How does speech analytics perform trend analysis? Isn’t it just counting how often words are used?


Trend analysis in interaction (speech and text) analytics solutions provides deeper insights than word count frequency. At its most basic, trend analysis is designed to allow users to define “things” they know to look for in voice and text-based interactions, which are then surfaced by the application. However, in recent years, more advanced solutions have started to include forensic capabilities that enable the application to self-identify (or discover) trends and concepts that a company didn’t know to ask about, which might otherwise go undetected. For example, if a competitor’s name or marketing program starts to show up in findings, the solution would automatically identify the new trend and notify system users. To provide this enhanced level of insight, interaction analytics (IA) vendors have begun adding more artificial intelligence technologies, particularly machine learning capabilities, into their IA solutions. 

IA vendors have also begun trying to assist their customers in understanding the meaning of trends. Initially, they did this by delivering new and more advanced reports and dashboards that show the relationship between words and phrases, associations between trends (words and keywords), heat maps that reflect the importance of issues, and much more. While displaying these results has increased the usefulness of the solutions, leading IA vendors have started to take the next step, turning their offerings into business intelligence tools to enable customers to apply the findings and realize a better return on their investment.