Challenges for Telecoms• To find the right strategy to increase the Average Revenue Per User - ARPU• To build the brand on higher level than the competition• To decrease the customer churn rate• To form efficient and consistence customer experience through different and complex customer channels.The operators need to convert data into business intelligence and connect the measureswith the objectives. This will help to anticipate customer’s requirements, improveresponsiveness, and provide more personalized level of service. Page 2
What is Crystal Qube™? Customer Experience management platform that enables telecom operators to have a complete profile of their customers and complete record of their past and current behavior. Closely connected with a smart deployment of business intelligence it helps to: •Understand the customer preferences •Predict the future behavior •Turn the customers’ insight into actions Outcome: Improved CX and loyalty, reducedchurn and increased ARPU and profitability Page 3
Where does Crystal Qube™ fit in?The Customer Experience Analytics deals with all aspects of customers’ behavior, as it enablesanalyzing past, measure present and forecast future trends. Customer analytics
Does CEM matter?In-depth interviews with senior executives within 20 service providers around the world weredone by TM Forum . The survey goal was to investigate how CSP approach customer experiencemanagement and how they are using or plan to use customer analytical applications
Our approach• Based on pipes and filters architectural pattern• Pipeline represents customer lifecycle• For every lifecycle phase the pipeline has different filters• Filters are based on various customer metric (customer value, churn probability, tariff simulation results, market segment belonging, predicted basket, etc.)• The output from the filters are customer profiles that are eligible for certain campaign• The campaigns are executed for such customers
Crystal Qube™ functionalitiesOur platform offers following integrated functionalities:• Various reporting on time, geography, offerings, segments, profitability, etc.• Prepaid and postpaid customer churn management• Postpaid customer risk scoring• Customer segmentation• Profitability & referral value analytics• Cross & up sell analytics• Campaign management• Tariff Simulation Page 8
Customer behavior• Based on customer profile data + mining model outputs• Forecasting bills, payments, refill and spending behavior for every customer• Detect significant change between actual and forecasted values• Record customer interaction with customer care (trouble tickets, inquires)• Perform bill checks• Perform health checks• Perform bill shocks detection Customer Profile Associated Market Segment FAX Postal Email Telephone Customer Facing Services Contacts Resource Involvment Individual Business Interactions Party Profile Party Customer Account RelationshipsOrganization Forecasted Profile Customer Customer Account Churn Stats Credit Profile Predicted Basket Customer Bills Customer Bill Payments Customer Refills Tariff Scores
Churn management• Use of customer analytics and data mining, to build predictive model based on customers’ past behavior (assign numerical/measurable score to individual customers.)• Address issues successfully to gain more loyal customers that can be prevented from churning.• Combine Churn Probability with CLV or CRV to get better results.• Forecast the churn rate in the next 6 – 12 months.
Churn management approach• Churn rate is measured from the history records• Churn probability is measured using decision tree/neural network mining models• Future churn rate is forecasted using time series data mining model. The forecast is done for the whole customer base and for every segment defined in the system.• The churn model is with more than 70% accuracy- the results are automaticly compared on daily basis with actual customer status. If the model performance drops below certain treshold, it is updated with new variables/rules.
Customer segmentation• Segment the customer base using different indicators like churn probability, risk score, CLV1..3, CRV, tariff, customer type, geography info, etc.• Track, analyze and forecast segment growth.• Target segments with personalized cross and up sale campaigns.
Profitability & referral value analytics• Find out which customers are best for your profit and worth for targeted retention efforts• Define the customers that have the biggest impact on other customers and prospects
Cross & Up sell analysis• Offer recommendation, basket analysis, next best action, real time offer management etc.• Make effective up sell /cross sale campaigns• Find offering candidates for bundles• Simulate potential basket and target the customers with the most probable offering they would buy.
Campaign management Design and execute campaigns using Iworkflow Campaing Components. Additional supported options:• Out of the box connectivity for USSD, SMS and IVR centres and email servers• Ability to integrate with Print House for sending personalized offers together with monthly bills• Select preferd customer channel based on customer usage preferences (VOICE, SMS, DATA, Email) and respone to previous campaigns. • Black listing of various customer segments (VIP, Big Business, State Employees, Police, etc.) • Spam Filter to set rules such as max numbers of offer sent to a customer in certain period • Use Facebook component to publish campaigns and surveys on company profiles • Support of provisioning and charging component that allows service activation and charging in time of offer acceptance.
Tariff simulation• Tariff models and options defined on internal product catalog• Simulate operators and competitors tariff model and options• Run simulations on single customer profile or on group (segment) level• Run simulation on the fly for single customer or during the night for whole customer base Product Offering Relationships Product Specification Product Offering Competitor Voice Price Zones Prod. Offering Fee Sms Price Zones Composite Prod. Offering Price Recurring Fee Mms price Zones Allowance Price Alter. Data Price Product Offering Price
Managed eco-system• Crystal Qube™ is managed service and runs in private cloud• Crystal Qube™ is secure – all data is encrypted during transfer• Crystal Qube™ is isolated, there are no two Crystal Qube™ platforms on the same HW. Page 18
System architecture• Information delivery – Portal with dashboards, simulators and power reports• Information & Process integration – iWorkflow server and designer• Data store and data mining – Microsoft SQL Server 2008, data model is based on SID. Portal (Information Delivery) Interactive Forms Information & Process Integration Dashboards Report iWorkflow iWorkflow Security & Parts Parts Audit Ad-hoc Tariff Server Designer query tool analysis Report Distr. Dashboards Browser iWorkflow Reports Data store, Data Mining and Data analysis Risk Scoring Provisioning Workflows Tran. Data CLV/ARPU/ RRPU cacl. Messaging simulation Behaviour Postpaid Prepaid change Queue Queue Churn Churn Tariff MBA ....... Worflows Definitions Worflows Configuration CEM.DM CEM.DB SID iWorkflow.DB Page 19
What to expect from Crystal Qube?• 360° powerful view of the customers’ profile and behavior• Easy to use and intuitive interface• Focused campaigns -identifying and targeting the right customers that will lead to maximized marketing ROI (better hit/return rate)• Operational cost reduction by enhanced customer care performance as problems are seen in real-time• Option to identify new revenue opportunities• Process automation - consistent, integrated, and appropriately modeled data even on daily basis to support the user’s reporting and analytic requirements• Managed environment