A Drilldown on Data Analytics as a Key Contributor to Customer Analytics
The practice of customer interaction analytics seeks to improve the overall customer experience by mining intelligence from all customer interactions, and is a growing subject of interest to enterprises and vendors supplying them with customer service applications. Blair Pleasant recently pointed out in her piece entitled Are You Leveraging the Voice of the Customer Holistically—Across All Touch Points? that, “As an element of UC Analytics, Customer Interaction Analytics helps organizations identify and address sales, service, and operational issues based on analysis and intelligence derived from all customer touch points (voice, email, blogs, social media, etc.) across the enterprise.” In her article she discusses how Verint, with the company’s Impact 360 Customer Interaction Analytics, captures, analyzes and integrates customer data from all touch points, using different analytics technologies, including speech, data and text analytics, and then provides a unified perspective on what is going on with the customer. The outcome is that regardless of the channel, the resulting picture provides the customer with actionable data to improve the business. This wealth of data runs the gamut from understanding customer sentiment, to solving problems within or outside of the contact center. The beauty of it is that in many cases, this mining creates new ways of proactively solving customer issues before they become big issues.
Nowadays we so often hear that companies should provide multiple channels of interaction so that customers can do business with a company in any manner they choose, and we hear how the landscape for use of those channels is changing dynamically. We also hear that we should be harnessing all of those “customer touch points” to provide a comprehensive picture of what is going on with the customer. As part of this industry chatter we hear a lot about speech analytics, and more recently text analytics. We also know that there is a core set of customer data and contact center data to be analyzed as well. One component of Verint’s Impact 360 Suite - Data Analytics - works as the glue to tie all of these together. Impact 360 Data Analytics can mine the data associated with voice calls to reveal specific scenarios that can help or hurt overall performance. Data analytics makes use of structured analytics that is numeric in nature, and when used in combination with other UC analytics solutions such as speech analytics, which is more unstructured, and customer feedback, Impact 360 Data Analytics becomes a critical component of a multi-channel analytics strategy that can help surface the root cause of performance issues.
When I talk with companies such as Verint, I always want to hear about clear and memorable examples of how these applications have really helped customers. Without the clear examples, talk of combining data to get actionable results just sounds like so much marketecture. So I asked Verint for real customer tidbits as proof points that this stuff really works. I asked them for examples of how mining this combined data did uncover hidden service issues within the contact center and in the back office, and what the result was for the customer. And most importantly, I asked how did customers use this data to proactively improve the customer experience, reduce costs, or streamline business operations.
Memorable examples are what I got, including some that were funny, but sad from a business perspective. For example, Verint worked with a catalog retailer who decided to drill down, using data analytics, into average hold time (AHT) to see what the root causes were for very long and very short calls. While not dramatic, they found that the very long calls were all driven by one agent, who in essence was adding an average of a second onto the whole group’s calls. Root causes for the short calls, however, uncovered “a while the cat’s away the mice will play” attitude among agents. It seems that the short calls were happening during weekends when there was less supervision. Since the reps were being compensated by the number of call completions, when the supervisors weren’t around they were hanging up after a second, on many calls to increase their numbers. Similarly, the system also uncovered that there were large numbers of short calls in the evening as well. The company turned to speech analytics to find the reason and discovered that the shipping department shut down an hour earlier than the contact center, so the agents were telling customer to call back when shipping was open. The company extended the shipping department’s hours and the problem was solved.
Another example was of a contact center in the UK that was using auto calibration to score agents calls. The company used data analytics to view the entire supervisor’s scoring, and pulled out the outliers – the high and low scores – to see what anomalies they could find. What they found was that agent scores were lower on Mondays and higher on Fridays. It turns out that on Fridays the supervisor’s would go out for a pub lunch, causing their scores to go through the roof on Friday afternoons. So they quit agent call scoring on Friday afternoons and the problem was solved.
There has been a lot of focus on social media lately, and how mining social media channels as part of multichannel customer experience can help improve customer relationships and business processes. Whereas we believe that social media’s inclusion into an overall strategy is beginning to be and will be an important aspect of customer contact, we shouldn’t underestimate the power of the analytics of existing data and real-time contact center data. As Verint rightfully pointed out, it is what happens before a social media interaction, particularly if a customer has had a bad experience with a company that results in the social media interaction, and mining to gain intelligence to prevent bad customer sentiment is crucial.
As you can see from the examples above, sometimes human behavior within the contact center just gets in the way, yet can impact customer service immensely. Using the combination of analytics tools available to uncover simple changes that can be made goes a long way in improving base performance of the contact center.
This paper is sponsored by Verint.