Using Big Data and

fearlessquickMobile - Wireless

Dec 12, 2013 (3 years and 6 months ago)

44 views

© 2012


PROPRIETARY AND CONFIDENTIAL INFORMATION OF CVIDYA

Using Big Data and
Analytics Insight to
Improve Customers'
Retention and Value


Gadi

Dr. Gadi Solotorevsky

CTO


cVidya Networks

Ambassador, Distinguished Fellow


TM Forum

2

A leading supplier of Revenue Analytics solutions to
communications and digital service providers

Founded:
2001

300

employees

in
15
locations
worldwide

Deployed
at
7

out of the
10

largest operators
in the
world

150

customers in

64
countries


Globally
processing
150
Billion
xDRs

per day,
55 Trillion
xDRs

per year

Contributing over
$12 Billion
to
providers annual
revenue

Partnering

with
world
leading vendors


What
You Should Know
ABOUT US

Africa

APAC

CALA

North America

East & Central Europe

West Europe

64
Countries

64
Countries

150
Customers Worldwide

3

USA


AT&T
, Sprint, Alaska Telecom,
Sovernet
, ATN,
Commnet
, Alltel,
Islandcom
,
CellularOne
, Choice

Canada

Bell Canada


South Africa


MTN, Vodacom

Uganda


Uganda Telecom

Zimbabwe


TelOne

Ivory Cost


MTN

Namibia


MTC, Telecom Namibia

Angola


Angola Telecom

Russia


RosTelecom
,
Uralsvyazinform

Ukraine


Kyivstar

Belarus

-

BelTelecom

Georgia


Magticom

Kazakhstan


KazakhTelecom

Turkey
-

Vodafone

Greece


Cosmote
, HOL (Hellas
Online
),
Forthnet

Albania


AMC (
Cosmote
), Plus

Romania


Cosmote

Serbia


Telekom
Serbia, Vip mobile

Slovakia


Orange Slovakia

Bosnia Herzegovina


Telecom
Srpske

Croatia

-

Croatian
Telecom T
-
HT

Bulgaria


Mobiltel

Poland



T
-
Mobile

Macedonia



Vip operator

Slovenia

-

Simobil

UK


BT,
O2, Orange
, T
-
Mobile
, Colt,
Three, Sky,
VirginMedia
, Cable &
Wireless

Germany


DT,
Vodafone, E
-
Plus
, 1und1

France


France Telecom,
Orange

Spain


Telefonica
,
Telefonica

Moviles
,
JazzTel

Italy
-

Telecom Italia ,
TeleTu

Switzerland


Swisscom

Austria


A1 (Telecom
Austria Group)

Sweden


Tele2,
Telia
,
Telenor

Denmark
-

Telia

Netherlands


KPN,
Telfort
,
Orange,
Vodafone

Ireland


O2

Lithuania


TEO,
Omnitel

Finland


DNA

Portugal


PT,
ZON

Liechtenstein



Mobilkom

Belgium

-

BICS

Brazil
-

Vivo,
TeleSP
, TIM,
Embratel
,
CTBC

Mexico

-

Movistar
/
Telefonica
,
Iusacell

Argentina

-

Movistar
/
Telefonica
, Telecom
Argentina

Peru

-

Movistar
/
Telefonica

Chile

-

Movistar
/
Telefonica

Ecuador



Movistar
/
Telefonica
, CNT

Colombia

-

Movistar
/
Telefonica

Venezuela

-

CANTV,
Digitel
,
Movistar
/
Telefonica

Guatemala

-

Movistar
/
Telefonica
,
Telgua

El Salvador
-

Movistar
/
Telefonica
, Claro

Panama

-

Movistar
/
Telefonica
, Cable & Wireless

Nicaragua

-

Movistar
/
Telefonica
, Claro

Costa Rica


ICE

Caribbean’s


Cable & Wireless,
Digicel

Puerto Rico


PRTC

Guyana
-

GT&T

Uruguay


Antel

India



Vodafone,
Airtel

Vietnam



MobiFone

Philippines



Bayan
, Globe

Australia



Three (H
3
G), Optus

Thailand



CAT

Singapore



MobileOne

Macau
-

CTM

Israel



Bezeq
,
Bezeq

International,

Cellcom
,
Orange,
Pelephone
,

YES
, HOT

Palestine


PalTel

Fiji


Digicel


Papua New Guinea



Digicel


Sri Lanka
-

Mobitel

4

What data is available to Telco’s

Usage Data
-

Events (CDRs, xDRs)

Location Data

DPI data

Customer information


(e.g. age, address, services, organization, id, email…)

Customer interaction information (payments history, complains history, churn)

Customer social information (who he calls, who calls him, when, duration.…)

Data from external sources


Facebook, tweeter, web...

Competitors plans




5

What can be done with this data

Revenue Assurance (detect and prevent Revenue Leakages)

Design
price plans & Best offers

Detect Fraud

Increase customer satisfaction and prevent churn

Cross and pre sale

Enable targeted advertising

And much more, plan traffic routes, prevent infections propagation,……



6

First Example


Revenue
Assurance

7

TM Forum Revenue Assurance KPIs Study
-

2011

98%

2
%

Detected Leakage % of
revenues

Revenues
Leakage detected
31%

69%

Recovered %

Recovered
Unrecovered
43
%

57
%

Recoverable and Unrecoverable
% of un
-
recovered revenues

Recoverable
Unrecoverable
8

IMS and
LTE Network

Internet

Evolved Packet

Core (EPC)

e
-
NodeB

e
-
NodeB

e
-
NodeB

e
-
NodeB

Radio Access
Network (RAN)

S
-
GW

MME

P
-
GW

HSS

PCEF

PCRF

SPR

ePDG

PDF

CSCF

AS

MGW

PSTN/PLMN

IMS

MGCF

OFCG

OCG

Wholesale

I/C Billing

system

CRM

Postpaid

Billing

system

Provisioning events

Accountable usage events

ERP

IP
Network

9

The Challenge


Daily input of ~12 billion records in various formats


Find discrepancies between sets of ~40 billion records


No exact matching (clocks differences, partial information, rounding
differences, similar but not exact values)


Need to detect also best match


Cost effective


Robust, and capable of manage significant delays in inputs


After discrepancies are detected, need to enable near
-
real time
analytical investigation


Naïve sort and match will not work

10

The Solution

Mediation

Events

Comparison
Engine

KPI, KRI





OALP

Events and
transactions
files Multiple
formats

Fragmented
info

Formatting

Enrichment

Correlation

Purging

Drill
down/trough

Records

Discrepancies

Aggregations

Records

Indexes
to files

Aggregations

DB

Events

Files

Second Example


Design and offer
Price Plans

12

Next
Best
Action

Unstructured

Advisor





Next Best Offer




Impact


Propensity to Accept


ARPU Impact Influence

Advanced Insights

Impact

CMOs need flexible
analytics
solution to
model multiple new
plans
and benchmark versus
competitors’ offering


Control
customers’ migration
and potential
income
dilution


Consider behavioral changes due
to price plan
change

Advisor


CMOs need “
Next Best Offer”
solution
to provide
personalized,
accurate, value
-
driven, pricing plan
recommendation per individual
based on operator’s financial
analysis


Based on individual historical
usage data enable the operator
to better understand the “future
invoice” per each price plan

Propensity to
Accept


Predicts
per customer


the
propensity to buy each of the
offers based on his usage and
profile

ARPU Impact Influence


Predict
for each
individual
the
ARPU change in case of
accepting an offer to buy each
of the
offer


Models are built by learning
how did ARPU of similar
customers have changed over
time after accepting each of
the offers


Advanced Insights


CMO want to utilize
clickstreams data and voice
transactions
to cross
-
sale and
up
-
sell their own as well as
3
rd

parties offering (e.g
. shows
tickets, loyalty programs, etc
.)

13

The Challenge


The goal
, increase customers’ satisfaction, reduce churn, and increase
revenues


The method:
Design new price plans, estimate their impact, and
propose next best offer to
customers


The volumes:
54 Billion events (30 Million subscribers, 10 events per
subscriber per day,
h
istory of 6 months)


The requirements:
cost effective (e.g. hardware costs), analyze multiple
price plans in parallel, offer the right plan to the right customer

Using a regular billing system will not do the trick

14

cVidya
Solution Overview

Price Plans

from billing

New Price
Plans by
Marketing

Competitors
Price Plans

Results database

Filter results based
on post rating
business targets
definitions

Priority definitions
for eligible options:
best,
2
nd best, …

Selection of price
plans, options and
customers for rating

Filter eligible rate
plans based on pre
rating parameters

Rating and bill
calculations

Pricing plan implementation

Process Audit & control

Price plan implementation

Scenario & Rating

definitions

Business rules

definitions

Reports, logs and
alerts

Extract eligible

subscribers data

Exports to CRM

& Campaigns

Rated
CDRs

Reports, detailed
results for each
subscriber

Web interface

Customer Data

Invoices Data

Business rules
simulations

Financial impact
simulations

15

Transform

Rating

Engine

Price plans

System Storage

Proposition
engine

Invoice

Customer


Details

CDRs


CDRs

Information Flow

Subscribers
Profile

Data repository

Campaign

Management

CRM

WEB

Self care

Rating

Engine

Rating

Engine

Rating

Engine

Rating

Engine

Pricing
Optimization

16

The result Advanced Insights


Cross
-
sale and
up
-
sale

You will save
10


by moving from
unlimited to
2
GB bundle

You will save
10


by moving from unlimited to
2
GB bundle. Register to this package and get
1
month of free access to “shape
-
up” gym club
near your office

You will save
10


by moving from unlimited to
2
GB bundle. Register to this package and get
3
months of free access to live sport streams at
premium speed

Website

Blog


www.ra
-
blog.org


www.cvidya.com

Your


Success


is Our


Business