How data science can help your company withstand the pandemic

Pandemic effects on business

Take, for example, our client ITYNS global (name anonymized). It is a U.S. Fortune 500 telecommunication company with multiple lines of business in broadcasting, entertainment, cable, internet, advertising and devices.

Imagine that you are the head of customer support and care in the U.S. You have six U.S.-based call centers and five international call centers. COVID-19 has hit ITYNS's broadcasting and advertising revenue drastically. Businesses are holding tight to their spend budget.

Due to the lockdown, your U.S. call center teams have shrunk to 70% of their original size, utilizing workers that could be transitioned to work from home. The other 30% are still on payroll but unable to work from home. Thus, they are off while the government-issued lockdown is in place. Your IT and operations teams are working day and night to provide more of these workers with remote setup and reliable internet, but there are major delays in devices as every other organization is in the same boat as yours.

You get an average of 10 million calls per month, but COVID-19 has created a 30% spike. With fewer call center workers available, you are routing more and more calls to an international and virtual location. However, the call handle time has increased dramatically with the double whammy of decreased agents and increased calls. The company's customer satisfaction score (CSAT) is dropping and the non-payment rates are increasing.

You had several long-term plans already in the making, including moving calls to chat and text and serving them via AI bots, improving interactive voice response, and moving certain calls to lower-cost vendor partners. However, none of those are functional enough to make a complete switch.

Hence, you need a stratified plan to handle the greatest number of calls most effectively with the most benefit to metrics that matter. What is the fastest way to learn about the biggest opportunities for shifting different types of calls? The answer is data.

Using data science to solve immediate problems

I heard an analogy recently and it's quite apt here: If you know you will be working in China shortly, wouldn't you prepare by learning some Chinese? Similarly, if you know you will be using data for your business, wouldn't you want to learn the language of data and business? Simply put, you need to know data science.

Taking the above-mentioned example of ITNYS, this would mean using a hypothesis-based approach to quickly analyze calls, CSAT and other call and customer-related metrics -- including Customer lifetime value (CLV) -- to come up with a Pareto chart of your call routing strategy. This can minimize call handle time and maximize CSAT, even with fewer agents. If you do it well, you can discover many key insights fairly quickly. These insights are data-based facts, as opposed to intuition, allowing you to rightly ascertain the issues affecting you.

For example, you may learn that one of the top reasons for the call spike is the customer's inability to pay in the current month. Having found the reason, you can now work toward tackling it. You could implement an autoresponder with customized messaging of payment extension dates based on a customer's CLV and income bracket. You have now remedied 60% of the spike in call volume.

However, that is a short-term solution for an immediate, acute problem. You used data science, and you reduced the impact.

Building a culture of data for long-term opportunity

Now imagine that it is September 2020 and workers have returned to their physical desks. But nothing is back to normal. The pandemic that resulted in huge global losses has eaten your margins and your business is no longer able to afford the inefficiencies it had accommodated for years or decades.

To survive and thrive in this new world, your company must become lean and mean, relying on data science. Your organization and all its decisions must be data optimized or you risk losing to competitors who know how to exploit data. But you can't manage this situation as you did with high call volumes during the initial COVID-19 response. You don't have enough data scientists to optimize every decision made in your organization and you must incorporate a long-term strategy and goals. Your entire organization must become data-driven by adopting the four Ds of a data culture.

In the current global economic situation, it is more urgent than ever that you develop a culture of data. The four Ds will serve you well as your guide.

1^st D: Data literacy

Data literacy is the ability to think about and use data appropriately toward making optimized decisions. It is not one-size-fits-all; the data enthusiasts in your organization need a different level of data literacy than citizen analysts or data scientists.

This means you need well-defined skills and capability mapping for every individual or user persona in your organization (perhaps mapped to job and function level). Additionally, you will need to create a learning path to help them reach their objectives.

Once the company's individual decision makers are data literate, they then need easy access to a single source of truth to derive insights and inform decisions.

2^nd D: Data maturity

An organization with a data maturity score of 7 or greater (on a scale of 0 to 10) typically has well-defined data sources with appropriate access levels by user persona. For example, in the case of ITNYS, a data enthusiast might have access to customer support dashboards via a self-service BI application such as Tableau or Microsoft Power BI, while a data scientist would access the same data at the transaction level via a Python connector to a data lake. The data would be accurate and mean the same thing whether accessed via the BI tool or directly through the lake.

3^rd D: Data-driven leadership

Data-driven leaders focus on building a culture of data-driven thinking and using data as a key asset for their organizations. They make significant investments to enable data maturity and to invest in data literacy -- both with time and money. They allocate resources to data during the budgeting process and hold their respective teams accountable for reinforcing data literacy. Essentially, building a culture of data starts with leadership.

4^th D: Data-driven decision-making process

A data-driven decision-making process ensures a systematic way of making decisions with full transparency and a systematic way of looking back, evaluating and learning from those decisions to improve future ones.

A culture of data is essential to doing business today. It will help your company reduce inefficiencies, take advantage of opportunities, and respond positively to economic challenges. But before you start, you must know that building a data culture is not a three-to-five-year prolonged project. Smart businesses can make a great deal of progress in less than a year and be quite data-driven by this time the next year.

The time to plan is now. Never before has your team been in this mindset of transition and open to possibilities. You have a precious opportunity for massive skill building -- don't let it go to waste.

About the author
A highly-regarded industry thought leader in data analytics, Piyanka Jain is an internationally acclaimed best-selling author and a frequent keynote speaker on using data-driven decision-making for competitive advantage at both corporate leadership summits as well as business conferences. At Aryng, she leads her SWAT data science team to solve complex business problems, develop enterprise-wide data literacy, and deliver rapid ROI using machine learning, deep learning and AI. Her client list includes companies such as Google, Box, Here, Applied Materials, Abbott Labs and GE. As a highly regarded industry thought leader in data science, she writes for publications including Forbes, Harvard Business Review and InsideHR.