Real-time Decisioning for Big Data
0 likes1,075 views
Openet's document discusses the evolution of real-time decision-making for big data, emphasizing the importance of leveraging service and customer data to drive revenue and reduce costs. It highlights the need for real-time analytics techniques for personalized services, particularly in audience measurement for video content across various platforms. The document also addresses the challenges associated with this data management and the potential benefits of improved customer engagement, satisfaction, and loyalty.
Real-time Decisioning for Big Data
- 1. Real-time Decisioning for Big Data: Evolving the Service with Audience Measurement Presented by Marc Price, CTO Americas July 2014
- 2. 2w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Openet delivers the infrastructure to enable Operator Innovation A global leader with more than 80 customers in 32 countries, Openet provides insight, interaction, control, and monetization within the world's largest and most complex networks. Openet helps control costs and increase revenue by enabling customers to innovate how people, machines and services engage with their network Openet helps increase revenue and control costs by enabling customers to innovate how people, machines, and services engage with networks Cable Mobile Machines People ServicesFixed Real-time transaction management
- 3. 3w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only • Service Data and Customer Data can be leveraged in unprecedented ways to drive revenues, increase loyalty, and reduce costs • New techniques for big data require real-time processing for low latency analytics, as well as pre-processing for batch analytics • An example of valuable analytics for service usage is audience measurement for video viewing across various platforms • New techniques enable streaming event processing, in-memory No-SQL storage, real-time decision triggers, and holistic data delivery for personalized services, recommendations, analytics, dashboards, and beyond Highlights
- 4. 4w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only The Digital Communication Age The volume and variability of immediately available offerings make the communication decision-making environment quite distinct from other billable services. The Challenge is to maximise every interaction Experience Interaction Types: • Ad hoc • Occasional • Informative • Social • Regional • National • International • Short term • Regular • Excessive Interaction Types: • Information • Entertainment • Purchasing • Business • Social Media • Messaging • Browsing • Posting content • Gaming • Tweeting • Communication
- 5. 5w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Growth of Personal Data This richness of big data provides an enormous potential to personalize services - yet customers still marvel at how little their service provider seem to know them The volume of personal data is growing exponentially & at a staggering rate to the actual number of subscribers Per Subscriber Personal Data My Plan My usage My spend My Contacts My devices My fav apps I’ve been a customer for years… I never receive any relevant offers… I want more interactive, informative contact
- 6. 6w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only 1998 Prepaid 2000 GPRS Unleash the power of real time personalization Despite the vast amount of customer data that exists, marketing continues to rely on traditional segmentation schemas and static profile attributes that provide an incomplete view of the customer • Traditional BI systems, data warehouses and CRM systems are ill-equipped to support dynamic customer profiling • Operators know they are not reaching certain groups of people, and may not even identify a “category” before it is too late • Legacy systems are not designed to deal with the volume, variety, velocity and volatility of big data to deliver value, relevance and immediate responses to market needs 2005 3G 2015 HetNets 2013 LTE 2014 IMS 1996 2G/ GSM Patchwork of legacy infrastructure supporting the evolved technology & services. Optimizing subscriber profit margins is now a fine art with rising infrastructure OPEX costs 2007 HSDPA 2010 WiFI OPEX
- 7. 7w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Seizing the Moment Analytics insights require useful and actionable information through timely actions to process data when it is most relevant Tier 1 (first order Analytics) Tier 2 (second order Analytics) Tier N (nth order Analytics) Tier 1 (first order Analytics) Location A Location B Window of Insight Window of Insight Window of Insight ADS Triggers Batch ie. Instructing Policy Use Cases, Recommendations Insights Insights Insights Status Scenario Elligibility Condition A Condition B Condition C Condition X <15 milliseconds 1 second 15 seconds For a particular subsriber: Continuous stream of data ADS: Advertising Decision Server ie. Ad Decisions ie. Lifecycle analytics
- 8. 8w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Operators are seeking Reporting and Analytics solutions in response to new business drivers Delivering value, relevance and immediate response • Intelligent Upsell use cases: creating more relevant, timely and personalized services Real Time Decisioning • Real time behavior visibility & detection of network service impacts and billing impacts Real Time Service Assurance • Enable content provider sponsored data, Relevant advertising, B2B revenue chains Sponsored Data and Advanced Advertising • Understanding viewing impressions and related applications usage to aid with marketing, upsell and content negotiations Audience Measurement Some Examples:
- 9. 9w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Detailed video viewing impressions and contextual usage statistics better support marketing and advertising campaigns as well as content negotiations Case Study: Audience Measurement Audience measurement incorporates analytics for: • Linear television program viewing • Video on demand viewership • Ad campaign viewership • Interactive content engagement
- 10. 10w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Detailed video viewing impressions and contextual usage statistics better support marketing and advertising campaigns as well as content negotiations Audience Measurement: Challenges Challenges include: • Normalization across traditional cable systems and IPTV and other platforms • Correlating usage for “Second screen” and multiple devices with linear viewing • Household vs. Individual user data • Protecting Personally Identifiable Information (PII) • Supporting Data Governance techniques
- 11. 11w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Enables operators to generate subscriber profile and segmentation with unprecedented depth and accuracy Audience Measurement: Highlights Usage information is validated, normalized across platforms, then enriched with marketing & subscriber information handling opt-in/out for behavioral viewing metrics. The solution collects and correlates second-by-second click stream, linear TV viewing, on-demand and interactive TV events. Delivers profile information across systems in a useful, anonymous format to internal and third-party marketers, advertisers and buyers to enable the planning and measurement of addressable advertising campaigns and marketing promotions
- 12. 12w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Audience Measurement: Architecture Data Sources: ActIngest Analyze Ad campaigns HSD Events Wireless Data Events Linear TV Tuning Events VOD/PPV Events iTV Events DVR Events Ad Inserts TVE Events Filtering Parsing Validation Translation Aggregation Correlation Enrichment Multi structured data for streaming and batch analytics Data Collection & Processing Content rights negotiation Marketing campaigns Recommendations Real Time Insight Data Analyzed in motion Storage
- 13. 13w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Measure and influence customer behavior • Measure: • Geographic/Demographic Trends • Screen/Device Usage trends • Ad Viewing Trends • Content Viewing Trends • Content Ratings • Influence: • Ad Placements and pricing and therefore Ad Revenue • Content Pricing driving content revenues • Targeted Offers driving improved service uptake and revenues • Recommendations driving increased usage revenue Big Data techniques applied to Video Viewership enable service providers to:
- 14. 14w w w . o p e n e t . c o m © Copyright 2013 Openet – Company Confidential For Use Under Non-Disclosure Only Conclusions • Improved Customer Satisfaction • Boost in Loyalty • Reduced Costs New techniques for managing service usage data enable: