If you’re following the analytics market, you’ve probably heard analyst firms like Gartner talking about the increasing importance of new forms of analytics, that look beyond traditional structured data. One key area has been speech analytics.
But for many data professionals who don’t have a background in conversational or voice analytics, this raises the question. What exactly is speech analytics? Speech analytics is the process of analyzing recorded calls to gather information to improve business and employee performance. Insights from speech analytics can help a business shift from a traditional view of call center performance to a view that is centered on the most important person for your business: the customer.
Let me start by telling a story.
For almost 20 years of my professional career, I worked for a large US-based wireless telecommunications company. I held roles as an individual contributor and a leader in teams that spanned across the business in Finance, Sales Operations, Marketing, and Customer Service. In many of those roles, I was tasked with measuring and improving a key metric for the business:Call-In-Rate. Call-In-Rate is the number of customer calls made to the customer service team divided by the average number of active customers. Call-In-Rate was both a customer experience measure (does anyone really call customer service unless they’re having an issue?) and a cost containment measure, given it was very expensive to have an army of employees taking these calls every month.
For the final 4 years at this telecommunications company, I was asked to lead a data analytics team.
Our task? Answer a question that had always challenged the business:
Why do so many customers call every month?
The tens of millions of calls per year made Call-In-Rate a billion dollar problem for the business every year. To determine the root cause of these calls, I asked my new team a deceptively simple question. Do we know why customers call every month? I was seeking to determine the root cause of calls and was hoping that we had evolved to the point where we could confidently answer that question. The answer was a shaky “maybe”.
I knew the business had rolled out a technology called “speech analytics” a few years back and thought it would enable the team to answer that critical question. And the team was using the speech analytics platform - but only to find specific calls for the team to manually listen to and code the call reason in a spreadsheet.
The team was proud of its efforts, and rightfully so. They listened to 2,000 calls per month, manually categorizing the call reasons and reporting them, quarter over quarter, to leadership. But while listening to 2,000 12-minute calls may seem like a lot of work (it was), the sample size was not enough to create meaningful initiatives to address the overall number of calls. Those 2,000 calls were less than 1% of the total monthly volume handled by customer service - not nearly enough to convince our leaders that my team understood why customers called. There was no way to scale these to meaningful numbers without an army of people.
To tackle our billion-dollar problem, we had to transform the way we worked, starting with how we used speech analytics. Instead of finding and listening to calls, we put our team of call listening experts to work training our speech analytics tool to do the listening and categorization for us. We leveraged their expertise by building speech analytics rules that reflected what they would listen for on calls. Those rules were part of categories built across many call subject domains – billing, pricing, account maintenance, equipment, troubleshooting, and disconnects. Each domain reflected customer pain points that represented the reason or driver of the call. Instead of 2,000 manually categorized calls, speech analytics categorized over 3 million calls monthly. This explosion of data created a gold mine of potential insights. We just needed to evolve to make sense of it all.
I love spreadsheets. I used to tell everyone that I opened one application every day for the first 18 years of my career. That app was Microsoft Excel. My new team, responsible for tackling our billion-dollar problem, loved Excel too, but it doesn’t handle 3 million calls very well. We needed a modern analytics platform that could handle large datasets to mine our data for insights.
Once we combined our speech analytics data with a modern analytics platform, the real magic began. We could provide business leaders with valuable insights to make better decisions and drive tangible business results. Instead of a quarterly Excel report, we had daily analytical dashboards that told a critical and influential story to hundreds of business leaders. We were exponentially more confident in why our customers called us because we had exponentially more data. We developed a better understanding of our business and knew where to allocate resources to improve the customer experience, reducing the need to call customer service.
The result was a 10% call reduction in each of the four years I was in my role. Put another way - we saved over $500 million dollars for the business.
Speech analytics can be a vital data engine for your business. It can take something unstructured (a voice recording) and produce something that can be analyzed (large amounts of customer-specific data) to help you make better decisions to drive better results.
Every customer interaction counts toward the success of your business, and with speech analytics, you can answer these all important questions:
- Why are customers calling?
- What is being discussed on each call?
- How do customers feel when they call?
- How are employees handling these calls?
- Are customers considering leaving for a competitor?
- What products and services are being discussed on calls?
- How was the voice quality on calls?
- Did our employee say the right things to our customers?
- What topics of discussion on calls leads to higher sales conversion or lower repeat calls?
Answering these questions can lead the way to business process improvements, greater sales and revenue, cost savings, and higher levels of customer loyalty.
Since my time in telecommunications, analytical platforms have continued to evolve to utilize artificial intelligence and machine learning to increase the value companies can derive from speech analytics. Platforms, like VoiceBase, make it easier than ever to implement enterprise-level speech analytics and get a full 360-degree view of customer interactions while analytical platforms, like ThoughtSpot, make it easy to mine these very large speech analytic datasets at the most granular level for valuable insights. Pairing these platforms together can help companies quickly unlock insights within their critical customer data and empower business users to deliver outstanding customer experience & improved business results.
At Axis Group, we help organizations overcome data challenges to reach their insights faster. Faster insights mean better decisions and better results. If my story sounds familiar to you and you would like to explore more about the challenges and successes I've had, I can be reached at [email protected] or you can learn more about our solution on the ThoughtSpot Atlas marketplace here.
Steve Vaughan is a Master Solutions Consultant with Axis Group.