Now let's turn our attention to the telecommunications industry. This involves perhaps not as many different types of businesses, but they include mobile operators, cable companies, satellite companies, separate device, infrastructure, and retail companies, and some kinds of conglomerates of all of the above. The kinds of data and analytics that they're most interested in tend to be those related to capacity, utilization and demand, pricing, network optimization, supplier management, regulatory issues, customer service, and customer fraud. The kinds of analytics that we'll often see revolve around understanding the potential of new product offerings, improving customer experience, improving service or reducing truck rolls, forecasting network capacity and demand more accurately and faster, implementing value-based network capacity planning, and of course reducing customer churn. Speaking of churn, this is exactly what XO communications wanted to do. So, they analyze 500 discrete data elements including support call patterns, late or delinquent payments, and other ongoing vital signs via predictive analytics solution. This helped them identify at-risk customers and tailor educational programs that address known sources of customer concern at an early stage before they morph into full-blown dissatisfaction. The 4-month implementation resulted in a 47% reduction in customer churn, protecting $15 million in revenue. It improved overall customer satisfaction and goodwill, and the predictive analytics concept has now spread organically to other parts of the company, including collections. The major telecommunications company in Pakistan, Mobilink was looking to utilize data and analytics for building customer trust, improving loyalty, boosting margins and decreasing churn. They built predictive analytics models that included clustering techniques and social media analytics to help them gain insights of customer behavior. This enabled them to achieve an 800 percent increase in uptake of customer retention offers, up from 0.5% annually to about 4% now, and at a fraction of the cost. They were able to also boost campaign response rates by 380%, thanks to social network analytics, and it now takes them less than a day to deploy new analytic predictive models. Globe Telecom also wanted to improve the performance of promotion offerings. They created a new platform that integrated customer and retailer data with real-time promotional analytics and incentive payments, achieving a 600% increase in promotional sales, a 95% reduction in time and cost of delivering new promotions, the ability to drive revenue from hundreds of simultaneous targeted promotions, and increased market share via improved customer experience and effective campaigns. Similarly, COX communications wanted to customize offers more quickly and accurately for more than 6-million subscribers and to double direct-mail campaign conversions. They deployed a solution for predictive analytics including segmentation, classification, regression, and data aggregation. The central analytics team streamlined market analysis covering all 28 regions. They now have a precise, accurate, and fast polling of 10 million observations in 800 variables to identify customer-related issues, including the propensity to purchase or churn. They've achieved an 80% reduction in model creation time and are able to generate models 42 times greater, creating 1,680 predictive models annually up from 40 previously.