For close to the past nine years, Charlie Thomas has led Razorsight, the producer of cloud computing-based big data and predictive analytics tools for telecommunication firms, as the Reston, Va.-based firm’s CEO.
Over that time, Thomas has grown Razorsight into an international firm with offices in Tokyo; Bangalore, India; and Singapore, in addition to its Virginia and St. Louis locations.
Thomas currently serves on the board of advisors for a number of technology firms – including GuidePoint Security, CFN Services and Base Technologies – and has co-founded and sold four companies over his career, including Net2000 Communications.
We recently asked Thomas a question:
How is Razorsight leveraging Data Science to help Communications Operators better serve their customers and manage their businesses?
He penned the following as his response:
There are 12 billion communications provider connections across the globe today serving over 3 billion distinct end users (consumer and business). The connections come in the form of Cable TV, Satellite, Mobile Devices (smartphones, tablets), Broadband and Wireline.
In meetings with nearly 200 Executives from different communications providers around the world over the last year, we’ve seen a common set of challenges:
- Market saturation
- Due to extensive penetration, most operators (except in developing nations) are not seeing subscriber or revenue growth
- This saturation results in margin compression for network operators
- Saturation and price compression result in a race for the best customers. Any hiccup in service or customer sat, or a proactive campaign from a competitor, could well result in customer churn.
- Revenue Flat
- Market saturation, churn and competition from Over-the-top (OTT) competitors with Internet cost of delivery economics results in flattening revenue and declining ARPU for network operators
- Increasing Capex & Opex
- The market need to support new technologies (4G/LTE) increases Capital and Operating expense.
Razorsight delivers new micro-segmented levels of insight to communications providers through advanced statistical analysis or predictive analytics to better pinpoint customer sentiment, potential churn, upsell opportunities, prioritize capital and overall to predict what will happen in the future.
Best of all, with Razorsight there’s no software to buy as these predictive insights are delivered via a private cloud.
For example, at a Tier 1, leading global wireless operator, we’re leveraging Razorsight’s predictive analytics platform to operationalize better programs for cross-selling and upselling their tablet (iPad, Galaxy) customers.
So, someone may have an iPad, but may not have activated the data plan for it. The provider may be able to up sell them a family plan for data usage that’s tied to their smartphone and tablet. They may also be able to bring family members onto the plan who may not be under that operator.
We’re leveraging our data scientists and our cloud analytics apps to help them improve in that area.
In another instance, we’re helping a Tier 1 global wireless operator understand how a discounted or subsidized price for a tablet or a particular device would impact pricing across their other solutions and products (elasticity).
Again we’re leveraging our predictive analytics to understand what the take rate might be on certain price plans and how that could impact pricing elsewhere, either up or down.
Another area where we’re experiencing significant growth is with customers, such as a prepaid wireless operator, that have very high churn rates in the 7 to 8 percent range. We help these operators reduce churn which translates into tens-of-millions of dollars a year.
Yet another example is with a leading MSO (Cable operator) helping reduce truck rolls by leveraging our vast amount of data on network QoS and calls to customer care to predict customer issues and network troubles in advance so they can proactively address issues and alert customers in advance – this results in great customer sat and higher retention rates as well as lower maintenance costs.
We analyze 50,000 different variables with a complex data model. We’re able to operationalize and give them very specific insights into which nodes are most likely to have trouble or fail next and which customers would be impacted.
If they had to make a repeat visit on one of the truck rolls, we help them understand why – the cause and effect – and point out where its most likely to occur again based on a set of correlated factors and demographics that we’ve ascertained through pouring through extraordinary amounts of data.
Those are some real-world examples of things we’re doing with the predictive analytics platform we have for various operators. We’re doing this for a marquee list of customers and the value is extraordinary. The value for their top line is significant and the return is fast – less than 6 months.