How to embrace Big Data A methodology to look at the new technology
How to embrace Big DataA methodology to look at the new technology
Contents
2 Big Data in a nutshell
3 Big data in Italy
3 Data volume is not an issue
4 Italian firms embrace Big Data
4 Big Data strategies and operations need enhancements
5 The “Big” misunderstanding
5 How to approach Big Data effectively?
6 The Reply value offering
7 The technological perspective
7 Big Data as a ‘Washing Machine’
8 Traditional architecture as a data source for Big Data analytics
8 Traditional and Big Data architectures working together
9 Business perspective
9 Can Big Data help in detecting insurance fraud?
11 Big Data to improve ‘churn’ analysis in the telecoms industry
12 New boundaries in customer profiling
13 Conclusion
How to embrace Big DataA methodology to look at the new technology
2
How to embrace Big Data. A methodology to look at the new technology
Big Data in a nutshell
In it’s short life Big Data assumed a wide range of mean-
ings: on the one hand it refers to the global phenomenon
of information growth, resulting from the proliferation of
activities and data generated on the net via the social net-
work, smartphones or machine to machine interactions.
On the other hand, Big Data describes a new generation of
technologies and architectures, designed to extract value
in a cost-effective manner from very large volumes of in-
formation, by enabling high-speed data capture, discovery
and analysis.
In the already well-established technical literature, ‘three
Vs’ are generally used to characterize Big Data:
Volume: the total amount of data to be managed
Velocity: the pace at which the data can be processed
Variety: the complexity and heterogeneity
of the data set
Please forget it all! Big Data solutions cannot be defined
by how you can measure data in terms of Volume, Velocity and Variety. The three Vs are just measures of data related
issues. One firm’s “big data” is another firm’s peanut as
velocity appreciation strongly depends on any single con-
text behavior.
So, what is Big Data? A nice definition says aloud: “The
frontier of a firm’s ability to store, process, and access all
the data it needs to operate effectively, make decisions,
reduce risks and serve customers”: that’s probably the real
essence of the paradigm change addressed by Big Data
technology.
Over the last twenty years, ideas of how to assemble a
decision support system have coalesced around the con-
cept of a data warehouse as a tool for navigating business
issues but today the real challenge for business intelli-
gence is to let emerge hidden value through intelligent
filtering of low-density and high volumes of information,
being them operational or unstructured data arising from
sensors, transactions or either the web. Unfortunately the
unstructured data sources may not easily and cheaply fit
in traditional data warehouses, which may not be able to
handle the processing demand imposed by unstructured
data which for this reason remains largely untapped.
To help connecting the dots of all the content that’s out
there by analyzing a huge data set and returning results
in seconds a new class of technology has emerged; it in-
cludes new tools as NoSQL databases, Hadoop and Map
Reduce. These tools form the core of an open source soft-
ware framework that supports the processing of very large
data sets across clustered systems. Let we show you how
and why these technologies are gaining a leading position
in the interest of companies
3
Big data in Italy
In January 2013, in collaboration with Forrester Consult-
ing, Reply carried out a survey to evaluate interest in adop-
tion of Big Data solutions in Italy. The aim was to explore
not just the acceptance of Big Data but also the stage of
maturity reached by organisations in building their strat-
egy towards Big Data implementation.
The study spanned vertical sectors across the country’s top
100 organisations, with concentrations in financial ser-
vices, telecoms, energy, utilities and waste management,
retail and professional services. Key findings included the
following, in such a way surprisingly, results:
Data volume is not an issue
The results of the survey confirm that a high percentage
of respondents do not have to deal with ‘petabytes’, ‘zet-
tabytes’ or ‘yottabytes’ of data. The most of the Italian
companies manage volumes of data that are relatively in-
significant in comparison with the vast size of major enter-
prises, being them social networks like Facebook, Twitter
or important retailers as Walmart and Target.
Anyway, Italian companies seem to have realised that vol-
ume is not the only or primary characteristic of Big Data.
The interest in Big Data technologies is then driven by
the necessity to acquire and process heterogeneous data,
while fastening computational time at a greater level of
accuracy. This is pretty similar to the findings of a recent
study conducted in the U.S., where it became evident that
the amount of useful data generated inside and outside
the company is not raising the hugeness of the major so-
cial network.
0 20 40 60 80 100
UNSTRUCTUREDDATA
ESTIMATE THE SIZE/VOLUME OF DATA WITHIN YOUR COMPANY
SEMI-STRUCTURED DATA
14% 24% 26% 25% 7% 2% 2%
STRUCTURED DATA FROM
TRANSACTIONALSYSTEMS
13% 15% 44% 18% 6% 3% 1%
11% 25% 24% 26% 7% 2% 3%
>1000TB 100-1000TB 10-100TB 1-10TB <1TB None Don’t know>1000TB 100-1000TB 10-100TB 1-10TB <1TB None Don’t know
4
How to embrace Big Data. A methodology to look at the new technology
Italian firms embrace Big Data
Reply identified a significant amount of interest in Big
Data technologies and solutions. It looks as the wish to
take a competitive advantage from the analysis and inte-
gration of unstructured data is driving companies to adopt
Big Data technologies.
Notwithstanding only around a quarter of respondents
have already implemented a solution, 40% were planning
to implement in the next 12 months and a further 28%
planned stretching out to a slightly longer time horizon.
These companies are struggling with the data coming into
their organisations and are looking for new methods to
better leverage data to improve their businesses.
Big Data strategies and operations need enhancements
Key goals focus firstly on data quality, followed by busi-
ness objectives. The business cases that companies have
developed do not measure concrete key performance indi-
cators. Moreover, Italian organisations aren’t ‘pushing the
envelope’ when exploring the potential of Big Data.
Although the demonstrated significant interest for the new
technology, Italian organisations need definitely to invest
more attention in improving their Big Data strategies and
operations. When asked about the most important goals
or drivers organisation cares while overall orchestrating
its business intelligence strategy, 34% of respondents
pointed to improving data quality and consistency. But
data quality is not the end goal. The whole idea of Big
Data is to improve business success, through factors as
customer insights, operational efficiencies and cost con-
trol. Business targets must come first and data quality is
a prerequisite for these.
Only 11% of respondents claimed to have a business case
for Big Data with concrete KPIs and proven ROI. They
represent the highest level of maturity in the Big Data
initiatives. A further 19% reported having a business case
with KPIs but no proven ROI. The majority (47%) have a
business case with intangible benefits.
As shown by the results, 70% of respondents are not yet
able to translate the advantages of Big Data initiatives into
tangible business benefits. This would indicate that fur-
ther expertise is needed to lead Italian companies into the
Big Data world. Starting small and demonstrating tangible
benefits will enable organisations to prove the ROI on a
small scale before ‘going big’.
IMPLEMENTED,NOT EXPANDING
EXPANDING/UPGRADINGIMPLEMENTATION
PLANNING TO IMPLEMENTIN MORE THAN 1 YEAR
PLANNING TO IMPLEMENT INTHE NEXT 12 MONTHS
INTERESTED BUT NO PLANS
NOT INTERESTED
DON’T KNOW
“BASED ON FORRESTER’S DEFINITION OF BIG DATA, WHAT BEST DESCRIBES YOUR FIRM’S CURRENT USAGE/PLANS TO ADOPT BIG DATA TECHNOLOGIES AND SOLUTIONS?” (SELECT ONE)
3%
19%
28%
40%
10%
0%
0%
WE HAVE A BUSINESS CASEFOR BIG DATA WITH MEASURABLE
KPIS AND ALREADY PROVEN ROI
WE HAVE A BUSINESS CASEFOR BIG DATA WITH MEASURABLEKPIS AND A PROJECTED BUT NOT
YET PROVEN ROI
WE HAVE A BUSINESS CASEFOR BIG DATA BUT WITH
INTANGIBLE BENEFITS ONLY
CURRENTLY WE HAVE NOBUSINESS CASE, BUT WE ARECURRENTLY WORKING ON ONE
MAT
URI
TY
WE HAVE NO EXPLICITBUSINESS CASE FOR BIG DATA
DON’T KNOW
“DO YOU HAVE A BUSINESS CASE FOR YOUR BIG DATA INITIATIVE IN PLACE?"
19%
19%
11%
4%
0%
47%
5
The “Big” misunderstanding
A cause of frustration for the customer trying to tap into
the ability to design and embrace a strategy about Big
Data is the fundamentally misleading view of Big Data as
a social phenomenon on the net, generated by millions or
even billions of pieces of information and backed by tech-
nologies that have been developed to extract the hidden
value of that data.
Too often technology key users (marketing or sales depart-
ment, product development team, security and fraud of-
fices, to mention just a few), are asking for solutions that
will never come because echoing the traditional approach.
It is not simply a matter of technology. Within the func-
tional organisation Big Data demands new processes, a
different way of interacting with the end customer, even
new skills to leverage the increased power of the analysis.
Simply Big Data requires a shift, in the corporate analyst
behaviors, to leverage the potentiality of new information
made available and in the IT departments, to deploy a
new array of IT architectures that will enable companies to
handle both the data storage requirements and the heavy
computational processes needed to handle cost-effective-
ly large volumes of data.
How to approach Big Data effectively?
The survey, in line with our overall understanding of com-
pany’s behaviours, suggests that the strategy to deal with
Big Data challenges will strongly differ depending on the
maturity of the organisation towards this topic. Reply has
developed a ‘Big Data maturity model’ to measure the or-
ganisation’s aptitude in approaching Big Data. The real
aim of this model is to help CxOs in better understanding
the company behavior alongside the new technology and
then properly identify the correct strategy for implement-
ing a coherent and profitable Big Data project.
We can segment company’s behavior into 4 blocks:
Inactive: Companies deal with Big Data issues as a
storage problem and essentially deny that there is a
problem. When issues come up, they just try to fix the
problem using standard techniques. This approach
results from a lack of business awareness and has
several failings: it is expensive and unpredictable.
Proactive: Companies have the technologies and the
infrastructures to deal with Big Data but they still
don’t have business cases with measurable KPIs and
proven ROI.
Reactive: Companies have business cases and the
maturity to start a Big Data project but lack of ability
and expertise to address technological issues.
Active: Companies view Big Data as an asset and
own the necessary human resources, processes and
technologies to gain insight into their data. These
companies looks at Big Data as an opportunity to dif-
ferentiate and gain competiveness, while well under-
stand it is not the last technological hype. The final
goal is then putting in place a comprehensive strategy
to maximize the data value to business purposes.
6
How to embrace Big Data. A methodology to look at the new technology
The Reply value offering
As the model demonstrates there are several challenges on
both technical and organisational side that must be care-
fully addressed to achieve the full potential of Big Data,
while finding the right solution involves more than a simple
evaluation of price/performance of any specific tool.
Reply has built an integrated a consistent methodology
to support clients in the development of suitable strat-
egies to let them able to benefit of best of breed solu-
tions. A multidisciplinary team of business analysts and
technologists has been established to address the main
issues related to a Big Data project implementation within
a comprehensive approach. Additionally, to help business
people in challenging value from data, has been founded
a data scientists team. The goal is to help customers by
proposing the most appropriate business and technology
model fittings to their needs.
This heterogeneous team can help companies at any stage
of the Reply’s maturity model:
Inactive: The first stage, where the organisation has no
expertise in Big Data. Technologists provide the archi-
tectural solution, while business analysts and data
scientists collaborate with business in discovering new
patterns from the data, to create a business case with
a proven ROI.
Proactive: Organisation has already gained experience
in the technology but do not know how to apply it to a
real business case. Reply’s business analysts can help
supporting the development of a Big Data roadmap,
to transform customer’s needs into a real business
scenario. Data scientists work with business analysts
in finding new insight and perspectives, helping the
company to improve data value.
Reactive: Organisation has established a business
case with measurable KPIs but it lacks the technical
experience necessary to develop a Big Data archi-
tecture. Reply’s technologists can help customers in
finding the best architectural solution. The first step
is to analyse the organisation’s data sources and IT
infrastructure. The overall overview of the technical
environment enables technologists to develop a Big
Data infrastructure tailored to the customer’s goals.
Active: Organisation has here reached a high maturity
level in both the technology and business issues. Fine-
tuning job can still be done, however, to make it easier
to better develop business opportunities: this is the
typical environment where data scientists can, use their
expertise to refine the logical approach to data discove-
ry and modeling to deliver more detailed insights.
In summary, Big Data may be approached by two different
roadmaps: starting from the business issues, using Big
Data as a very powerful tool to redesign and improve data
analysis processes and from a technological perspective,
looking at Big Data solutions to reshuffle best practices
and infrastructure in order to provide faster and cheaper
results.
7
The technological perspective
The most important value offered by the Reply’s approach
to Big Data implementation lies in the development and
delivery of solutions which strongly fits with the custom-
er’s technological architecture.
The goal is addressing the resolution of specific business
problems while maximizing the safeguard of the current
investments in technology, through a gradual integration
of the Big Data architecture into existing legacy systems.
Founding on the distinctive competencies and wide op-
erational expertise of the Group companies, Reply devel-
oped a framework to tailor Big Data implementation in
three different scenarios.
Big Data as a ‘Washing Machine’
A major problem when approaching a data warehouse solu-
tion is represented by extracting data from outside sources,
transforming and loading it into the data warehouse. ETL
processes (extract, transform, load) can involve consider-
able complexity while significant operational problems can
occur with improperly designed ETL systems, whereas – for
the business purposes – they represents a null-value, ex-
pensive and time consuming set of activities.
Big Data can solve this problem by substituting the tra-
ditional ETL process with a new kind of storage architec-
ture and, on top of this, a processing layer able to quickly
transform data and load it into a data warehouse.
This approach can appreciably lower the overall time
needed to satisfy the necessity of building the base of
reporting/analytics value chain, at a fraction of the cost
incurred by the traditional approach. It will also introduce
a faster management of the quality and coherence of the
data ‘ignited’ into the systems.
DATA
IN
GEST
ION
&
STO
RAGE
DATA
DATA
AN
ALYT
ICS
&
VISU
ALIZ
ATIO
N
PRESENTATION
CALCULATION ENGINE
DATA MART
DWH
STAGING AREA
CALCULATION ENGINE
DATA STORAGE
Unstructured Structured
8
How to embrace Big Data. A methodology to look at the new technology
Traditional architecture as a data source for
Big Data analytics
Traditional data warehouses can handle many situations
but they do have limits. The volume of data imported into
a data warehouse is a critical issue in terms of costs to up-
grade the system and in data elaboration time. The higher
the volume, the greater the impact on processing perfor-
mance. The usual solution for this problem is to back up
the data but in most cases this is tantamount to losing
the information.
Big Data architectures can load data from existing data
warehouse systems and process it along with data from
sources such as data streams or unstructured data that
are not easily managed by the traditional data warehouse.
There are many benefits from using this approach; it is
possible, for example, to combine classical structured data
with other sources, enabling new insights and achieving a
better granularity in the data analysis. Furthermore, hav-
ing a Big Data storage structure means that data coming
from the data warehouse will never be lost; it will always
be possible to use historical data and analyse it.
Traditional and Big Data architectures working together
In some cases, thinking of replacing or supplementing IT
architectures can be valued as a disruptive approach, so
that the Companies prefer to keep their incumbent sys-
tems despite the loss of the information that - if properly
used - could dramatically upgrade their competiveness or
the revenue streams.
Big Data architecture, more than others solutions, allows
companies to implement a parallel infrastructure to ex-
ploit new data sources in counterpart with the traditional
B.I. ones. The cost-effective hardware jointly with the
open source software, which represents the foundations of
the Big Data solutions, enable a company to manage both
scenarios at a very marginal differential cost. Moreover,
some of the tools that belong to the Big Data ecosystem
(e.g. analytics, presentation and data integration layers)
can be substituted by or integrated with resources already
present in the traditional architectural stack.
DATA
AN
ALYT
ICS
&VI
SUAL
IZAT
ION
DATA
IN
GEST
ION
&
STO
RAGE
DATA
PRESENTATION
CALCULATION ENGINE
DATA STORAGE
DWH
STAGING AREA
ETL - ELT
Unstructured Structured
DATA
AN
ALYT
ICS
&
VISU
ALIZ
ATIO
N
DATA
M
ANAG
EMEN
T &
STO
RAGE
DATA
PRESENTATION
ANALYTICS
DWH
STAGING AREA
ETL - ELT
CALC. ENGINE
DATA STORAGE
Unstructured Structured
9
Business perspective
Our daily confrontation with CxOs make clear that many
organisations start in claiming how business intelligence
solutions are failing to meet their current business needs;
this is the major push to accept looking at Big Data as
the ultimate instrument to design a new, more effective
information strategy. From being the sort of tool that was
only needed for meteorology or mathematical simulations,
Big Data has pretty recently moved into the industry main-
stream as the easiest and cheapest way to overcome tra-
ditional costs and implementation times of complex data
management systems, essential to encompass and manage
heterogeneous and multi source data sets.
Not all industries are likely to benefit from Big Data projects
equally and not surprisingly, the first movers were internet
companies; in fact, the most popular Big Data platforms
has been built on top of software originally used to batch
process data for search analysis but now retail, telecom,
financial services and media sectors are quickly recovering
while manufacturing and process industry are definitely ap-
proaching.
But just having the Big Data tools isn’t enough: enterprises
need to know what questions to ask, actually ask them and
then translate that into strategy or tactics. It will be impor-
tant for enterprises to develop new policies around privacy,
security and intellectual property. Big Data isn’t just about
technology and employees with the right skill sets, it will
also require businesses to align work flows, processes and
organization to get the most out of it. It is important to note
that enterprises should not concentrate on destructured
data at the expense of “current data” or business informa-
tion as normal. There is still a lot of value to be extracted
from the information inside their traditional databases!
Reply can help customers in designing and addressing the
right path to define an appropriate strategy, by identify-
ing business cases where a Big Data approach can create
a true difference to meet unsolved organisation’s needs.
Below are summarized some of the most common usage
patterns explored by Reply; while the explanation of the
usage patterns may be industry-specific, the rational basis
can be applied across industries to bring new sparks that
ignite the change.
Can Big Data help in detecting insurance fraud?
The technology that most insurers have currently in place
to help to fight frauds is a blend of business rules and
database searches, where the results rely heavily on the
sensitivity of the claims auditor. While these techniques
have proved being successful in detecting known fraud
patterns, insurers today need to invest in new analytical
capabilities to help them to spot unknown and complex
fraud activities. These analytical capabilities include in-
congruity detection, predictive modelling, unstructured
data mining and social network analysis.
Anomaly detection aims at discovering fraud by identify-
ing those elements that vary from the norm. Key perfor-
mance indicators associated with tasks or events are base-
lined and thresholds set. When a threshold for a particular
measure is exceeded, then the event is reported. Outliers
or anomalies could indicate a new or previously unknown
fraud pattern.
10
How to embrace Big Data. A methodology to look at the new technology
Predictive models use past fraud events to produce fraud-
propensity scores. Adjusters simply enter data and claims
are automatically scored against the likelihood of them
being fraudulent. These scores are then made available
for review. Use of predictive modelling makes it possible
to understand new fraud trends.
Since around 80 percent of claims data is unstructured,
the use of tools able to mine unstructured data enables
insurers to analyse information arising from medical
chronicles, police records, external and internal database
sources or even e-mails.
Social network visualisation tools allow investigators to ac-
tually see network connections so they can uncover previ-
ously unknown relationships and conduct more effective
and efficient investigations.
By using Big Data technologies companies are able to
manage all of these issues and to ‘learn’ from experience
to improve their fraud detection and pattern identification
capability. This learning characteristic enables the soft-
ware to adapt and increase in sophistication as more and
more intelligence is gathered over time. The more analyti-
cal the tools, the higher the chance of detecting fraud in
the early stages and predicting potential areas of abuse
before fraudsters discover the opportunity themselves. Au-
tomation also means less reliance on the human element,
and provides greater accuracy and homogeneity in fraud
discovery activity.
Reply has established a proven methodology to apply a
Bayesian model in fraud recognition combined with Big
Data analysis techniques. This is a comprehensive ap-
proach, which includes data discovery through all the
available internal and external structured and unstruc-
tured data sources, combined with the powerful computa-
tional capabilities of a Big Data infrastructure to support
the claims manager in every phase of the investigation.
First of all, a network analysis will identify any histori-
cal relationship between the actors in a specific claim,
revealing any connection in the past that could suggest a
propensity to commit a fraud. Then a clusterization of the
actors and related behaviors based on a self-learning sta-
tistical model let emerge similarities in the data model, to
better represent relations and attitudes to plausible fraud
existence.
While this technology is still in its early stages, the bottom
line is that new Big Data analytics can be used to explore
large volumes of networked data, using high-speed pro-
cessing with configurable data entry from multiple internal
and external sources, to reveal fraudulent behaviour. Can
you imagine how far you could go using a so strong para-
digm change in tracking frauds?
Risk/TarifEvaluation
Internet data base
Externaldata baseClaims
ManagementsFraud
Monitoring
Real time evaluation
BRMS
Work�ow MgmtFraud reporting
SOGEI
MCTC
ANIA / ISVAP
Others
Case Analysis
Case Assignment
Case Manager Dashboard
Scoring
Data Certi�cationClustering
Network analysis
Big Data Analytics Data Matching
USERS ACTUARIALCLAIM MANAGER
RISKMANAGER
ContractsCustomersClaims
Fraudsblack list
Big Data to improve ‘churn’ analysis in the telecoms industry
Today’s customers want competitive pricing, value for mon-
ey and, above all, a high quality service. They won’t hesitate
to switch providers if they don’t find what they’re looking
for. So particularly in mature markets or where regulations
and service dematerialisation makes ‘churn’ easier, it is ab-
solutely crucial to put in place a sustainable and robust
strategy for customer retention to preserve customer life-
time value. The telecoms market provides a good example
of why the high acquisition costs and slim profit margins for
each customer make churn analysis vital to help companies
identify and retain the most profitable among them.
In this context, the paradigm change ‘more is more’ is in
tune with the main aim of Big Data analytics. The uncov-
ering of hidden value, through the intelligent filtering of
low-density and high volumes of data, can become a real
differentiating factor. The more data you have, and the
more recent and accurate it is, the faster you can learn
from it and the more predictive you can be.
The value of Big Data can then be exploited in two dif-
ferent directions: to decrease the capital expenditure
(CAPEX) or operational expenditure (OPEX) associated
with the computational infrastructure needed to address
the huge amount of data used to feed predictive analyti-
cal models; and/or to increase the data sources used for
the integration and leverage of new kinds of unstructured
information, enabling companies to better describe and
understand customer behaviour.
One method now emerging to enable an operator to move
from reactive churn management to proactive customer
retention is to use predictive churn modelling based on so-
cial analytics to identify potential ‘churners’, thereby ena-
bling the operator to act on such predictions, rather than
waiting for explicit trigger points (e.g. credit on prepaid
card running down), by which time the churn is most prob-
ably inevitable, irrespective of any act or offer on the part
of the operator. Big Data analytics offer the opportunity
to process and correlate new data sources and types with
traditional ones, to achieve better results more efficiently
and receive insights that will set alarm bells ringing before
any damage has been done, so giving companies the op-
portunity to take preventive measures.
Pricing analytics and ‘next best offer’ recommendations
in particular are classic examples of how, by analysing
structured data (such as CDRs) and unstructured or semi-
structured data types (such as log files, IVR tracked calls
to call centres, clickstreams and, ultimately, text from
e-mails), telecoms operators can provide more accurate,
personalised offer recommendations.
Last but not least is the issue of timing. It is true that
traditional business intelligence solutions have allowed
enterprises to move forward by consolidating data sources
into centralised data centres. However, this data is used
‘simply’ for reporting. We are now moving into a new era
where information can and must be converted into real-
time actionable insight, to enable the company to respond
in real-time to behavioural changes in the customer mind-
set or to react quickly to threats on the competitive hori-
zon. This is exactly why and where Big Data analytics can
win the battle against ‘old’ BI tools.
VOICENETWORK
DATA
MOBILE WEBNAVIGATION
DATA
CUSTOMERINTERACTION
DATA
CELL TOWERSDATA
CRMTOOLS
AD SERVER
CAMPAIGNMNGT
CALL CENTER
HDFS &MAP
REDUCEREAL TIMEANALYTICS
BIG DATA PLATFORMOPERATIONAL
STACK
INSIGHTS
Feedback
11
How to embrace Big Data. A methodology to look at the new technology
12
New boundaries in customer profiling
Customer analytics start with data. To get better customer
insight, most companies begin by analysing their struc-
tured transactional data, which typically includes infor-
mation such as demographics, purchase history, com-
plaints and retention information. Statistical algorithms
can help companies to create meaningful segments and
gain insight into buying patterns. These insights and ten-
dencies are then encapsulated in models which are used
as a basis for future predictions; basically, an extrapola-
tion of past history. Is this enough in today’s markets?
Probably not!
In recent years every one of us has become a powerful
‘walking data generator’, delivering personal information
(that reflects daily changes in our habits) through many
different channels. Information sources include call cen-
tre records, email communications and transactional data
as well as usage patterns on company websites. Very few
enterprises, however, are in a position to probe this ‘gold
mine’ of information.
In their quest to make these models more accurate, com-
panies are starting to embrace new sources of data; but
most of this data is unstructured and it is quite expensive
to have it integrated into traditional data warehouse and
data-mart infrastructures, both in terms of cost and time.
Moreover, analytical algorithms are continuing to evolve to
deal with the changing landscape brought about by new
trends (such as mobility, social media and e-Commerce),
while the need for a very fast computational time is in-
creasingly becoming a necessity to help companies to seg-
ment their customer base more effectively, attract more
profitable customers, improve campaign handling or re-
duce customer churn. Propensity models are also becom-
ing more dynamic to deal with the geo-spatial and tempo-
ral dimensions, acknowledging the fact that location and
time events impact people’s propensity to react to external
stimulation; in this case, the ability to react in real-time or
near real-time becomes a ‘must have’ feature.
As demonstrated by a recent Reply project, Big Data
technologies provide a very powerful tool-set to address
all of these issues. The ability to digest and elaborate in
real-time huge amounts of data as single cash lines in
till receipts, and compare them with the purchase his-
tory of each customer in order to generate promotions in
real-time is without any doubt a capability that would be
extremely hard to achieve using traditional analytics solu-
tions - which would in any event be prohibitively expen-
sive. The more data and information to be analysed, the
longer the process required (days); while Big Data solu-
tions allow retail companies to analyse huge volumes of
data, with more granularity, in a shorter period (hours vs.
days). Retailers can now get insight into customers’ sea-
sonal trends and use it to improve the management of
stock or create tailored pricing and promotions.
While embracing this new customer approach companies
must be aware there is a very fine line between using cus-
tomer analytics to create value by serving customers with
customised precision, and destroying value by surprising
customers with actions that erode trust. Privacy policies
and a consistent execution across the enterprise are es-
sential and must be properly performed to understand the
ever-narrower segmentation of customers and so deliver
much more precisely tailored products or services. It is
worth it, however, and the reward will surely overcome
best expectations.
Conclusion
While other business metrics come and go, growth con-
tinues to be the most important criterion used to meas-
ure companies value, the measure by which the market
assesses companies and managers evaluate their perfor-
mance compared to competitors. Daily we appreciate as
competiveness passes more and more through a better
understanding of the huge amount of data organizations
collect and store about employees, customers, finances,
vendors, inventory, competitors and markets, to name only
a few. The amount of data needed is important because
people generally make better decisions if they have more
data available to them.
In parallel, even more in the coming years we will ap-
preciate the increasing volume and detail of information
captured by enterprises as the rise of multimedia, social
media and the Internet of Things will fuel exponential
growth in data for the foreseeable future. The real issue is
data have swept into every industry and business function
and are now an important factor of production, alongside
labor and capital.
As organizations will definitely understand this pattern
and invest to become more dependent on information,
the processes of gathering, managing, and utilizing data
will become more central to operational success, because
data is only as valuable as our ability to access and extract
meaning from it. This is probably the main reason why Big
Data solutions have definitely left their primordial field of
application, entering to its own right the industrial world.
Also if there could be reasons to be skeptical about the
Big Data expansion we can say without risk of contradic-
tion that a disciplined, targeted approach to Big Data de-
serves a very focused attention; when organizations will
recognize that Big Data’s ultimate value lies in generat-
ing higher quality insights looking in a different way to
available data to enable better decision making, interest
and related revenues will accelerate sharply. Albeit in this
field Big Data is still in its infancy, the rapid and constant
growth of attention to this technology suggests that indus-
try begin to embrace the challenge and is ready to take on
transformative measures, using the next generation of Big
Data industrial solutions.
Then, the final and most important question is: are you
ready to harness the power of Big Data?
13
www.reply.eu