8/12/2019 090113 Ananto Whitepaper Telecom v3 1
1/13
Ananto
HARNESSING BIG DATA ANALYTICS FOR TELECOM
January 2013
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
2/13
Ananto
1. BackgroundAccording to a recent Cisco forecast, global mobile data traffic is estimated to grow at a
compounded annual growth rate (CAGR) of 78 percent, reaching 10.8 exabytes per month.
All of this being carried by telecom players across the globe is forcing them to mitigate
threat to voice based business plans they had with improved data handling capabilities and
associated charging mechanisms. While being carriers of data, Telcos have mostly focused
on their CRM and usage patterns to derive customer behavior and profiles. The challenging
business environment that they face today, presents a classic Michael Porters five force
model. Threats from new entrants, substitution of voice based business models with data,
strong bargaining power of customers, a supplier base that associates as vendors rather
than partners and a strong intensity of competitive rivalry are all creating extreme survival
needs and declining bottom-line.
In these times, reversal to basics of business is most important. Basic fundamentals of
increased association and relevance to consumers, excellent quality of service, quick and
easy customer care, accurate and measured financial planning and innovative solutions for
tomorrow are the need of the hour.
According to a recent survey by The Economist Intelligence Unit:
a) Companies that use data-directed decision-making enjoy a 5-6% boost in
productivity.
b) 53% of companies leverage only half of their valuable data, and one-fourth of
respondents noted that vast quantities of useful data go untapped.
We, at Ananto, believe that data has the ability to throw at us significant actionable insights.
If the framing of business issue to be tackled is right and data is being collected from
multiple sources and analyzed with rigor, valuable insights about associations and
behaviors can be easily unearthed. However this increases the data volume, data types and
velocity of generation poses impediments that traditional methods of analysis are unable to
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
3/13
Ananto
keep pace with in the current business environment. Telcos today can no longer afford to
ignore the next competitive advantage i.e. consumer should buy more, consumer should beretained for longer and consumer should get his family and friends to buy the same brand.
Dynamic decision ability with real-time actionable analytics helps improve and stream-line
tactical processes while keeping in mind the holistic goal of optimal customer experience.
2. Challenges of data explosionService providers have immense access to data when it comes to subscribers, their usage
transactions, network performance data, cell-site information, device level information as
well as back office and consumer interaction logs. However the data resides in various
forms i.e Structured (already in the IT systems like CRM, Sales Force, and Billing etc),
Unstructured (voice calls, videos, social media, email, and chats), Static and Dynamic.
Traditional approaches of offline analysis or using business intelligence with siloed
datamarts cannot keep pace with the variety and velocity of data that operators have to
deal with today. Increase in data loading time, querying time and hardware shortage, willbe major hindrances when they are struggling to remain profitable and remain preferred in
the competitive eco-system.
Fig 1: DATA Explosion
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
4/13
Ananto
In order to deal with the above challenges and at the same time deliver new, compelling,
revenue generating, customer satisfying service without overloading networks and withoutcosts running out of control, Telcos need to find innovative ways to run their process with
fine-grained service control. This can be achieved by the intelligent use of Analytics,
particularly Big Data Analytics. Analytics not only provides insight about existing behavior
of business but can also deliver models that can predict future outcomes far more
accurately. We believe Telcos current business environment does not leave them with any
room for error and hence there is a need to adopt sophisticated analytics capability that
can unearth valuable correlations. These insights, coupled with business acumen of theleaders in the trade are the future differentiating strategy for any telecom player today.
3. Using Analytics to gain valuable insightsIt is indeed a challenge to blend multiform voluminous data from a variety of sources when
it is being generated at a great speed. With the help of brushing and mining certain basic
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
5/13
Ananto
patterns along with low-latency solutions, enormous value and return on investment can
be delivered and that is what enterprises require. Businesses must be able to access,analyze, and act on business information faster than ever.
In order to deliver new compelling, enhanced revenue generating, cost effective customer
experience without network congestion and huge cost implication, Analytics services using
Big Data is the new way to harness efficiency and effectiveness along with real time
actionable intelligence.
While all business environments are different, we make an attempt to help you with someof the known areas of improvement using big data analytics:
3.1. Segmentation and Subscriber profilingSegmented customers based on the current and potential value so as to enable
customized value based customer treatments and strategy.
Fig 3: Customer Segmentation
The first step here is to ensure that one is able to get data about the customer from
various sources. Billing, CRM, Usage etc are more commonly known but social media,
location based information, interactions with customers service etc are some other
sources gaining prominence.
Micro segmentation is a more advanced form of segmentation that groups small
numbers of customers into extremely precise segments, based on various factors,
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
6/13
Ananto
including behavioral predictions. Marketers can then direct specific marketing actions
and campaigns to each micro-segment (down to the individual consumer level) tomaximize the effectiveness of every contact with each customer.
Segmentation is a continuous activity, using advanced cluster analysis approaches, a
pattern about Prized Customers are continuously obtained and applied for predictive
analytics. Campaign results both in terms of its effectiveness and its impact on customer
experience are evaluated, helping modelers with a list of impacting variables, thereby
improving predictive outcomes.
Fig 4: Micro Segmentation
A recent experience explains the gap between what a customer needs vs. what a Telco is
able to deliver. An existing mobile user with heavily used multiple connections wanted
to get another connection. The CRM system of provider was unable to see him as an
existing user and thereby started treating his request for a new connection as a fresh
buyer. This results in a paltry credit limit vs. established credit limit of significant
amount due to other connections. The customer is left with extra work to manage hisnew connection and extra bill and would surely result in extra work for providers
marketing department to figure our complete value of this customer. Such processes
cause error in segmentations, resulting in irate customers and cost to provider in
understanding his customer profiles.
http://www.optimove.com/predictive-behavior-modeling.aspxhttp://www.optimove.com/predictive-behavior-modeling.aspx8/12/2019 090113 Ananto Whitepaper Telecom v3 1
7/13
Ananto
3.2. Customer Retention & Experience Enhancement:
3.2.1. Churn Prediction:a. Future voluntary churners so as to facilitate proactive retention
efforts yielding better retention performance.
b. Future non-payers and bad debtors so as to customize the dunning
policies and collection efforts for different risk groups and enable
early retention actions.
Fig 5: Actionable Analytics with Churn Prediction
3.2.2. Trend & Pattern Detection along with Identification ofa. Major recharge patterns in the prepaid base and segmenting customers
based on the dominant recharge behavior so as to enable customized
recharge stimulation actions.
b. Multi SIM users and customers using primarily competitor lines and
enabling share of wallet acquisition actions.
c. Household customers and missing products to enable mobile, fixed and
multi-play penetration opportunities
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
8/13
Ananto
3.2.3. Propensity modelTo prioritize, cross sell and up sell offers to customers based on the
acceptance probability and expected revenue increase, eventually predicting
the next best offer (BNPO) for the given customer based on their call usage &
behavior pattern.
3.2.4. Tariff OptimizationIdentification of potential spending of subscribers and competitor tariffs so
as to calculate the best tariff for each customer to be used for retention and
test performance & relevancy of tariffs. Analyzing the price elasticity of
subscribers and calculating the impact of price changes and financials should
identify tariff.
3.2.5. Experience measureMeasuring experience from contact center information, unstructured voice
of consumers from social media and structured survey data helps in
Identification of inherent yet prudent Brand imagery and subscribers Share
of Mind (SOM).
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
9/13
Ananto
Fig 6: Advanced Analytics & predictive modeling
3.3. Proactive Network MonitoringThe existing transactional system detects critical outage when it occurs. They were
rarely deliberate to analyze network performance over a long period of time to
understand how and where service issues are trending, how that affects most profitable
customers and ultimately effect customer retention. Network lifecycle events have a
large influence on customer satisfaction, as there are many moving parts that will affect
customer experience. The residual effect of suboptimal network performance will result
in longer contact center calls, higher customer support costs and unsatisfactory
customer experience. Optimized and integrated information can enhance the customer
experience.
The efficient network monitoring solution adds data from network monitoring tools
with CDRs in real time, giving an updated picture of network operations every four to
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
10/13
Ananto
five minutes. The network operations center can quickly see issues as they arise and
proactively resolve them, before customers start calling. Network quality problems canbe identified in minutes or seconds rather than by the next day. The statistical
analysis would assist in the prioritization of network investments and improvements by
calculating the potential value adds and customer experience impact of network in each
location.
3.4. Buy More Strategy:
Commonly referred to as cross-sell and up-sell opportunities, it is by far becoming the
most irritating interaction for the customer. The callers or the front desk has little or no
knowledge of the consumer profile, causing a dump of possible options, hoping to hit
the bulls eye. From carpet bombing to intelligent and customized offers makes a world
of difference to customer experience. Ananto believes that being relevant and aligned is
most important in this effort to increase consumer lifetime value (CLTV). It should
appear to the customer that he is being serviced and not being sold to. Data insights and
predictive modeling techniques can provide significant value to make this process adelight for both customers and providers. One of the most used example is an old
Harvard study which shows a correlation between a single father buying beer when he
comes to buy diapers. Intuitively it sounds right when one understands the need to take
care of the baby and watch a game along with beer, but it took many years of basket
analysis to identify this correlation. To make such folks buy more, beer is normally
kept on the way from the diaper section to the checkout counters.
3.5. Revenue leakage management:CDR data amounts to multiple terabytes in a day, yet this data is the most important
input to the revenue management of a service provider. By analyzing such data on the
fly, patterns of missed/incomplete records can be identified and corrected in near real
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
11/13
Ananto
time basis. This requires sophisticated modeling and continuous updates and results in
millions of dollars of saved revenue. By doing it on the fly, it reduces the cost of storage.Big data techniques are very apt in managing such volume and velocity of data. Such
techniques help the operator maintain an accurate tracking of service usage. This
accuracy is crucial in minimizing revenue leakages, especially at inter-carrier exchange
points.
3.6. Google Business model:
This concept is in its infancy, yet we find it relevant to be mentioned, considering the
stressed business plans most of the Telcos have today. Google mined its search service
to generate behavior of the users that it could intelligently offer as differentiated
advertising strategy to business. Can Telcos mine the data flowing in their pipes to
offer intelligence of consumption behavior, propensity to purchase, ability to use to
potential business as qualified target groups? This is not merely a service that Telcos
should provide but it needs a bold step to change the business plan completely. Just like
in Googles case it should not appear as though Telcos are trying to sell information orlists, rather it should be a comprehensive approach using mobile devices as means to
channel intelligent offers.
In summary, these are few uses of analytics in managing business issues for a provider.
There are number of other specific uses that we have seen, for example, deployment of
mobile towers, evaluating front line staff, predicting future traffic patterns due to a specific
event, evaluating low revenue yet profitable customers, reducing cost of outbound
campaigns and cluster based modeling.
4. The Next StepAnalytics is an art that uses technology as its tools. Use of platforms and technology is to
make the job easier and accurate when done repeatedly. The most important resources are
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
12/13
Ananto
the people who can see patterns in data. While the science, software and computing power
has increased tremendously, it still needs a fine eye to detect possible correlations, unlikelypatterns, outliers and important insights. Its the people and platform combination that
provides one with proactive actionable intelligence.
Essence of building and implementation strategy lies in Five C framework of Ananto.
i Curiosity Ask the right question? Leaders in near future will not have to worry
about providing the right answers but about asking the right question that can help
them gather correct insight.ii Capture Capture data from multiple sources to create an encompassing image. The
variety of data flows and the velocity of data flow are important considerations in
building right capture strategy. This function needs to be automated to manage
velocity coming in from multiple sources.
iii Curate Data, when compiled through multiple sources needs careful treatment. It
needs to be put in similar plane, errors need to be corrected, and it needs to be
converted from an unstructured format to a common structure.
iv Crunch At this stage, business issue needs to define the shortlisting of data set that
needs to be processed. Sophisticated tools, various permutations by data scientists
and model generation and validation are some common tasks in this phase.
v Create Visualization Ananto believes that insights from analytics are worthless if
leaders in Telco are unable to play with it and then take decisions. Visualization is a
very important aspect of all the hard work done in earlier phases. It not only
displays information gathered but also has ability to drill down and ability to run
what if scenarios. All this on the finger tip of these leaders, on a device of their
choice.
Ananto also recommends a crawl-walk-run strategy in implementation. It is not
important to start investing in significant infrastructure and people. We have seen
8/12/2019 090113 Ananto Whitepaper Telecom v3 1
13/13
Ananto
relatively low hanging fruits in most business cases where some specific problem can be
solved for credible business gains. Since the big bang approach requires significantchanges to working business processes, it usually faces significant change management
challenges and stakeholder buy-ins. By using the crawl-walk-run strategy, where the
investment is small initially, there is natural buy in from critical business owners and a
clear impetus to invest more dollars.
Ananto is your partner in this journey from data to dollars.
About Ananto
Ananto is a Big Data Analytics Solutions Company, supporting firms with Data Insight,
Predictive Intelligence, Technology Integration and Advisory and Professional Services.
The word Ananto means Infinite, depicting the very nature of endless possibilities of
business solutions that can be achieved using effect analytics.
Our mantra is to convert your data to dollars for you. Data itself has no value, unless it can
be harnessed smartly and transformed into deeper customer insight and channelized
towards enhancement in productivity and efficiency. This deep insight into data can help
organizations to Re-focus, Re-align, Re-engineer and be Relevant and Resilient.
Started by customer experience industry veterans with global experience in technology,process management, consulting and business solutions. Aparup Sengupta, in his last role
was the CEO and MD for a $1bn business services organization while Anil Modi was the
CMO for the same organization.