Unformatted Attachment Preview
Amazon's Big Data Strategy
Adapa Srinivasa Rao
Debapratim Purkayastha
IBS Hyderabad
IBS Hyderabad
"As Amazon's recommendation team, we asked ourselves, how do we take this data and make a little bit more
money, how do we apply it in this channel differently and it was really neat. Now it is being called 'big data' in the
marketing and media world, but at the time we were doing this stuff, it was just kind of putting one foot in front of
the other." 1
—David SeBinger, Former Software Manager (Customer Behavior Research), Amazon.com , Inc. (Amazon), in October 2013.
•
•
eading e-commerce company Amazon corn, Inc.
(Amazon) and its subsidiary Zappos were ranked
among the top ten retailers in the National Retail
Federation Foundation/American Express Customers' Choice Awards a for two years (2010 and 2011)
in a row. Industry observers felt that the coveted recognition was the result of Amazon's use of its big
datab resources to provide superior service quality.
Right from the time it had emerged as a dominant
provider of Internet services in the early 2000s, Amazon had started to focus on big data to improve its
performance. Along with many other major Internet
companies, it realized the importance of big data in
the early 2000s, and had since then, focused on properly utilizing the huge databases of people who were
shopping on its e-commerce portals.
Amazon leveraged its big data sources to give
its customers good product recommendations and
L
thereby improve the relationship with them. It utilized its big data resources to meticulously upgrade
its famed customer recommendation system. Data
on past purchases made by customers was used to
give them highly customized product suggestions.
Analysis of past customer data also helped Amazon
in giving suggestions to new customers who were
buying from its portal for the first time. Big data
helped Amazon in developing 360 degree customer
profiles and to create hyper-personalized marketing messages regarding the products based on the
needs and preferences of individual customers. On
the customer side, Amazon also utilized its big data
resources to improve the quality of its customer care.
Easy access to the profiles of customers and their
past purchasing/browsing habits made it easy for the
company's customer service executives to provide
quick solutions to the complaints of customers. The
a The National Retail Federation Foundation/American Express Customers' Choice Awards were designed to know the consumer attitudes
toward retailers' customer service. The survey collects data by polling consumers and is conducted by the marketing intelligence firm
BlGinsight.
b Big data refers to the growth and availability of large volumes of data,
both structured and unstructured. Such an exponential volume of data
could not be analyzed by the traditional software used to handle databases. The latest trends in technology allowed decision making to be
done largely based on data and analysis instead of past experience and
intuition. According to a definition given by industry analyst, Doug
Laney, big data spans three key dimensions, viz. Volume (amount of
data generated), Velocity (speed at which data is streamed), and Variety
(formats in which data comes in).
This case was written by
Adapa Srinivasa Rao,
under the direction of
Debapratim Purkayastha,
IBS Hyderabad. It was
NYDER ABAD
compiled from published
sources, and is intended
to be used as a basis for class discussion rather than to illustrate either
effective or ineffective handling of a management situation.
© 2014, IBS Center for Management Research All rights reserved.
To order copies, call +91 9640901313 or write to IBS Center for
Management Research (ICMR), IFHE Campus, Donthanapally,
Sankarapally Road, Hyderabad 501 203, Andhra Pradesh, India or email:
I S
info@icmrindia.org , www.icmrindia.org
C-404
PART 2
Cases in Crafting and Executing Strategy
acquisition of Zapposc by Amazon in the year 2009
further facilitated the use of big data in improving
customer service quality. Big data resources were also
put to some innovative uses like checking fraud at the
organizational level. Product catalogue data was analyzed thoroughly to identify which of the items were
more likely to be stolen. The results of this analysis
were fed back to the warehouses of the company to
limit the theft of items.
Other than improving its own performance,
Amazon also helped other smaller e-commerce companies by allowing them to use its big data resources
and improve their performance. An innovative service called Amazon Webstore, launched in 2010,
allowed smaller companies to build their portals
around Amazon's e-commerce platform. Users of
Amazon Webstore could advertise their products on
Amazon's portals by paying a small part of the sales
proceeds as a commission to Amazon For a fixed
monthly fee for utilizing the service, partnering
businesses could use Amazon's big data resources.
Amazon Webstore was quite successful and was
adopted by both small as well as big retailers such as
Timex and Samsonite who did not want to have their
own e-commerce system. Amazon's suite of cloud
based Internet services known as Amazon Web
Services (AWS) had also come out with solutions
for small companies so that they could implement
big data easily. A new service known as Kinesis,
announced in November 2013, could process high
volumes of data flowing into AWS on a real time
basis. According to some industry observers, this
was Amazon's bid to close the loop on its integrated
cloud stack and deliver an end-to-end solution for
collecting and processing data. They felt that just by
taking a relook at the various aspects of its big data
capabilities and effectively leveraging on these, the
company could emerge as a threat to the entire analytics eco-system. 2
BACKGROUND NOTE
Amazon was founded in the year 1994 by Jeffrey
Preston Bezos (Bezos). It started its operations at a
time when the reach of the Internet was increasing
and the Internet was being considered as a potential
business medium. Understanding the trend, Bezos
came up with the idea of selling books through the
Internet. He felt that books were the best products
to sell online as millions of titles were in print and
an e-commerce site could house and sell many more
books than the conventional brick-and-mortar bookstores. Bezos calculated that the common brick-andmortar stores could not house more than 200,000
books at a time3 and aimed to build a large online
bookstore which would be bigger than any physical
bookstore in the world. Amazon was initially funded
with the money that Bezos borrowed from friends
and relatives. Bezos and his wife, along with some
employees, built the website and tested it for over a
year before launching it (Refer to Exhibit 1 for Timeline of Amazon).
Amazon was finally opened to customers in the
year 1995. Like many other technology giants, it
was initially run from a garage—the one in Bezos'
Washington home. At the time when Amazon started
its operations, the book retailing market was highly
fragmented and there was no major player except
Barnes & Noble, Inc. d Barnes & Noble had onetenth of the total market share but no online presence. Amazon thus got the first mover advantage
and faced very little competition in its initial days
of operation. Right from when it began its business
operation, Bezos focused on customers and believed
that customer loyalty was the key to penetrating the
market and increasing sales Amazon started to ship
goods to all the 50 states in the US and 45 other
countries within a month of its launch—and all this
while still working from Bezos' garage. Amazon's
popularity grew through word-of-mouth as customers recommended it to others. Within four months
of its launch, Amazon was selling more than 100
books a day. The company's impressive performance attracted investors and Amazon got its first
big investment of US$ 100,000 from Madrona Venture Group, Inc.e in 1995. The company reported net
sales of US$ 511,000 during the first six months of
its operations and Bezos' confidence that he could
make a success of the company increased.
"
Zappos, headquartered in Las Vegas, Nevada, USA was a leading
online retailer of shoes and clothing products. This online shopping
portal was founded by Nick Swinmum in the year 1999.
•
Barnes & Noble, Inc., headquartered in Manhattan, New York City,
USA, is the largest book retailers in the United States.
Madrona Venture Group, Inc., headquartered in Seattle, Washington,
USA, is a venture capital which primarily focuses on investing in earlystage technology companies.
•
•
•
CASE 28
EXHIBIT 1
C-405
Timeline of Amazon.com
Year
Month
Event
1994
July
Amazon incorporated in Delaware
1995
1996
1997
July
July
May
1998
June
March
Amazon.com launched. Sells its first book, "Fluid Concepts & Creative Analogies: Computer
Models of the Fundamental Mechanisms of Thought"
Amazon.com Associates Program launched
Announces IPO and begins trading on NASDAQ
Launches music store
1999
Launches Amazon.com Auctions, the company's Web auctions service
December
Jeff Bezos named Time Magazine "Person of the Year"
Launches Amazon.fr (France)
2001
August
November
April
2002
July
Launches Amazon Web Services
November
Opens Apparel & Accessories Store
2000
•
Amazon's Big Data Strategy
Launches Amazon.co.jp (Japan)
Amazon partners with Borders Group to run the company's online bookselling business
2003
2004
June
Launches Amazon Services, Inc. subsidiary
April
2005
2006
2007
February
September
November
Opens Jewelery Store
Introduces Amazon Prime
Launches digital video download service, Amazon UnboxTM
Launches Amazon Kindle
Announces Frustration-Free Packaging initiative
2008
November
2009
May
Introduces Kindle DX
2010
2011
2012
2013
April
July
February
August
Amazon moves to new HQ in South Lake Union, Seattle
Market capitalization of Amazon tops US$ 100 billion
Amazon launches Sports Collectibles Store
Jeff Bezos buys Washington Post
Source: Compiled from various sources.
•
At the beginning of 1996, Amazon moved to
new headquarters—a small warehouse in Seattle. The
company employed 11 people and offered 2.5 million
book titles. Following the Japanese model, Amazon
had very limited inventory and thereby kept its costs
under control. It started an innovative affiliate marketing program called Amazon Associates Program
in July 1996. The program allowed third party websites to sell books through links to Amazon posted
on their sites for a commission of 15 percent on the
total sales made. The program was a huge success
and helped in expanding Amazon's reach without
the company having to spend much on advertising.
Experts opined that the program not only generated
traffic to Amazon but enhanced the brand's presence online as these third party sites carried Amazon's logo on their pages. The Amazon Associates
Program was later extended to all the products sold
on Amazon's portal. Amazon went public in the year
1997 and offered 3 million of its shares for sale. The
shares opened at US$ 18 a share and the IPO raised
US$ 54 million for the company. In the year 1998,
Amazon started selling DVDs with the opening of
its video store which was followed by the launch of
Amazon.com auctions in March 1999. 4
Amazon's success attracted many new competitors like Book Stacks and Book Zone to the market
and this led to higher competition for the company.
To counter the competition effectively, Amazon
introduced new features like online product reviews
where customers could write their own book review
as well as read reviews written by others. By the
year 2000, Amazon had made a big change in its
business model and started selling other products.
In 2000, it also expanded its presence and launched
sites in France and Japan. In the year 2001, Amazon
C-406
EXHIBIT 2
PART 2
•
Cases in Crafting and Executing Strategy
Financials of Amazon from 2008-2012 (in US$ Millions)
Total net sales
$61,093
$48,077
$34,204
60,417
47,215
676
40
862
61
32,798
1,406
51
(92)
(65)
(39)
(80)
(132)
76
79
29
72
91
32
59
1,161
901
(247)
Total operating expenses
Income from operations
Interest income
Interest expense
Other income (expense), net
Total non-operating income (expense)
544
Income before income taxes
Provision for income taxes
Equity-method investment activity, net of tax
Net income (loss)
934
(291)
(12)
(428)
(155)
$
(39)
$
631
1,497
(352)
7
$ 1,152
$24,509
$19,166
18,324
842
23,380
1,129
37
(34)
83
(71)
47
(253)
(6)
$
902
$
(9)
645
Source:
allowed other retailers to sell their products through
its site and took a part of the sales proceeds as its
commission. In the last quarter of the year 2001,
Amazon reported its first profit. 5
Amazon's operations were further expanded in
the year 2003 as it opened new websites in AsiaPacific and European countries. In the year 2006, it
launched a key subsidiary called Amazon Web Services.6 Amazon Web Services provided an array of
cloud based remote computing services to its customers. The advent of the digital era was changing
the content consumption patterns of people. Many
people started reading books and magazines on their
desktops and laptops instead of buying physical copies. Responding to this change, Amazon introduced
an e-book reader called Kindle in the year 2007.
Kindle was a big hit in the market and heralded a
new era of digital reading. Kindle was later released
as an app for other devices working on operating
systems like Android and i0S f. By the year 2011,
the market capitalization of Amazon had reached
the US$ 100 billion mark, making it one of the leading technology companies in the world. For the fiscal year 2012, Amazon had revenue of US$ 61.09
billion and a net loss of US$ 39 billion (Refer to
Exhibit 2 for the financials of Amazon).
(
Android and iOS are the two leading mobile OS promoted by Google
and Apple respectively. They are used in mobile devices like smartphones and tablets.
BIG DATA AT AMAZON
Over the years, Amazon had evolved from being a
pure e-commerce player into a giant Internet services firm which offered a large range of services
for individuals and corporations. It started to focus
heavily on big data and embarked on its transition
from a pure online retailer into a giant big data
company.7 Amazon along with other major Internet
giants like Yahoo! Inc.g (Yahoo) and Twitter, Inc. h
•
(Twiter)alzdnhy20staed
huge amounts of data about their users which they
could put to valuable use. 8 While the other companies did not concentrate on the importance of big
data, Amazon was quick to cash in on the invaluable
database of people who shopped on its e-commerce
portals around the world. The product recommendation team at Amazon thought of innovative ways in
which it could use the data accumulated by the company.9 The result was the big data revolution which
transformed the way Amazon did business.
As an e-commerce giant, Amazon's success had
always depended on making the right products available to its customers. Making the right products available in turn depended on understanding the precise
products that customers wanted. Understanding the
g Yahoo, Inc., headquartered in Sunnyvale, California, USA, is a leading
multinational Internet company.
h Twitter, Inc., headquartered in San Francisco, California, USA, is a
leading social networking and microblogging service. It allows its users
to send and receive text messages which are limited to 140 characters.
•
•
CASE 28
needs and tastes of customers involved doing proper
market research as well as analyzing its own customer
base. Since its inception, Amazon had been renowned
for its product recommender system which provided
product suggestions to customers depending upon
their past purchasing behavior. Data collected from
its customers was the primary driving force behind
Amazon's recommender system. Being the leading
e-commerce player, Amazon had a large bank of data
regarding the likes and the past purchasing behavior of its customer base. It had used this data bank
to build its recommender system. Its earlier recommender system had been based on showing more
items similar to the ones which were being looked for
by its customers. This item-by-item similarity method
was built on the basis of collaborative filtering' and
was hugely successful in deepening the relationship
with its customers. Its recommender engine had since
been improved and perfected to give better results.
Amazon later started utilizing the historical
purchase data of consumers as well and the clickstream data of all its customers to show webpages
with uniquely customized information. 10 Using such
data helped Amazon in many ways other than showing the related and alternative products that the consumers had been looking for. Mining the vast amount
of data helped in understanding the inner feelings
and likings of customers which they could not
express themselves. Commenting on the importance
of data in understanding the behavior of customers,
Michael Driscoll of Datasporai said, "You can ask
people what influences their desire to renew their
cell phone contract, and what people say and what
they do are often very different. Data is the key to
differentiating between what people say in terms of
sentiment and what they do in terms of actions." 11
UTILIZING BIG DATA
Amazon leveraged on big data to improve its relationship with its customers and provide superior
customer service. The online retailer built a vast
database of its customers and their buying preferences over a long period of time. It was one of
•
The Collaborative filtering system is a technique used in many recommender systems of e-commerce portals. Collaborative filtering involves
filtering information or usage patterns through techniques involving
multiple data bases, viewpoints, and agents.
Dataspora, headquartered in Cambridge, Massachusetts, US is a leading
big data and analytics consultancy.
Amazon's Big Data Strategy
C-407
the first e-commerce companies to start using the
cross-selling/up-selling method. This customer recommendation system was later augmented by utilizing its big data resources. Using big data, Amazon
started analyzing the past product purchases made
from its online store by its customers and the other
items that were purchased along with them. Data
collected from its customers was used to give silent
but highly customized suggestions to make them
buy more. This analysis of data helped Amazon
give product suggestions to existing as well as new
customers who might not have otherwise bought a
complementary product. This was the reason why
Amazon's sites displayed 'Other customers who
bought this item also purchased that item' kind of
cross-selling recommendations. 12 Product related
recommendations were also customized based on
many factors such as the customer's location and
demography. "It can even cross-correlate buying
behavior between home and garden sales," 13 said
Jeff Kelly, lead big data analyst at Wikibon k.
The bewildering range of products that were
showcased on e-commerce portals made them seem
unwieldy and incomprehensible to many customers. According to analysts, mining the treasure trove
of information and providing relevant product recommendations could make e-commerce sites feel
smaller and more intimate to the consumers. 14 Big
data also helped Amazon in the development of its
personalized marketing strategy—a tactic in which
it excelled. Many e-commerce firms resorted to
generic mass emailing of the products and offers
that were available with them. This strategy led to
the wastage of marketing efforts of many firms and
the labeling of e-commerce mails as spam Amazon
created 360 degree customer profiles which tracked
and stored everything related to customers like their
browsing history, social data, tastes and preferences, past purchase history, etc. These 360 degree
customer profiles facilitated the identification of
discerning groups of customers who could be well
targeted. 15 Amazon could create hyper-personalized
marketing messages regarding the products based on
the individual customer's needs and interests. 16
Amazon also relied on big data to improve the
quality of its after sales service to its customers.
k
Wikibon, headquartered in Marlborough, Massachusetts, USA, is a
community of practitioners and consultants on technology and business
systems that use open source sharing of free advisory knowledge.
C - 408
PART 2 Cases in Crafting and Executing Strategy
Most American customers were known to have a
largely negative experience in their service interactions. 17 Amazon tried to solve this problem by leveraging on the large data it had regarding its customers.
Having the right data helped it have a favorable
exchange with its customers and to solve their problems quickly. Amazon's customer service executives
had speedy access to data regarding the past purchases and browsing history of its customers. This
enabled the company to provide quicker solutions to
the problems and complaints of its customers. Complainants did not have to spell out their details like
last names, contact numbers, and addresses repeatedly before their queries/problems were solved. This
unique approach of utilizing big data to improve service quality gave spectacular results. After having
a positive experience with Amazon's support team
in one such encounter, Sean Madden, a top business
blogger, said, "After nearly a decade of ordering
stuff from Amazon, I never loved the company as
much as I did at that moment." I8
One of the factors which facilitated the use of
big data for customer service was Amazon's acquisition of Zappos, the largest online retailer of shoes,
in the year 2009. Amazon acquired Zappos for US$
1.2 billion to expand its reach in product categories
in which it was not strong. 19 Zappos was famed for
using its customer database to provide a personal
touch to its customers and turn them into its fans and
cheerleaders. 2° Amazon adopted the customer service
strategies of Zappos after it took over the company.
The application of big data for improving customer
EXHIBIT 3
service made Amazon and its subsidiary Zappos to
rank among the top 10 retailers in National Retail
Federation Foundation/American Express Customers' Choice Awards for the years 2010 and 2011
•
(Refer to Exhibit 3 for top 10 retailers in National
Retail Federation Foundation/American Express
Customers' Choice Awards for 2010 and 2011). 21
Rather than using big data to just provide better
product suggestions and improve the quality of service,
Amazon used it to check fraud in the organization. An
interesting area where Amazon benefited through using
big data was in preventing warehouse theft. At any
given point of time, Amazon had 1.5 billion items in its
catalogues across its 200 fulfillment centers across the
world. Theft of these items was a big threat to Amazon. The problem with identifying which of these
items were more sought after by thieves was that both
expensive and low-priced items were stolen. Inexpensive items too were often stolen due to reasons like
their scarcity. To solve this problem, Amazon used big
data and updated its product catalogue data nearly 50
million times a week. 22 Product catalogue data was
collected, stored, and analyzed to identify which of
the items were more likely to be stolen and the information was fed back to the warehouses (Refer to
Exhibit 4 for the five components of big data process).
This helped Amazon in preventing the theft of items
in its catalogues. Werner Vogels (Vogels), Chief Technology Officer and Vice President of Amazon.com ,
felt that data and storage should be unconstrained.
"In the old world of data analysis you knew exactly
which questions you wanted to ask, which drove a
Top 10 Retailers in National Retail Federation Foundation/American
Express Customers' Choice Awards for 2010 and 2011
1
Amazon.com
Zappos
2
L.L. Bean
Amazon.com
3
Zappos
L.L. Bean
4
Overstock.com
Overstock.com
5
QVC
Lands' End
6
Kohl's Department Stores
JCPenney
7
Lands' End
Kohl's
8
JCPenney
QVC
9
Newegg
Nordstorm
Nordstorm
Newegg
10
Source: "Customers' Choice Awards."
•
•
CASE 28
Collect
Collecting and getting the data to the place where the process can be started.
Store
Storing the collected data before it is put to proper use.
Organize
Controlling the quality of data by knowing which data to include in the stream. Organizing also
involves validating data in order to make sure that correct data is used.
Analytics
Analysis of well-organized data to create usable information.
Share
Information that is created through analytics is shared with those who need it.
very predictable collection and storage model. In the
new world of data analysis your questions are going
to evolve and change over time and as such you need
to be able to collect, store, and analyze data without
being constrained by resources." 23
•
C-409
EXHIBIT 4 Five Components of the Big Data Process
Source: Eric Savitz, "CeBIT: Amazon CTO Werner Vogels Talks Big Data,"
•
Amazon's Big Data Strategy
AIDING OTHER COMPANIES
WITH BIG DATA
Other than utilizing big data for improving its own
performance, Amazon also helped other e-commerce
portals to leverage its big data resources. This it
achieved through an innovative service for smaller
e-commerce businesses called Amazon Webstore.
Amazon Webstore, launched in 2010, allowed retailers
to build their portals around Amazon's e-commerce
platform.24 Amazon Webstore was an independent
store outside of Amazon's official e-commerce store
built on the third party domain name and brand. Users
of Amazon Webstore could place Amazon Product
Ads' on Amazon's portals. People who clicked on
these product ads would be redirected to the site of the
partnering site where the sale would be completed. As
part of this partnership, Amazon allowed its partnering sites to use its big data while maintaining their
independent identity as a small online e-commerce
store.25 Amazon charged its merchant partners a fixed
monthly fee as well as a fixed commission for using
its resources and big data resources. 26
An interesting case where Amazon Webstore
improved the performance of a small retailer was
Anaconda Sports. Anaconda Sports, a successful
Amazon Product Ads was an advertising program that allowed sellers
to promote their products on the official e-commerce portals of Amazon in different countries.
, August 3, 2012.
sports retailer from New York, USA, found itself
stuck with an inefficient and expensive e-commerce
system with issues like inability to store all the customer information, lack of unique experience based
on customer preferences, and poor customer service
quality.27 Modifying its e-commerce portal through
Amazon Webstore made it possible for it to develop
an efficient store which solved all the problems it had
been facing as well as increase its sales substantially.
Commenting on the benefits of Amazon's Webstore for small and medium businesses, Scott Pulsipher, director of Amazon Webstore, said, "By
leveraging Amazon's technology and infrastructure,
Amazon Webstore levels the playing field for smalland medium-sized businesses, helping them quickly
and easily build their businesses and improve the customer experience."28 Amazon Webstore was implemented even by big brands like Timex, MTV, Boeing,
and Samsonite which allowed them to improve their
engagement with their customers (Refer to Exhibit 5
for the screenshot of Samsonite's website built using
Amazon Webstore). Commenting on how Amazon
Webstore helped to increase sales and cut costs at
Timexm, its e-commerce director Cal Crouch said,
"When we launched our new Amazon Webstore, we
saw an immediate lift of 40 percent in revenue and
average order size. And on the support side, we have
gained the flexibility to make most changes to content as well as brand ourselves—saving us thousands
(of dollars) in development costs." 29
Amazon's Amazon Web Services (AWS) helped
a lot of companies to develop better applications,
deploy new products and services, and cut their costs
m Timex, headquartered in Hoofdorp, Netherlands, is a maker of timepieces and luxury goods.
PART 2 Cases in Crafting and Executing Strategy
C-410
EXHIBIT 5 Screenshot of Samsonite's Website Built Using Amazon Webstore
•
Sams nite
SELECT PRODUCTS BUY 1 GET 10% OFF, BUT 2 GET 15% OFF, RUT 3 GET 20% OFF • FREE STANOARO SHIPPING ON ALL LIGHTWEIGHT SPINNERS
Online only. Ends 1/2. use Promo Code: NEWYEARS Offer Derails.,
LUGGAGE
BUSINESS & LAPTOP
BACKPACK & MESSENGER
DUFFLE & SPORT
ACCESSORIES
COLLECTIONS
Lightweight Spinners,
Four Wheel Freedom.
SHOP LIGHTWEIGHT SPINNERS
HIGHLY RATED
TOP SELLERS
Featured
LATEST RELEASES
Prod
r•••It
III
Samsonite
Samsonite EZ Cart
Samsonite Lift
HypErSpaCe Spinner
2S"
Wheeled Boarding
BPS
Boarding Bag
SamsoniteVizAlr
Laptop Backpack
•
Samsonite Carbonl
DUX 20' Spinner
Pink
Source: http://webstore.amazon.com/client-showcase/b/6254207011.
(Refer to Exhibit 6 for AWS architecture). Amazon offered its solution using familiar tools such as
Oracle Database and Microsoft SQL Server, while
also pioneering and promoting new platforms such
as DynamoDBt1, Hadoop°, and RedshiftP. 3° "One
of the core concepts of Big Data is being able to
evolve analytics over time. For that, a company cannot be constrained by any resource. As such, Cloud
Computing and Big Data are closely linked because
for a company to be able to collect, store, organize,
" DynamoDB is a managed NoSQL database service which makes it
simple and cheap to store and retrieve large amounts of data.
° Hadoop is an open-source software framework for storing and processing large data-sets.
and powerful data warehouse servicc which is a Pare
P Redshift is a fast
of AWS.
analyze, and share data, they need access to infinite
resources,"31 said Vogels.
Small companies faced a lot of difficulties in
adopting and deploying big data due to the limited
resources at their disposal Amazon came out with
solutions for such companies so that they could
implement big data easily. In November 2013, Amazon Web Service announced a new service for real
time processing of big data. The service known as
Kinesis, processed the high volumes of data flowing into Amazon's web-based storehouses on a real
time basis. The tool had the capability to accept any
number of data sources and could process terabytes
of data per hour. It was intended to allow developera to create
acycateatierts that -a/el -Ste...I ea a
bumir. as
basis for tasks like website traffic arnalysis,
110
•
CASE 28 Amazon's Big Data Strategy
C-411
EXHIBIT 6 AWS Infrastructure
AWS Simple
Storage Service
Logs
Users
Elastic Compute
Cloud Instance
Elastic Compute
Cloud Instance
RESTful API
Web Server
Storage
(Images,
Video, etc.)
Job Worke A
Backup
Source: tp://d36cz9buwruitt.cloudfront.net/pixnet_diagram_2.jpg.
transactions related to marketing and finance, social
media data, and logs (Refer to Exhibit 7 for Kinesis'
architecture). Commenting on the flexibility Kinesis would bring to businesses, an analyst at Neoviseq
Paul Burns said, "Sometimes people spend hours or
days just collecting the data, then coming back and
processing it, so it's out of date. . . . So Amazon said
we'll take care of all that for you, just write your
own program and connect to us." 32 The ability to
create big data apps through Kinesis was expected
to remove one of the biggest bottlenecks for smaller
companies in adopting big data for their businesses.
But one limitation of Kinesis was that all the data
processing would be done at the data centers of
Amazon itself instead of at the clients' location.
LOOKING AHEAD
Analysts came up with suggestions on more ways in
which Amazon could benefit from big data. Having
•
q Neovise, headquartered in Fort Collins, Colorado, USA, is an IT industry
analyst firm.
its roots in selling books, Amazon had built a review
system for the books sold through its website.
Amazon's review system was mainly based on text
reviews written by customers and the number of
stars (from one to five) given to a book or author.
This review system allowed Amazon to build a community and a loyal customer base. Over the years,
there were allegations that many authors had found
a way to manipulate Amazon's review system and to
get paid reviews for their books. Such paid reviews
tended to be biased and in turn, they impacted the
reliability of the review system. 33 To solve this problem, some industry experts suggested that Amazon
create a big data solution which would allow readers to give a vast range of additional feedback and
comments which could be used to check the veracity of the reviews. Analysts opined that apart from
improving the reliability of its review system, a big
data based review system would also make it possible for Amazon to show more relevant reviews to
the customers just as it suggested relevant products.
Another suggestion regarding the use of big
data to further Amazon's prospects was in giving
C 412
-
PART 2
•
Cases in Crafting and Executing Strategy
EXHIBIT 7 Kinesis' Architecture
Appl
(Aggregate &
De-Duplicate)
App2
AWS
Endpoint
Availability
—"–
Zone
(Metric
Extraction)
App3
(Sliding Window
Analysis)
App4
(Machine
Learning)
Source:
http://arstechnica.com/information-technology/2013/11/amazon-wades-into-big-data-streams-with-kinesis/.
better competition to other big Internet companies
like Google Inc.` and Facebook, Inc.' The business
models of many of these Internet companies were
based on online advertising. And according to an
estimate by Google, 30 to 40 percent of its revenue
from search advertising came from e-commerce
sources. 34 The growth of Amazon as the Internet's
one-stop shop and its increasing product base made
it the primary destination for product searches,
clearly bypassing Google. This unique position left
Amazon in possession of more shopping data of people than any other Internet company. Some analysts
Google Inc., headquartered in Mountain View, California, USA, is a
leading Internet-related products and services firm.
Facebook, Inc., headquartered in Menlo Park, California, USA, is a
leading social networking service.
The Washington Post, headquartered in Washington, D.C., USA is a
leading American newspaper.
•
were of the view that using big data, Amazon could
beat Google and Facebook in the long term. Bezos
acquired The Washington Post` for US$ 250 million
in August 2013.35 This acquisition sparked speculation among industry observers that Bezos would use
big data to revitalize the news business and find new
revenue sources for the ageing business. The use
of big data analytics could give better insights into
the readers of The Washington Post. Amazon could
thereby integrate the likings of readers in developing new products in the news business. According to
analysts, Amazon's expertise in big data could transform the online news business in the same way as
Bezos had transformed the 500-year-old book publishing business. 36 According to Wikibon's big data
analyst Jeff Kelly, as of end 2013, Amazon had all
the pieces of the big data puzzle but the firm would
have put these together effectively to emerge as a
dominant player in this space. 37
•
•
END NOTES
•
Teresa Novellino, 'At War with Amazon? Rich
Relevance Offers Big Data Weaponry,"
, October 9, 2013.
2 Maria Deutscher, 'Amazon Closes the Loop
on Big Data," , November 22, 2013.
3 "History of AMAZON.COM ,"
1
4
"Timeline History Amazon.com ,"
5
"History of AMAZON.COM ,"
6
"History and Timeline,"
7
"How Amazon is Leveraging Big Data,"
Bill Vorhies, 'A Brief History of Big Data Technologies — From SQL to NoSQLtpHadoop and
Beyond," , October 31,
2013.
9 Teresa Novellino, At War with Amazon?
Rich Relevance Offers Big Data Weaponry,"
, October 9,
2013.
10 "How Amazon is Leveraging Big Data,"
CASE 28
Amazon's Big Data Strategy
Adria Saracino, "Interesting Ways Businesses Use Big Data to Improve Personaliza, April 23, 2013.
tion,"
15 Lisa Desjardins, "How Amazon Uses Marketing Personalization,"
, October 7, 2013.
16 Lisa Desjardins, "How Amazon Uses Marketing Personalization,"
, October 7, 2013.
Sean Madden, "How Companies like Amazon Use Big Data to Make You Love Them,"
, May 2, 2012.
18 Sean Madden, "How Companies like Amazon Use Big Data to Make You Love Them,"
, May 2, 2012.
19 "History and Timeline,"
14
8
" "Q&A: What Can Non-IT Companies Learn
from Amazon and Facebook about How to
Leverage Big Data?"
12 Jodi Beuder, "How Can Big Data Improve
the Customer Experience this Holiday Season?" , November 5,
2013.
13 Constance Gustke, "Retail Goes Shopping
Through Big Data," , April 15,
2013.
•
Justin Amendola, "What Zappos.com can
Teach you About Turning Customers into Mega
Fans,"
May 19, 2012.
21 Frank Reed, 'Amazon Retains Top Spot in
Customer Service Poll, Zappos Third,"
, January 19, 2012.
22 Ryan Lawler, "How Amazon Uses Big Data
to Prevent Warehouse Theft,"
, October 18, 2011.
23 Roberto V. Zicari, "On Big Data: Interview with
Dr. Werner Vogels, CTO and VP of Amazon.
com ,"
, November 2, 2011.
24 Chris Crum, 'Amazon Launches New Webstore E-Commerce Product,"
, May 24, 2010.
25 "Power of Amazon," ,
.„-,ostore.amaaon
20
-your-brand/
26
Amazon Webstore Pricing,"
C-413
ap://webstore.
amazon.com i arnazar.-webstore-pricing/b/
6960 7501•.
Adria Saracino, "Interesting Ways Businesses Use Big Data to Improve Personaliza, April 23, 2013.
tion,"
28 Chris Crum, 'Amazon Launches New
Webstore E-Commerce Product,"
, May 24, 2010.
29 Chris Crum, 'Amazon Launches New Webstore E-Commerce Product,"
, May 24, 2010.
30 Doug Henschen, 'Amazon's Vogels:
Big Data Belongs In The Cloud,"
, April 19, 2013
31 Roberto V. Zicari, "On Big Data: Interview with
Dr. Werner Vogels, CTO and VP of Amazon.
com ,"
, November 2, 2011
32 Andy Patrizio, "Why Amazon's Kinesis Tool is
a Big Deal for Working with Big Data,"
, November 22, 2013.
33 Amir Kurtovic, 'A Better Review: Why Amazon
Should Embrace Big Data to Fix Its Ratings
System,"
34 David Hughes, "Big Data is the Only Way to
Compete with Google,"
July 18, 2013.
35 Amazon Boss Jeff Bezos Buys Washington
Post for $250m,"
, August
6, 2013.
36 Karyl Scott, "How Bezos Could Apply Big Data
and Other Amazon Tactics to the News Busk
, August 14, 2013.
ness:'
37 Maria Deutscher, 'Amazon Closes the Loop
on Big Data," , November 22, 2013.
27