Verizon partners with Intent HQ to improve sector-leading churn prediction accuracy by up to 3.5%.
Challenge
Verizon already had a top-performing churn model, refined over two decades, but saw an opportunity for improvement by leveraging behavioral data and Intent HQ’s AI platform.
Results
The combination provided significantly more accurate churn predictions, which will improve Verizon’s ability to find and retain at-risk customers.
Accuracy improved by 3.5% in the highest-risk decile
Churn model accuracy improved by 2.1% in the top 3 deciles
Forecasted value over five years: $180 million
Verizon partners with Intent HQ to improve sector-leading churn prediction accuracy by up to 3.5%.
“I challenged the team to become ‘customer first’ by delivering exquisitely personalized services and experiences.”
Ronan Dunne
VP and CEO of Verizon’s Consumer Group
Operationalizing Churn Improvement
Verizon leads the U.S. wireless telecom market because it constantly innovates to maintain its competitive advantage and drive brand preference.
For a company generating annual revenues in excess of $100 billion and serving over 100 million customers with a range of voice, data, and video services, delivering better operational performance can be incredibly challenging.
But Verizon’s sustained success at the top of the U.S. wireless telecom market is due to their relentless drive to innovate and improve along all fronts, including customer acquisition, base management, and customer retention.
Telcos are at an inflection point. For years, they’ve faced mounting pressure from better-funded, more innovative tech competitors, and seen a return on investment, profits and market share slowly dwindle.
This is the dream: you put vast troves of data to work, accurately predicting each customer’s wants and desires, creating individually tailored marketing, sales, and service. You solve problems before they become problems. You operate like a nimble and personalised small business, but at massive scale. Your customers love you. Your competitors wither away.
Pucket is a Scala library which provides a simple partitioning system for Parquet. But what is Parquet and why does it need partitioning when it already supports filtering? In this post I will attempt to explain Parquet, partitioning in Hadoop, and the motivation and design of Pucket. If you’re not interested in the background, you can skip straight to some simple code examples or go to the GitHub repository.
By no means all of the regulatory action under GDPR has been accomplished through fines on organizations. It is expected that the EU legal consensus regarding privacy will strongly influence their behavior. Since GDPR’s introduction, most regulators have taken a consultative stance, giving advice and looking for incremental improvements in behavior.