Find Insights and New Opportunities with Unsupervised for Telco
January 13, 2021
Telecom companies are in a fierce battle for market share. Customers want support for the latest network technology and devices without making any sacrifices to quality or service. In the face of poor experiences, customers can easily churn to one of many competitors.
This forces them to understand their customers' needs today to maximize their experience. But at the same time they must prepare for the future, constantly upgrading technology and infrastructure to support new breakthroughs such as fiber optics, 4k streaming, or 5g to name a few.
When customers churn, should it really take months to find out why? Will it even be possible to save these accounts months from now?
Between customers, call centers, and complex network configurations, telecom companies collect massive amounts of data each day. In most cases, this information is siloed by department or division if analyzed at all. However, the true value and competitive edge buried in this data only emerges when blended and enriched across all sources and structures.
Telco leaders turn to Unsupervised as a new way to embrace their data complexity and find actionable insights to build a competitive advantage. Unsupervised goes beyond business intelligence dashboards and can adapt to your business and needs, surfacing insights from both structured and unstructured data sources.
Unsupervised is a shift in the status quo, allowing businesses to reallocate nearly 80% of time spent today on data management to analysis and tasks that lead to ROI while letting our AI handle the dirty work. Let’s take a look at how this works for a telecom company.
Finding Simple Action in Complexity
In this example, the company’s data includes customer data from their CRM, service coverage information, and their subscription packages. Unsupervised further enriches this data and builds new features such as ranges, unstructured text patterns, and more that are used to find and rank significant insights.
For this company, Unsupervised was able to enrich and analyze all of their provided data. One major finding was that customers that have active accounts for more than 4 years, but have not received a promotional offer in the last 12 months are twice as likely to churn. With this knowledge, the telecom company was able to act and offer the thousands of accounts falling into this bucket new promotional offers and substantially reduce month-over-month account churn as a result, and adjust their promotion strategy to mitigate this risk in the future.
The 4-year and 12-month figures were found by the AI. Nothing in the data suggested that these ranges are significant to tenure or promotions, respectively; however, because our platform is based on Unsupervised learning, it found these values by engineering and evaluating every possible range and element to include in the insight and surfaced the most significant result for human review.
To find such an insight in a BI dashboard, analysts manually test potentially millions of combinations of data elements, values, and ranges in order to find an insight like this one - taking weeks or in some cases months to find the equivalent of one pattern in Unsupervised.
Whenever your data updates, Unsupervised finds entirely new patterns to act on. BI requires reinventing the wheel to find new trends. What’s more, Unsupervised becomes even smarter and more curated to your needs, improving ranking and prioritization based on user behavior with each update to your data.
In telecom, time is always of the essence. With fierce competition, changes to technology and evolving customer needs, it’s critical that telecom companies invest in technology to monetize their data to build an advantage today so that the lights stay on tomorrow.