The dawn of Big Data

IBM Smarter Analytics:
Big Data
The dawn
of Big Data
Businesses need to adapt
rapidly if they are to prosper in
a world where massive, high-
velocity data is everywhere.
Smart businesses are already
racing to harness its power
in order to gain competitive
advantage. Do advanced
analytics hold the key to
profiting from the Big Data
explosion? Research reveals
a widening divide,with those
organisations already using
analytics more than twice as
likely to outperform their industry
peers1
. What does your business
need to become one of them?

IBM Smarter Analytics: Big Data
Our ability to generate data has
always outstripped our ability to
organise it. Paper emerged around
100AD but it took another 1,800 years
to come up with the filing cabinet.
Today’s businesses can’t wait that
long. By 2011, it was estimated that
90 per cent of the world’s data in
circulation had been generated
in the preceding two years2
.
How can advanced analytics
unlock the insights buried in the
almost unimaginable amounts of
data now being generated? The
ability to analyse Big Data has
already transformed the world of
science, unravelling the human
genome and identifying the elusive
Higgs boson. Now it is transforming
the way the world does business.

By applying advanced analytics, businesses can tap into their own
Big Data streams to gain deeper insights, enhance the efficiency of
operations and deliver business transformation. It also allows organisations
to differentiate themselves with new services for customers and new ways of
working for the business itself.
This has implications for businesses of all types, especially those already
working in a data-rich environment. For example, mobile telecoms providers are
able to capitalise on the deluge of data from smart phones to provide customers
with new services. Telecoms can access and analyse Big Data – such as
browsing logs, download histories and GPS location data – to create better
products and align services with customer needs.
Big Data also puts the customer centre stage.
What makes Big Data analytics so valuable is its ability to
pinpoint the needle in the haystack – and to do so in real
time. This creates unique opportunities to deliver highly
granular personalised services.
For example, network operators know that dropped
calls and poor reception are drivers for customer churn;
they also know who their high value customers are.
What if they could identify a dropped call, identify the
customer – and then offer that customer a personalised
apology and a compensatory discount the moment
something goes wrong? That would have been virtually
impossible five years ago.
Today, Big Data provides the raw materials needed for businesses
to develop a much deeper understanding of their customers through
brand insight and reputational analytics. Businesses can now gauge customer
sentiment and behaviour easily, and tailor products and services to achieve
deeper market penetration.
In the past, the only way this could be achieved was through expensive market
research and by using basic information provided by customers themselves. In the
case of banks and financial services providers, that includes information such as
when an account was opened and the frequency and nature of transactions.
Traditional structured data still has immense value – Big Data
analytics takes this to the next level. It adds to the equation a whole raft of
semi-structured and unstructured data that had previously been too big and too
fast to handle. Now, virtually every interaction can be considered.
For example, when a customer contacts a call centre, the content of the
conversation can be analysed; if the customer has exchanged information using
the bank’s website, that interaction can also be taken into account. It is also
possible to scan what customers are saying on blogs and social media websites.
This could provide organisations with a 360° view of the customer for the first time.
The insurance sector is also set to benefit from the Big Data revolution.
Traditionally, insurers have invested in conventional data warehouse architecture
to help them price products and provide management information. These
systems provide business value, but they don’t provide the big picture view
needed to deliver real transformation.
For insurers, one of the benefits of Big Data usage could be the ability to
combat fraud more effectively, something that can be achieved by correlating
claims with social media activity and GPS location data. This can provide a far
more accurate view of what really happened in individual incidents
than has previously been possible.
Big Data analytics also makes it possible to deliver
pay-as-you-drive (PAYD) insurance for motorists. PAYD
hinges on the ability to match premiums to actual driving
behaviour, rather than statistical abstractions based on
a driver’s history and risk profile. To deliver PAYD
effectively, insurers need to make sense of massive data
flows from individual vehicles.
Big Data could also meet the needs of the
entertainment industry, making it possible to measure the
likely success of upcoming film releases by analysing social
media postings made by potential cinemagoers. Data about
specific films could be filtered from tens of thousands of Twitter feeds,
providing a clear indication of which films are getting talked about and which
aren’t, to provide an accurate measure of audience sentiment.
Insights gained can be used to reallocate promotional resources based on
needs, with last-minute fine tuning to pre-release campaigns that provide films
with an extra push before launching at the box office. By leveraging Big Data
from social media streams, marketers would be tapping into what are, in
effect, the biggest focus groups on earth.
The ability to dig deeper, faster, has important implications for businesses that
routinely generate huge amounts of data. Big Data analytics allowed one major
multinational car hire business to transform customer operations by redeploying
staff to meet demand surges at peak times. This insight was gained by centralising
the processing of customer satisfaction surveys – something that had previously
been conducted locally – and applying advanced analytics to parse data from
multiple sources to identify patterns and problems that were previously invisible.
The ability to makes sense of Big Data has significant implications for oil & gas
businesses as well. The industry has no shortage of data – an oilfield, for
WHAT CAN BIG DATA DO?
90%+
of Fortune 500 companies
are expected to have
Big Data initiatives
in progress
in 20133

example, can produce around one terabyte of data every day. What is in short
supply are actionable insights.
Deploying advanced analytics can help engineers and other stakeholders
to get their hands on the insights they need, identifying key patterns and
converting data into actionable insights to support better decision making.
With oil & gas companies increasingly exploiting reserves in highly complex
geological environments, the ability to make sense of growing amounts
of data is likely to become a differentiator.
Some of the most exciting potential advances, though, could hinge on
collaborative ventures between different data custodians such as telecoms
providers and retailers. Businesses like these already have an intimate electronic
knowledge of their customers’ habits.
By combining this conventional structured data with additional unstructured
data streams – what a customer is saying on Facebook, Twitter, LinkedIn and
blogs, for example – it is possible to reveal an even clearer picture of that
customer’s tastes. By further combining that with real-time location data, the
prospect of targeting the right offer, at the right time, to the right person,
becomes a real one.
Contextual targeting of this sort is already a possibility and it should become
easier to deliver – and more powerful – as Big Data analytics becomes ubiquitous.
But businesses that go down this route need to make careful judgements about
the reputational impacts of deeper engagement: the line between an unbeatable
offer and unacceptable intrusion is, potentially, a thin one.
UNDERSTANDING BIG DATA
The term “Big Data” first emerged in the public arena during 2008. The word
“big” reflects the fact that the volume of stored and streaming data had
become too big for conventional analytics and business intelligence systems to
handle. Big Data is generally considered to have three characteristics: volume,
variety and velocity. Behind each of these three is a technology that is making
analysis possible.
Volume refers to the scale of data being generated, currently estimated to be
in the region of 2.5 quintillion bytes of data each day4
. However, it should be
noted that individual businesses continue to manage high-volume data
successfully, with complex terabyte-scale models capable of handling billions of
observations. These already exist in conventional data warehousing and
business intelligence, along with software that can handle semi-structured data
and text analytics. The difference today is that data warehouses are evolving,
offering ever-increasing power to tackle volumes that are only going to grow.
Variety is a reflection of the upsurge in heterogeneous data sources, most of
them generating semi-structured or unstructured data – digital gibberish – that has
traditionally proved resistant to easy analysis. These include feeds from sensors
embedded in plant and vehicles, social media streams, audio, video and text.
To make sense of this volume and variety of data, the analytics involved must
be distributed across a large number of machines rather than relying on a single,
centralised database. From IBM’s perspective, Apache™
Hadoop®
leads in this
sphere, an open source software framework that allows for large data sets to be
processed across clusters of servers.
Velocity adds a further layer of complexity. Traditionally, data has been batch
processed and scrubbed clean before being used by businesses. Big Data, by
contrast, is a moving target, with mass flows from a huge array of sources.
Because there is too much to store, insight must be acquired in real time. One of
the things that makes Big Data so useful to businesses is the value that can be
extracted from it while it is still fresh.
Stream computing allows businesses to analyse and act on high-velocity,
high-volume data – such as social media and GPS feeds – while it is in motion.
This represents a fundamental shift away from traditional analytics, which
depends on first accumulating and structuring data. Stream computing is
intended to analyse and identify facts on the fly. This is an essential function of the
Big Data environment because in many cases the sheer volume of raw data
would be too great to store in the first place.
The combination of these three elements – Volume, Variety and Velocity
– creates a unique opportunity for businesses to tap into the true potential of Big
Data and allow organisations to create value in two ways:
1. Continue to develop actionable insight through business analytics but on a
wider and deeper scale. Many consumer-facing organisations already employ
2.5quintillion
(that’s 2.5×1018
) bytes of data have been created
every day, as of 20125

analytics in order to gain insights into their customers. Big Data analytics
approaches could increase the depth and breadth of this capability.
2. Create new opportunities previously unimaginable or not practical due to
technology constraints. Not only should organisations be able to analyse data
previously out of reach, but they will be able to act upon the discovered insights in
near real-time
BIG DATA TECHNOLOGY
IBM’s view is that Big Data should be treated as a platform in its own right. It is not
a single solution, however, but a suite of technologies, concepts and practices.
Big Data does not mean moving away from traditional business intelligence
applications or abandoning recent investment in predictive analytics
applications and it certainly does not suggest shutting down the enterprise
data warehouse. On the contrary: all of these can be plugged in to the Big Data
platform. A number of technologies are shaping the shift to Big Data.
Attempting to manage Big Data without solutions like Hadoop®
is a
potentially risky and expensive strategy, particularly if that data is
heading straight for a conventional data warehouse. This type of
framework helps by allowing organisations to offload that
data into a separate cluster where advanced analytics are
applied. This can reduce the cost and complexity
associated with conventional data extraction,
transformation and loading (ETL), and can create a new
way of enabling exploration and insight discovery. This
could dramatically reduce the time to insight when
considering the potential value of previously unexploited
data sources.
Hadoop®
is a framework rather than a solution. That
means tools are needed to drive processes and derive insights.
Data scientists – experts tasked with solving complex data problems –
are increasingly being relied on to write the processes for analysing the data,
which means it can be time-consuming to get Big Data analytics up and running.
Individual data scientists work in everything from data engineering to
mathematics, statistics and computing, but their expertise may be limited to
one or two of these areas – it’s unlikely for any one data scientist to cover all the
bases. As a consequence, data science tends to involve teams and there have
already been concerns raised over skills shortages in this area6
.
However, the emergence of high-level tools – such as IBM InfoSphere®
BigInsights™
– is playing a decisive part in democratising distributed analytics.
BigInsights™
is built on top of Hadoop®
, and allows much more complex
analysis to be written by developers in a familiar user interface. The result is that
business users and business analysts can have direct interaction with their
data, without needing anybody to write new code for them.
Frameworks such as Hadoop®
and analytics solutions such as BigInsights™
can leverage the power of data at rest to deliver insights. But what about data on
the move? One of the most interesting and promising developments in Big Data
is the ability to carry out real-time analytics on data as it enters the organisation.
IBM InfoSphere®
Streams meets the need for real-time stream analytics,
providing a high-performance computing platform that allows user-developed
applications to rapidly ingest and correlate real-time information from
thousands of different sources. Solutions of this sort have the capability to
handle data rates that are measured in petabytes per day and they are critical if
businesses are to leverage the full potential of Big Data.
The capacity to make sense of massive amounts of chaotic and unstructured
real-time data is a game changer. Around 80 per cent of the world’s data falls into
this category. Any digital data stream is a potential candidate. These include
everything from freeform text and images to audio, voice traffic generated by call
centres, VoIP, video, web traffic, email, GPS data, financial transaction data,
satellite data and sensor outputs. All of these digital streams can be turned into
insight – often instantly – and delivered into the business ecosystem.
HOW DO YOU PUT BIG DATA TO WORK?
Businesses that have already successfully embraced
business analytics have a head start. To succeed with Big
Data, as with analytics, organisations need to develop
three core competencies.
1. The will to drive transformation
Management commitment is needed to translate
insights and outputs into action and to ensure that those
insights are readily accessible. Both the practitioners who
carry out the analysis and the leaders of the organisation
around them have a vital role to play in achieving this.
Big Data analytics need to be embedded in business
processes. Businesses may be swift to adopt analytics, but the effort is
wasted if hard-won insights are simply shunted into reports that are filed and
forgotten. This is an opportunity for organisations to change the way things are
done and to expand into new opportunities, triggering new actions across the
business on the path to greater value. The insights gained should link to
business strategy in a way that is easy for end users to understand.
Big Data demands a new approach, with exploratory and iterative analysis
to cope with semi-structured and unstructured data. This requires a new level
of dialogue between IT and business stakeholders. IT departments need to
understand the technologies and what is possible; business stakeholders need
to come forward with questions and use cases. Simply building a Big Data
platform and sitting back to see what happens is not the right approach.
2. Understanding the tools and methods
The ability to understand and apply analytics skills, tools and methods to data
is of critical importance. Businesses need to recognise that the information
94%of Hadoop®
users
perform analytics
on large volumes
of data that were not
possible before7

they hold – and are acquiring – is a unique strategic asset. Building a Big Data
platform is just the first step. High-level frameworks such as Hadoop®
can only
be exploited with access to the right tools: organisations that want to get
ahead of their competitors are either investing in high-level tools or coding
for themselves, in some cases from scratch.
Organisational capability is a factor and Big Data provides an opportunity
to make data scientists out of existing teams by taking advantage of the
accelerators that IBM has to offer that reduce barriers to entry and time
to insight.
3. An approach to veracity and relevance
Approaches to information management must evolve to take account of
Big Data opportunities. Organisations will need to recognise the fundamental
differences between traditional “refined” data and “raw” Big Data, which is
much harder to handle.
Today, most of the data used by businesses is highly
structured. Data is subjected to ETL processes in which
it is first extracted, then transformed to meet operational
needs and finally loaded into an end target such as
a data warehouse or database.
Approaches of this sort will continue to be
required for conventional high-value data. However,
the time lag between acquisition and data being
converted into something that is useful for the
business – a data mart or a dashboard, for example
– can sometimes be lengthy.
It’s important to remember that there is another
“V” in the equation – Veracity, which introduces the idea that
not all Big Data is high quality.
What’s more, organisations need to embrace the idea that not all of their
Big Data is going to be relevant all of the time. In fact, the bulk of any given
Big Data set may be of low or no relevance at all so it ought to be discarded
or pre-processed into higher order information structures (eg, routes rather
than raw location).
These Big Data sets we encounter allow us to preserve the concept of a
high integrity, high data and information quality architecture as well as a core
single version of the truth for the business.
WHERE NEXT?
l Decide what sort of capabilities you want to have as a business.
Then use a third-party partner to reduce the time to insight and help you get
your proof of concept going.
l Recognise that your Big Data project is likely to be the first of its kind.
Big Data is virgin territory, so think creatively about data sources both internal
and external. New data sources create new opportunities.
40%growth in global data
generated per year
is expected versus only
5% growth in global
IT spending8

l Information governance is key. Organisations are entering new territory,
particularly with regard to how they use their customers’ information and how
they deliver products and services. This raises important questions about
data security, privacy and – equally important – customers’ perceptions of
how their personal information is being managed.
l Businesses must understand the legal constraints governing the way
data is used and they must prepare for an open a dialogue with customers to
explore the mutual benefits of Big Data usage.
l The skills needed to bring Big Data to life are already in your
organisation. Big Data is an opportunity for businesses to unshackle their
in-house analysts. They are vital to the success of Big Data in your enterprise.
Today’s business analysts are tomorrow’s data scientists: allow them to be
the drivers and proponents of the data revolution.
HOW IBM CAN HELP
IBM has the broadest Big Data capability in the industry with a strategy that
is dedicated to helping businesses go further. To do this, IBM provides
market-leading software and the hardware needed to run it. It also provides
services and consultancy to help businesses benefit from Big Data – a
capability that is unique in the marketplace.
IBM’S CORE BIG DATA CAPABILITIES INCLUDE:
l Hadoop®
-based analytics to process and analyse any data type
across commodity server clusters.
l Stream computing to enable continuous analysis of massive volumes
of streaming data with sub-millisecond response times.
l Data warehousing to deliver deep operational insight with advanced
in-database analytics.
l Information integration and governance that allows organisations
to understand, cleanse, transform, govern and deliver trusted information
to enterprise-critical initiatives
In addition to this, IBM provides a range of supporting data platform services.
These include:
l Visualisation and discovery to help end users explore large, complex
data sets.
l Application development to streamline the process of developing big
data applications.
l Systems management to monitor and manage Big Data systems for
secure and optimised performance.
l Accelerators to speed-up time to value with analytical and industry-
specific modules.

© Copyright IBM Corporation 2012
IBM United Kingdom Limited
76 Upper Ground
South Bank
London
SE1 9PZ
The IBM home page can be found at ibm.com
IBM, the IBM logo and ibm.com are trademarks
or registered trademarks of International Business
Machines Corporation in the United States,
other countries or both. If these and other IBM
trademarked terms are marked on their first
occurrence in this information with a trademark
symbol (® or ™), these symbols indicate US
registered or common law trademarks owned by
IBM at the time this information was published.
Such trademarks may also be registered or
common law trademarks in other countries.
A current list of IBM trademarks is available on the
web at “Copyright and trademark information” at
www.ibm.com/legal/copytrade.shtml
References in this publication to IBM products or
services do not imply that IBM intends to make
them available in all countries in which
IBM operates. Copying or downloading the images
contained in this document is expressly prohibited
without the written consent of IBM. This publication
is for general guidance only.
All rights reserved.
NIW03034
RECYCLED FIBRE CONTENT
50% post consumer waste
25% pre consumer waste
25% virgin fibre
References
1. Analytics: The widening divide – How companies are achieving competitive advantage through analytics. IBM Global Business Services
in collaboration with MIT Sloan Management Review, 2011.
2. “Where angels will tread”. The World In 2012. The Economist. Nov 17th 2011.
3. “Big data market set to explode this year, but what is ‘big data’?” Joe McKendrick. Smart Planet. 21 February 2012.
http://www.smartplanet.com/blog/business-brains/big-data-market-set-to-explode-this-year-but-what-is-8216big-data/22126
4. “What is big data?” Bringing big data to the enterprise. IBM. http://www-01.ibm.com/software/data/bigdata/
5. Ibid
6. “Big data skills shortage imminent”. Karl Flinders. Computerweekly.com. 17 January 2012. h
ttp://www.computerweekly.com/news/2240114012/Big-data-skills-shortage-possible.
7. “Ventana Research Survey Shows More Organizations Use Hadoop to Perform Data Mining and In-Depth Analytics”. Business Wire. 26 July 2011.
http://www.sys-con.com/node/1920943
8. “Big data: The next frontier for innovation, competition, and productivity”. James Manyika, Michael Chui, Brad Brown, Jacques Bughin,
Richard Dobbs, Charles Roxburgh, Angela Hung Byers. McKinsey Global Institute. May 2011.
IBM Contacts
Jonathan Batson
Associate Partner
UKI Advanced Analytics Leader
IBM Business Analytics &
Optimisation (BAO)
+44 (0)7710 189256
jonathan.r.batson@uk.ibm.com
Matin Jouzdani
Senior Managing Consultant
UKI Big Data Leader
IBM Business Analytics &
Optimisation (BAO)
+44 (0)7841 495021
matin.jouzdani@uk.ibm.com

The dawn of Big Data

More Related Content

What's hot (20)

Similar to The dawn of Big Data (20)

More from The Marketing Distillery (20)

The dawn of Big Data