end. The number of attendees was four times what
we thought it would be, and the response was so positive that

we rolled it out to six more cities. We timed these additional
events to coincide with the launch of our first-ever AWS Start-
up Challenge. Startups were asked to submit a summary
business plan and explain how they were using AWS. We
received over 900 entries from across the country before
awarding a golden hammer (symbolic for being able to destroy
all of your servers without worries) to Ooyala, a video concern
that took home $50,000 in cash, $50,000 in AWS credits, and an
investment offer from Amazon.
In addition to the start-up challenge, AWS developed key
partnerships with large platform companies to increase the
visibility of its offerings in the developer community. For
example, Facebook developers could use AWS’s computing
infrastructure to scale their applications, MySQL Enterprise
(owned by Sun) subscribers received database support for EC2,
and Oracle Corporation database and middleware solutions
could be deployed to the AWS cloud or backed up to S3.
While developers and technical experts were generally positive
about AWS, reactions from the press and Wall Street analysts
were mixed. The majority of analysts covering Amazon’s stock
either took little notice of AWS or felt that it was a diversion
from the company’s core business and a drain on profitability.
Piper Jaffray analyst Safa Rashtchy commented, “I have yet to
see how these investments are producing any profit. They’re
probably more of a distraction than anything else.”27 Even
analysts from Deutsche Bank Securities, Inc., who looked
favorably upon AWS, noted, “When we take a step back, we
find it difficult to reconcile how an online retailer such as
Amazon spends
$600-plus million annually on technology and content spending.
In fact, this number should be closer to $260 million . . . unless
the company was investing in technology initiatives that had
little to do with its core online retailing operations and
everything to do with Web services for other companies and
startups in the market.”28 Jassy noted:
A lot of people ask us, “Why is Amazon doing this

business?” While it is very different from our traditional
consumer business, we believe that there is a very large market
and that it is a very large opportunity for the company. A lot
of people still don’t realize it, but we really are a technology
company. We’re good at it because we had to be good at it in
order to run our massive, worldwide Web business. We don’t
expect that it’s going to be successful overnight, but we do
believe it will be a very significant, differentiated free-cash-
flow-generating business for the company.
Bezos’ response to those skeptical of AWS was quite succinct:
“We’re very comfortable being misunderstood. We’ve had lots
of practice.”29
27 Robert D. Hof, “Jeff Bezos’ Risky Bet.”
28 Jeetil Patel and Herman Leung, “Amazon.com: Upgrading to
BUY.”
29 Robert D. Hof, “Jeff Bezos’ Risky Bet.”
12
(
2021.
While most industry observers felt it was too early to predict
the success of AWS, some praised its service offerings and saw
substantial potential for the business. In their coverage of
Amazon, analysts at Stifel Nicolaus & Company, Inc. wrote,
“We look at Amazon’s new offerings as a play on the Internet

continuing on its path toward open-source technology and that
Amazon is well-positioned in such an environment.”30 Trip
Chowdhry, an analyst with Global Equities Research, noted,
“Amazon is years ahead of anyone else when it comes to ‘cloud
computing.’ Tomorrow’s computing environment is being
dictated by Amazon.com.”31
Capacity Management and Service Releases
Amazon launched S3 and SQS in a general availability format,
which meant that the service was available to anyone. Jassy
explained the ramifications of this approach:
It’s not like working with traditional hosting facilities where
you call your favorite vendor, tell them about your application,
describe what you want to store and what you want to host, and
then negotiate a price. The whole process could take a few
weeks. With our service, you read the documentation and you
go. As a result, we just don’t know exactly how many new
customers will come on board. In the beginning, it’s challenging
to know how much capacity to order and hold. We want to have
capacity for both new developers and for existing customers to
grow, but we don’t want to significantly over-provision. If we
sit on a lot of excess capacity for a long time, there are real
costs associated with that under-utilized hardware.
To manage the uncertainty surrounding the response to a new
service launch, AWS management also used unlimited public
beta and limited public beta releases. An unlimited public beta
launch allowed the service to be available to anyone who
wanted it, but with the understanding that the service was still
in beta mode. A limited beta release, where the service was still
in beta, controlled the number of users by opening up the
service gradually, a few hundred users at a time, rather than
through a full commercial offering. Jassy’s team decided to use
a limited beta release for EC2. Within four hours of the
service’s 2:00 a.m. launch, all of the slots offered to developers
were taken. Over the course of the next year, Amazon continued
to open a few hundred slots at a time before releasing the beta
on an unlimited basis in June 2007. Jassy reflected on the

launch:
We found that EC2 being in beta release did not stop people
from using the service. I think users understand that beta means
there may be some fine-tuning we’re doing, but they also
assume that when you’re talking about a company like Amazon,
the services are going to work. The good news is that releasing
EC2 in a beta form allowed us to manage the scale in a more
staged fashion than with S3.
In December 2006 and January 2007, Amazon’s S3 suffered
some minor performance glitches, due largely to problems with
faulty hardware installed during an upgrade.32 In an online
discussion forum, an Amazon representative explained, “The
Amazon S3 team has been adding large amounts of
hardware over the past several weeks in order to meet and stay
ahead of high and rapidly
30 Stiefel Nicolaus & Company, Inc., “Internet Consumer
Services, Amazon.com, Inc.,” March 1, 2007, p. 3.
31 Patrick Seitz, “Firms See Growth in This Web Tech;
Applications, Savings, and Security Are a Couple of Pluses that
Keep Outfits Plugged In,” Investor’s Business Daily, June 4,
2007, via Factiva, accessed July 2007.
32 Martin LaMonica, “Amazon’s Hosted Service Hits Bump;
Performance Glitches Make Some Customers of Amazon’s
Nascent Web Service for Storage Grouchy,” CNET News.com,
January 8, 2007, via Factiva, accessed July 2007.
13
(

2021.
increasing demand. Unfortunately, our most recent order
contained several substandard machines.”33 In February 2008,
AWS suffered an additional temporary outage. The company
offered a status report dashboard for AWS (see Exhibit 5 for an
example) and a service level agreement that guaranteed 99.9%
uptime for S3. Jassy commented on the issues Amazon faced
with regard to scale and capacity:
A lot of people assume that we started AWS primarily as a way
to leverage excess capacity, when in fact, we are long past that
point and have bought lots of additional bandwidth and servers.
In fact, bandwidth consumed by AWS customers is more than
that consumed by Amazon’s retail businesses. (See Exhibit 6.)
With S3, the demand was such that early on we quickly found
ourselves in a situation where we just couldn’t get hardware in
fast enough. I think we probably underestimated how good we
had to be at capacity planning. We had many capabilities from
our core retail business. We had added servers and had gained
experience operating and scaling them. However, we hadn’t had
a service that allowed people to get started seamlessly and
add capacity and scale without a lot of manual intervention.
Our AWS offerings are truly self-service products. Customers
don’t even have to talk to us. They can go to our website, read
the information, and start making Web services requests to add
data, download that data, etc. We first learn about new
customers when we look at the activity reports for the day. Very
quickly, new customers can be sending us significant amounts
of volume.
Jeff Barr, AWS’s lead evangelist, blogged about an example
with start-up client Animoto:
Built on top of Amazon EC2, S3, and SQS, [Animoto] allows

you to upload a series of images. It then generates a unique,
attractive, and entertaining music video using your own music
or something selected from the royalty-free library on the site
After the images and
the music have been uploaded, proprietary algorithms analyze
them and then render the final video. This can take an
appreciable amount of time and requires a considerable amount
of computing power.
Animoto introduced the Animoto Videos Facebook
applicationand it had done pretty
well . . . so [they] . . . decided to step it up a notch. . . . They
made a subtle but important change to their application: they
auto-created a user's first Animoto video.
That did the trick! They had 25,000 members on Monday,
50,000 on Tuesday, and 250,000 on Thursday. Their EC2 usage
grew as well. For the last month or so they had been using
between 50 and 100 instances. On Tuesday their usage peaked
at around 400, Wednesday it was 900, and then 3,400 instances
as of Friday morning. . . . [See Exhibit 7.] We are really happy
to see Animoto succeed and to be able to help them to scale up
their user base and their application so quickly. I’m fairly
certain that it would be difficult for them to get their hands on
nearly 3,500 compute nodes so quickly in any other way.
Other clients that took advantage of AWS’s elasticity in scaling
included the New York Times (the Times). Derek Gottfrid,
senior software architect for the Times, wrote about the
paper’s experiences with S3 and EC2 in his blogs. In the fall of
2007, the company needed to convert and store in digital format
its archive of full-page scans of the newspaper from 1851
through 1980. The Times used AWS to upload the data from
11 million articles onto S3. Next, Gottfrid wrote the code to
read the data on EC2 and create PDF versions of the articles.
The results were then stored back on S3, where the PDF
33 Ibid.

14
(
2021.
files were available to the general public. The project was
completed in less than 24 hours, using 100 EC2 instances and
generating 1.5 terabytes of data stored in S3. In May 2008, the
Times used AWS again to create a browseable interface for the
data in its archives. Gottfried wrote, “I was using some very
new and not totally proven pieces of technology on a project
that was very high profile and on an inflexible deadline.
Now that this adventure can be called a success, I can’t imagine
how we
might have done it without Amazon S3/EC2. The one caveat I
will offer . . . is that it is highly addictive.”34
Looking Ahead
As Jassy reflected on the evolution of AWS, he was proud of
the business’s success to date and convinced that it had
enormous potential. Earlier that summer, analysts at Gartner
Research noted that “cloud computing heralds an evolution no
less influential than e-business.”35 Similarly, writer Nicholas
Carr predicted cloud computing would put most IT departments
out of business.36 Jassy was confident that Amazon’s
technological competence, ability to operate reliable and
scalable services cost-effectively, first-mover advantage in this

nascent space, and ability to execute quickly positioned AWS
well. Most observers agreed AWS was the front runner in
providing a platform of “cloud” services. One prominent
venture capitalist commented, “Amazon is out-Googling
Google.”37
At the same time, AWS would continue to face competitive
threats. While Google’s App Engine was initially limited to a
single programming language and narrower than AWS in
technical implementation, Google was expected to add
languages and features over time. Other competitors such as
Microsoft and IBM were also rumored to be developing Web
services products directly comparable to those AWS offered.
These technology firms had long-established brands, existing
enterprise customers, and large sales forces. AWS, by
comparison, had focused initially on developers, and Amazon
had a brand that was primarily associated with the company’s
retail and e- commerce business. Further, Amazon continued to
face substantial scrutiny from capital markets that were already
skeptical about the company’s technology expenditures and
profitability levels. AWS’s overall business model was geared
toward driving high volume and utilization of its products and
services to reduce costs and increase free cash flow; Jassy
recognized the need to expand and broaden the mix of AWS’s
customer base.
In spite of AWS’s success, at least one analyst felt that
“Amazon still has work to do to make its services ready for
corporate customers.”38 Nonetheless, Jassy and his AWS team
continued to move aggressively forward with the belief that
AWS had the core assets and unique set of skills necessary to
continue its leadership position in cloud computing, and in turn,
develop a meaningful free cash flow business for Amazon.
34 Derek Gottfrid, “Self-Service, Prorated Super Computing
Fun!” Open—All the Code that’s Fit to Print, New York Times,
November 1, 2007.

35 Chris Nuttal, “The IT World Pokes Its Head in the Cloud,”
FT Report—Digital Business 2008, Financial Times, July 9,
2008,
p. 2, via Factiva, accessed July 2008.
36 Bill Snyder, “Cloud Computing: Not Just Pie in the Sky,”
CIO, March 5, 2008, via Factiva, accessed July 2008.
37 Spencer Reiss, “Cloud Computing. Available at Amazon.com
Today,” Wired Magazine, April 21, 2008.
38 Mylene Mangalindan, “Business Technology: Small Firms
Tap Amazon’s Juice—Web Services Unit Gains Popularity
Renting Storage, Server Capacity,” Wall Street Journal, January
15, 2008, p. B3, via Factiva, accessed July 2008.
15
(
2021.
Exhibit 1Amazon Historical Financial Data ($ in millions)
For the Fiscal Years Ended December 31,
2002
2003
2004
2005

2006
2007
Operating Data:
Net Sales
$3,933
$5,264
$6,921
$8,490
$10,711
$14,835
Gross Profit
992
1,257
1,602
2,039
2,456
3,353
Operating Expenses:
Fulfilment
405
495
601
745
937

1,292
Marketing
129
128
162
198
263
344
Technology and Content
252
257
283
451
662
818
General and Administrative
95
104
124
166
195
235
Other Operating Expenses (income)
47
3
(8)
47
10
9
Operating Income
64
270
440
432
389
655

Income Before Income Taxes
(151)
39
355
428
377
660
Net Income (loss)a
(150)
35
588
333
190
476
Balance Sheet Data:
Cash, Equivalents, and Marketable Securities
$1,301
$1,395
$1,779
$2,000
$2,019
$3,112
Total Assets
1,990
2,162
3,248
3,696
4,363
6,485
Long-Term Debt

2,230
1,919
1,835
1,480
1,247
1,282
Total Shareholders' Equity
(1,353)
(1,036)
(227)
246
431
1,197
Source: Amazon.com, Inc. website, Investor Relations page.
a Before cumulative effect of change in accounting principle.
16

(
2021.
609-048
)Amazon Web Services
Exhibit 21997 Letter to Shareholders
1997 LETTER TO SHAREHOLDERS(Reprinted from the 1997
Annual Report)
To our shareholders:
Amazon.com passed many milestones in 1997: by year-end, we
had served more than 1.5 million customers, yielding 838%
revenue growth to $147.8 million, and extended our market
leadership despite aggressive competitive entry.
But this is Day 1 for the Internet and, if we execute well, for
Amazon.com. Today, online commerce saves customers money
and precious time. Tomorrow, through personalization, online
commerce will accelerate the very process of discovery.
Amazon.com uses the Internet to create real value for its
customers and, by doing so, hopes to create an enduring
franchise, even in established and large markets.
We have a window of opportunity as larger players marshal the
resources to pursue the online opportunity and as customers,
new to purchasing online, are receptive to forming new
relationships. The competitive landscape has continued to
evolve at a fast pace. Many large players have moved online
with credible offerings and have devoted substantial energy and
resources to building awareness, traffic, and sales. Our goal is

to move quickly to solidify and extend our current position
while we begin to pursue the online commerce opportunities in
other areas. We see substantial opportunity in the large markets
we are targeting. This strategy is not without risk: it requires
serious investment and crisp execution against established
franchise leaders.
It’s All About the Long Term
We believe that a fundamental measure of our success will be
the shareholder value we create over the long term. This value
will be a direct result of our ability to extend and solidify our
current market leadership position. The stronger our market
leadership, the more powerful our economic model. Market
leadership can translate directly to higher revenue, higher
profitability, greater capital velocity, and correspondingly
stronger returns on invested capital.
Our decisions have consistently reflected this focus. We first
measure ourselves in terms of the metrics most indicative of our
market leadership: customer and revenue growth, the degree to
which our customers continue to purchase from us on a repeat
basis, and the strength of our brand. We have invested and
will continue to invest aggressively to expand and leverage our
customer base, brand, and infrastructure as we move to
establish an enduring franchise.
Because of our emphasis on the long term, we may make
decisions and weigh tradeoffs differently than some companies.
Accordingly, we want to share with you our fundamental
management and decision-making approach so that you, our
shareholders, may confirm that it is consistent with your
investment philosophy:
· We will continue to focus relentlessly on our customers.
· We will continue to make investment decisions in light of
long-term market leadership considerations rather than short-
term profitability considerations or short-term Wall Street
reactions.

17
(
2021.
· We will continue to measure our programs and the
effectiveness of our investme nts analytically, to jettison
those that do not provide acceptable returns, and to step up our
investment in those that work best. We will continue to learn
from both our successes and our failures.
· We will make bold rather than timid investment decisions
where we see a sufficient probability of gaining market
leadership advantages. Some of these investments will pay off,
others will not, and we will have learned another valuable
lesson in either case.
· When forced to choose between optimizing the appearance of
our GAAP accounting and maximizing the present value of
future cash flows, we’ll take the cash flows.
· We will share our strategic thought processes with you when
we make bold choices (to the extent competitive pressures
allow), so that you may evaluate for yourselves whether we are
making rational long-term leadership investments.
· We will work hard to spend wisely and maintain our lean
culture. We understand the importance of continually
reinforcing a cost-conscious culture, particularly in a business
incurring net losses.
· We will balance our focus on growth with emphasis on long-

term profitability and capital management. At this stage, we
choose to prioritize growth because we believe that scale is
central to achieving the potential of our business model.
· We will continue to focus on hiring and retaining versatile and
talented employees, and continue to weight their compensation
to stock options rather than cash. We know our success will be
largely affected by our ability to attract and retain a motivated
employee base, each of whom must think like, and therefore
must actually be, an owner.
We aren’t so bold as to claim that the above is the “right”
investment philosophy, but it’s ours, and we would be remiss if
we weren’t clear in the approach we have taken and will
continue to take.
With this foundation, we would like to turn to a review of our
business focus, our progress in 1997, and our outlook for the
future.
Obsess Over Customers
From the beginning, our focus has been on offering our
customers compelling value. We realized that the Web was, and
still is, the World Wide Wait. Therefore, we set out to offer
customers something they simply could not get any other way,
and began serving them with books. We brought them much
more selection than was possible in a physical store (our store
would now occupy 6 football fields), and presented it in a
useful, easy-to-search, and easy-to-browse format in a store
open
365 days a year, 24 hours a day. We maintained a dogged
focus on improving the shopping experience, and in 1997
substantially enhanced our store. We now offer customers gift
certificates, 1- ClickSM shopping, and vastly more reviews,
content, browsing options, and recommendation features. We
dramatically lowered prices, further increasing customer value.
Word of mouth remains the most powerful customer
acquisition tool we have, and we are grateful for the trust our
customers have placed in us. Repeat purchases and word of
mouth have combined to make Amazon.com the market leader

in online bookselling.
18
(
2021.
By many measures, Amazon.com came a long way in 1997:
· Sales grew from $15.7 million in 1996 to $147.8 million – an
838% increase.
· Cumulative customer accounts grew from 180,000 to
1,510,000 – a 738% increase.
· The percentage of orders from repeat customers grew from
over 46% in the fourth quarter of 1996 to over 58% in the same
period in 1997.
· In terms of audience reach, per Media Metrix, our website
went from a rank of 90th to within the top 20.
· We established long-term relationships with many important
strategic partners, including America Online, Yahoo!, Excite,
Netscape, GeoCities, AltaVista, @Home, and Prodigy.
Infrastructure
During 1997, we worked hard to expand our business
infrastructure to support these greatly increased traffic, sales,
and service levels:
· Amazon.com’s employee base grew from 158 to 614, and we
significantly strengthened our management team.

· Distribution center capacity grew from 50,000 to 285,000
square feet, including a 70% expansion of our Seattle facilities
and the launch of our second distribution center in Delaware
in November.
· Inventories rose to over 200,000 titles at year-end, enabling us
to improve availability for our customers.
· Our cash and investment balances at year-end were $125
million, thanks to our initial public offering in May 1997 and
our $75 million loan, affording us substantial strategic
flexibility.
Our Employees
The past year’s success is the product of a talented, smart, hard-
working group, and I take great pride in being a part of this
team. Setting the bar high in our approach to hiring has been,
and will continue to be, the single most important element of
Amazon.com’s success.
It’s not easy to work here (when I interview people I tell them,
“You can work long, hard, or smart, but at Amazon.com you
can’t choose two out of three”), but we are working to build
something important, something that matters to our customers,
something that we can all tell our grandchildren about. Such
things aren’t meant to be easy. We are incredibly fortunate to
have this group of dedicated employees whose sacrifices and
passion build Amazon.com.
Goals for 1998
We are still in the early stages of learning how to bring new
value to our customers through Internet commerce and
merchandising. Our goal remains to continue to solidify and
extend our brand and customer base. This requires sustained
investment in systems and infrastructure to
19

(
2021.
support outstanding customer convenience, selection, and
service while we grow. We are planning to add music to our
product offering, and over time we believe that other products
may be prudent investments. We also believe there are
significant opportunities to better serve our customers
overseas, such as reducing delivery times and better tailoring
the customer experience. To be certain, a big part of the
challenge for us will lie not in finding new ways to expand our
business, but in prioritizing our investments.
We now know vastly more about online commerce than when
Amazon.com was founded, but we still have so much to learn.
Though we are optimistic, we must remain vigilant and maintain
a sense of urgency. The challenges and hurdles we will face to
make our long-term vision for Amazon.com a reality are
several: aggressive, capable, well-funded competition;
considerable growth challenges and execution risk; the risks of
product and geographic expansion; and the need for large
continuing investments to meet an expanding market
opportunity. However, as we’ve long said, online bookselling,
and online commerce in general, should prove to be a very large
market, and it’s likely that a number of companies will see
significant benefit. We feel good about what we’ve done, and
even more excited about what we want to do.
1997 was indeed an incredible year. We at Amazon.com are
grateful to our customers for their business and trust, to each
other for our hard work, and to our shareholders for their
support and encouragement.

Jeffrey P. Bezos
Founder and Chief Executive Officer Amazon.com, Inc.
Source: Amazon.com.
20
(
202
)1.
Amazon Web Services609-048
Exhibit 3AideRSS Savings from Amazon Web Services

Startup math
Dedicated Route
Development:
Testing cluster (on-demand)
+ 5 CPUs / 5 months
+ 250mb/s uplink (Scary!)
+ Own cloud service
(5 servers x $100) x 5 Bandwidth + Cloud
------------------------------
at least $2,500 / month
Production:
Dynamic cluster for updates
+ 20-100 CPUs
+ 250mb/s uplink (Free!)
+ SQS-alike for messaging
(20-100) x $150 / month Bandwidth + Messaging
greater than $15,000 / month
Amazon AWS Platform
Development:
Testing cluster (on-demand)
+ 5 CPUs / 5 months
+ S3 for messaging
$180 total
(1,400% savings!)
Production:
Dynamic cluster for updates
+ 20-100 CPUs

+ SQS for messaging
-------------------------------
less than $5,000/ month
(> 300% savings / month!)
Source: Company.
21

(
2021.
Exhibit 4Pricing Structures for Selected SSPs
Online Storage
Solution
Pricing
Box.net
Free – 1 GB storage space/10 GB of bandwidth/10MB file size
limit
$80/year – 5 GB storage space/unlimited bandwidth/1 GB file
size limit
$199/year – 15 GB storage space/unlimited bandwidth/1 GB file
size

Mozy
Free – 2 GB storage space
$55/year – Unlimited space
Omnidrive
Free – 1 GB storage space/5 GB of bandwidth
$40/year – 10 GB storage space/20 GB of bandwidth
Xdrive (AOL)
Free – 5 GB storage space
$120/year – 50 GB storage space
Windows Live Folders
Free – 500 MB storage space/50 MB file size limit
Source: Adapted by casewriter from Harrison Hoffman, “Six
Places to Store your Files Online,” CNET News.com, June
29, 2007, http://news.cnet.com/8301-13515_3-9736064-26.html,
accessed July 2008.

22
(

2021.
Exhibit 5AWS Service Health Dashboard
Source: AWS website.
23
(
2021.
Exhibit 6AWS Bandwidth

Source: Amazon.com presentation, “Annual Meeting of
Shareholders,” May 29, 2008, p.35.
24

(
This
document
is
authorized
for
use
only
by
Esther
Babalola
in
BUSO732

-
Summer
2021
taught
by
JOSEPH
P.
BAILEY,
University
of
Maryland
from

Jul
)
(
2021.
Exhibit 7Aminoto’s EC2 Instance Usage
Source: Amazon.com presentation, “Annual Meeting of
Shareholders,” May 29, 2008, p.38.

(For the exclusive use of E. Babalola, 2021.9 -

Recommended

Recommended

More Related Content

Similar to (For the exclusive use of E. Babalola, 2021.9 -

Similar to (For the exclusive use of E. Babalola, 2021.9 - (20)

More from VannaJoy20

More from VannaJoy20 (20)

Recently uploaded

Recently uploaded (20)

(For the exclusive use of E. Babalola, 2021.9 -