Making VoC Text Analytics Count: From Directional to Diagnostic

For a number of years in the Customer Experience Management (CEM) space, we’ve relied on text analytics to make the first steps in turning unstructured, but highly rich customer verbatim comments into measurable packets of insight that organizations can use to gain never-before-seen insights into the expectations and perceptions of their customer base.

Back when I first set out in this industry and the VoC market was a little less mature, it was sufficient to be able to leverage text analytics to categorize customer comments into pre-defined buckets to understand what a customer is referring to. Example – customer says “The agent was great but your price plans suck” – analytics engine says Staff Conduct & Pricing. Taking this use case and extrapolating across hundreds of thousands, even millions of pieces of feedback every year and you’re left with a very credible barometer as to what’s on your customers’ minds. Tactically you can reward and recognize staff, recover customers based on the analytics output and strategically  aggregate to assess underlying trends and spikes. That is how it was.


Yet from conversations I have with customers every day there is no doubt that the market is maturing rapidly. Organizations across verticals and geographies are taking a more comprehensive, enterprise-wide view of Voice of the Customer (VoC) – a response to the growing importance of customer centricity on the executive’s priority list. Consequently, the required analytics to manage this shift are fundamentally different. We’re no longer just talking about analyzing direct customer verbatims, we’re talking about mining and operationalizing social feeds and incorporating indirect voice, email and chat channels into the VoC hub, to name just two.

This plethora of direct, indirect and inferred VoC feeds requires analytics to be both scalable and comprehensive. With regards to scalability, as the ongoing collection and analysis of VoC feeds moves towards big data proportions, we need to have the infrastructure to be able to maintain the speed and slickness of real-time analytics. At NICE, this is a given. More interestingly, the role of text analytics is moving from the purely directional analytics of yesteryear to the diagnostic analytics required for next generation Voice of the Customer solutions. It is no longer enough to be able to direct business functions as to where there might be an issue uncovered by VoC – we need to be able to leverage powerful analytics to drill down to the second and third level of detail to hit that holy grail – root cause.

Herein lies the shift from directional to diagnostic; the ability to not just identify what issues are most commonly referenced, but the ability to generate automatic clusters to track common themes within those categories and sub-categories. With the ongoing integration of NICE Fizzback’s pioneering Natural Language Processing engine with NICE’s leading ‘hot topics’ and ‘root cause’ analytics – we can begin to appreciate the power and potential of VoC analytics in helping organizations leverage the Voice of the Customer to successfully transform their business.

Andrew Robson

For more info on NICE Fizzback please click here.