The document discusses how communication service providers can leverage big data analytics to improve customer satisfaction and extract business value. It recommends implementing an Analytics Big Data Repository to provide automated insight into customer usage and performance in real-time. This would allow customizing marketing campaigns and offers for each customer based on their real-time usage patterns and behaviors. It would also enable real-time network optimization and customer experience management through issues detection and resolution. The repository is presented as a solution to data replication challenges that can support multiple use cases and analytical systems simultaneously.
An Investigation on Scalable and Efficient Privacy Preserving Challenges for ...
Etiya White Paper_ABDR
1.
Analytics Big Data Repository: Key to
Customer Satisfaction
An introductory white paper by Etiya
2.
Contents
EXECUTIVE SUMMARY
....................................................................................................................................
0
INTRODUCTION
................................................................................................................................................
1
ABOUT BIG DATA
.........................................................................................................................................................
2
RETHINKING DATA MANAGEMENT
...............................................................................................................
3
CALL DETAIL RECORDS (CDR) AND TRAFFIC ANALYSIS
.......................................................................................................
3
Traffic
Analysis
.....................................................................................................................................................
4
Analytics
Big
Data
Repository
.............................................................................................................................................
4
EXTRACTING BUSINESS VALUE
.....................................................................................................................
5
CAMPAIGN MANAGEMENT
.............................................................................................................................
6
REAL TIME
...................................................................................................................................................................
7
CONCLUSION
....................................................................................................................................................
7
3.
Executive Summary
The new technology advancements, quality issues and cost pressures are among the key
drivers forcing communication service providers (CSPs) to shift toward more accurate ways
for controlling network planning and enhancing quality of service. With the evolution of
networks and access devices, new information has become available, turning extremely rich
material into a valuable competitive differentiator for CSPs.
New tools could help CSPs differentiate themselves in a very competitive environment. More
sophisticated and real-time analytics solutions as well as tools to manage big data have been
hailed to decrease churn and to understand how their customers consume their services. By
doing so, they are able to increase average revenue per user and offer more customized
services, generating new revenues.
The authors of this white paper believe that analytics data repository is a tool and a measure
that is made available to CSPs, and recommend that the big data repository is implemented
as to provide detailed and automated insight into transaction paths and performance.
In this white paper, it is aimed to provide a broad understanding of telecom data (structured
and unstructured), the challenges of big data analytics and possible solutions to these
challenges. In addition, while the term big data is used in the context of the CSPs, same
concept applies to Information Technology (IT) organizations.
4. 1
Introduction
From the beginning of the Information Technology, the demand for storing and processing
data has always been ahead of the capability of the technology and tools available. It may be
the first time in the history that the data storing and processing technologies have advanced
beyond the “simple” needs of people and industry. The past few years have witnessed
immense commercial investments in solutions that address the processing and analysis of big
data. Big data opens a vast array of applications and opportunities in multiple vertical sectors
including, but not limited to, retail, hospitality, energy, media, utilities, healthcare and
pharmaceutical, telecom, manufacturing, government and transportation, homeland defense
and security, and the emerging technology fields.
In the past era, the most of the Communication Service Providers (CSPs) invested heavily on
the big data technologies. They have been having challenges on unlocking the data potential
accumulated in their environment. In order to create value from the data deluge, big data
management and strategies need to be aligned with business strategies and plans.
Leveraging big data is sought after. CSPs need to store what is important and extract results
to serve their own business objectives. Most of the CSPs are currently seeking ways to utilize
the big data, converting hard copies to soft copies and immigrating this information into
cloud, turn them into value for their businesses in the form of customer satisfaction, reduced
operational costs, increased service quality, and increased profits. The CSPs are extracting the
means of knowing more about their customers (their shopping habits, what ticks them off to
5. 2
make a purchase) to decrease churn, maintain smarter networks, and to generate new
revenues.
Big Data refers to a massive volume of both structured and unstructured data that is so large
that it is difficult to process using traditional database and software techniques. Telecom
structured data sources are many and varied. Some sources are completely static or semi-
static while others are very dynamic in nature.
The telecom industry was an early adopter of data mining technology and hence many data
mining applications exist; such as fraud detection, customer profiling and network fault
isolation. The CSPs maintain a great deal of data about their customers. In addition to the
general customer data that most businesses collect, the CSPs store the call detail records,
(CDR) which precisely describe the calling behavior of each customer. This information can be
used to profile the customers and these profiles can then be used for marketing and/or
forecasting purposes.
Big data offers CSPs a real opportunity to gain a much more holistic view of their operations
and their customers, and to further their innovation efforts.
About Big Data
Big data represents a challenge in the technological capabilities to store and analyze data, in
accordance with 3 V’s of big data: Volume, velocity and variety. The telecom industry
generates and stores a tremendous amount of data, which is difficult to calculate manually.
Data volume that is considered to be big data now will be comparably small data in the near
future. Thus, “big” does not describe the real size of data but its relative size to the capabilities
of technology of the present day. Policy and rules are just the beginning of new era where
telecom networks will adapt themselves in real-time to subscriber’s needs, profile, and quality
of experience (QoE) management.
Big data comes from a greater variety of sources, both structured and unstructured, in the
form of text, audio, video, and human language, as well as semi-structured data including
XML and RSS feeds, and multi-dimensional data from a data warehouse to add historic
context to big data. The type of these data are called call detail data, which describes the calls
that traverse the telecommunication networks, network data, which describes the state of the
hardware and software components in the network, and customer data, which describes the
telecom customers.
Telecom data pose several interesting issues for data mining. The first concerns scale, since
databases may contain billions of records and are amongst the largest in the world. A second
issue is that the raw data is often not suitable for data mining. Because many data mining
applications in the telecom industry involve predicting very rare events, such as phone fraud,
rarity is another issue. The fourth data mining issue concerns real-time performance: many
data mining applications, such as fraud detection, require that any learned model/rules be
applied in real-time.
6. 3
Telecom networks are extremely complex configurations of equipment, comprised of
thousands of interconnected components. Each network element is capable of generating
error and status messages, which leads to a tremendous amount of network data. This data
must be stored and analyzed in order to support network management functions. Due to the
enormous number of network messages generated, IT technicians cannot possibly handle
every message.
Rethinking Data Management
Customer loyalty is becoming difficult to manage since the Telecom customers are easily
attracted to other CSPs by a more attractive offer, plan, or new device. Traditional revenue
streams, such as short message service (SMS) and international direct dialing (IDD) calls, are
being eroded by over-the-top (OTT) players such as WhatsApp, Skype, Tango, SnapChat, and
Viber. Meanwhile the explosive use of mobile devices creates complexity and big demand on
network capacity. This section will highlight call detail records and how they can be used to
get all the insight into key call quality metrics and variables.
Call Detail Records (CDR) and Traffic Analysis
Each time a call is placed on a telecom network, descriptive information about the call is
saved as a call detail record. The number of CDR generated and stored is huge. Most CSPs
store several months of call detail online, leading to the storage of tens of billions of call detail
records to be stored at any time. CDRs include adequate information to describe the
important characteristics of each call at real-time. At a minimum, each call detail record will
include:
• the originating and terminating phone numbers,
• the date and time of the call,
• the duration of the call.
CDR is available to be mined and analyzed near real-time. Call detail records are not used
directly for data mining, since CDR holds data at the call level. CDRs needs to gathered and
summarized into a single record to describe a caller’s behavior. The choice of summary
variables is critical in order to obtain a useful description of the associated customer. Some of
the variables can be listed as follows:
• average call duration
• % no-answer calls
• % calls to/from a different city code
• % of weekday calls
• % of daytime calls
• average # calls received or originated per day
• # unique area codes called during a specific time period
These variables can be used to build a customer profile.
7. 4
Traffic Analysis
In an Etiya study, a month’s one Terabyte network data was analyzed for a leading CSP in
Turkey to forecast wireless network traffic and predict anomalies. Etiya performed customer
churn analysis on IVR and Call Center data. All tasks were performed in real-time for three
scenarios. The first aimed at the single view of the customer and to create a customer
storyboard. Etiya enriched the customer data with demographic information and then,
calculated Customer Experience Index (CEI) using pre-defined network KPIs. Predictive
analytics were followed by specific customer mobility on usage map. The second scenario
was to create a proactive customer monitoring profile by predicting Quality of Service (QoS)
and Quality of Experience (QoE). An alarm-based solution was created to notify the operation
upon a decreased QoS or QoE. Etiya was able to provide and predict the next ten minutes of
future events. The third aimed at tasks against a subset of customers; i.e. VIP subscribers.
When predefined errors were analyzed, it was found that QoE with number of errors and
delays could be forecasted per hour. If human error, call-centers could be conditioned against
these possible mistakes.
Analytics Big Data Repository
At Digital Disruption event of TM Forum in December 2014, laid the groundwork for the
creation of a common data repository to eliminate the need for multiple copies of data. At the
third phase of the project, Etiya team is a part of developing an innovative Analytics Big Data
Repository (ABDR) to help CSPs advance customer experience and reach business growth by
avoiding data replication, saving on extract, transform and load (ETL). ABDR provides a
unified layer that can support multiple use cases and analytical systems while addressing the
challenges such as hardware space and energy costs, etc.
In the Etiya use case, the scenario is created as follows:
Customers decide on their threshold level for usage types. CRM system receives the selected
thresholds. Data is collected per usage type of the customer in the Analytics Big Data
Repository: Voice Usage, Internet Usage, and SMS Usage. Usage data is then fed to the
Complex Event Processor (CEP). CEP compares the accumulated usage against the selected
thresholds and triggers actions to notify customers real-time if they exceed the limits they
decided.
8. 5
Extracting Business Value
Communication service providers conduct analytics programs that enable them to use their
internal data to boost the efficiency of their networks, segment customers, and drive
profitability with some success. The potential of big data poses a challenge on how to
combine large data sets across the entire telecom value chain, from network operations to
product development to marketing and customer service and to monetize the data itself as
data as service (DaaS). In order to gain leverage, CSPs manage the four key attributes of big
data:
• Volume - the quantity of data to be captured continues to grow exponentially.
• Velocity – the bits and bytes have to be processed at high speed.
• Variety - data comes in many formats, from diverse sources.
• Value - data needs to be converted into meaningful insights.
9. 6
Campaign Management
In the advertising business, sensory inputs
help the marketers and advertisers correct
course of action in order to meet their one
specific goal to drive people into a store, to
buy a product or service, or register for an
event. The internet acts as a collector of
these sensory inputs from multichannel,
providing a huge variety in data. In an
effort to cut through the advertising clutter
and ensure that consumers take a note of
their ads, some marketers try to increase
by buying variety of ad space within the
same environment (plane, bus, etc.) or the
sensory input by adding empty space in
order to accentuate the brand logo or
product illustration. In marketing, the
challenge comes from activating new
campaigns as the capacity of a measuring
one campaign may not produce the same
reading when measured under the same
prescribed conditions. In today’s
competitive landscape, the latency is no
longer acceptable. Latency is also caused
by inefficient approval flow with CSPs. The
campaign management needs to stream
digital data and run a real-time analysis.
Customer Experience Management (CEM)
emphasizes the concept of offering
targeted and tailored campaigns to
customers. Finding the best offer for the
customer, requires a detailed analysis on
customer behavior. There are technologies
and search engines that assemble real time
offers to shoppers in 200 milliseconds,
using advanced analytics based on
location, age, gender, and both historical
and immediately preceding online activity,
along with the most recent responses of
other customers. In today’s campaign
management systems, there is a lack of
unified customer view for marketing
campaigns. Customer behaviors stored
data may be stored in different databases.
10. 7
Therefore, there is a gap between how often the customers use one service and how much of
the data or voice is transferred, using their quota of services and the amount paid for these
services. Helping the customers find the optimal price plans mean helping the operator itself.
The business implication of happy customers is the increased revenue through increased
ARPU, increased acquisition rates and decreased churn rates.
Most of the customer related analytics on the quest for finding the best offer relies on the
usage behavior of the customer. Usage data is processed and stored in high volumes,
therefore analyzing the usage data for each customer is in the scope of the big data
technologies. Currently most of the campaign offers are determined through usage analytics
by manual processes. Many customer service providers employ ‘Business intelligence’ teams
to analyze the information stored in their enterprise Data Warehouses and distribute the
campaign proposals to their outbound sales channels periodically. To automate this process,
a powerful analytics engine is necessary as well as a good business strategy to implement
proper analytic techniques.
Real Time
Big data technology is capable of processing high volume of data in a very short time. Usage
data have an enormous potential to understand the shortcomings and needs of the
customers if monitored and understood correctly. Event-based analytics engines are proving
very useful in the offering the best campaigns and additional services to the customer at the
right time. The power of real-time campaign offers lies in catching the correct timeframe that
customer is in need. Customers are much more likely to accept the upsell or cross-selling
offers when their quota is about to end, or they tend to buy additional packages when they
use their service in a different location or from a different device. The real-time usage event
analysis is used to capture these moments to offer the best campaign for the customer. The
automation of this process is another benefit for the sake of reducing operational costs.
Conclusion
Customers decide on their threshold level for usage types. CRM System receives the selected
thresholds. Data is collected per usage type of the customer in the Analytics Big Data
Repository (ABDR). There will be real-time network optimization due to real-time detection of
dropped calls, error perturbation, sensors not reporting data, etc. There will be real-time
customer experience management with real-time dropped call resolution, or real time VIP
campaign monitoring, etc. With ABDR, last but not least, there will be real-time mobile spam
and fraud detection. Big data analytics paves the way for the real-time marketing based on
where the customers have been, where the customers are, and a prediction of future
behaviors to determine their needs; supplying real-time top-ups and dynamic charging, and
eventually predicting what they are purchasing next and knowing the ‘when’.
11. 8
About the Authors
Bengi Tozeren, MBA
Marketing Specialist II,
Business Development
Tel. +90 (212) 483 – 7101
bengi.tozeren@etiya.com
Rukiye Cetiner, MBA
Sr.Product Manager,
Product Development
Tel. +90 (212) 483 – 7101
rukiye.cetiner@etiya.com
Information on ABDR Catalyst Project:
Abdulkerim Mizrak: abdulkerim.mizrak@etiya.com
Rukiye Cetiner: rukiye.cetiner@etiya.com
Request a demo at sales@etiya.com
Published by Etiya
http://www.etiya.com
Yıldız Teknik Davutpasa Kampüsü Teknopark,
Cifte Havuzlar Mahallesi, Eski Londra Asfaltı Caddesi
151/1 B No:301 İstanbul, Turkey
Information
as
of
2015
The
written
statements
on
this
whitepaper
concerning
Etiya
solutions
and
projects
as
well
as
topics
around
big
data
are
forward-‐looking
statements,
which
involve
a
number
of
risks,
and
uncertainties
that
could
cause
actual
results
to
differ
materially
from
those
in
such
forward-‐
looking
statements.
The
risks
and
uncertainties
relating
to
these
statements
include,
but
are
not
limited
to,
risks
and
uncertainties
regarding
fluctuations
in
earnings,
intense
competition
in
IT
services
including
those
factors
which
may
affect
our
cost
advantage,
challenges
and
enablers
of
written
topics,
reduced
demand
for
technology
in
the
key
focus
areas,
disruptions
in
telecommunication
networks,
political
instability,
IT
security
and
information
governance,
legal
restrictions
on
ARPU,
OPEX
and
CAPEX
concepts,
and
unauthorized
use
of
our
intellectual
property
and
general
economic
conditions
affecting
Telecom
and
ICT
industries.