Conversation Analytics and the future of analyzing phone calls for customer experience
“Speech recognition technology has been around for what seems like forever,” says Kevin Mann, CallRail Co-Founder & CTO. “But it’s been pretty rudimentary until fairly recently.”
Mann notes that even though the tech has been around for two decades, it wasn’t until the past few years — when large companies like IBM’s Watson started investing time and money in it — that it really started to make serious progress.
Technology behind call tracking analytics
Conversation analytics involves machine learning and natural language processing (NLP) technology that evaluates the voice data from phone calls (who’s talking, what he or she is saying, what keywords have been said, etc.) and makes conclusions based on programmed analysis. Literally, analyzing a conversation.
Natural language processing is key to conversation analytics. According to Matt Kiser of Algorithmia, “NLP is a way for computers to analyze, understand, and derive meaning from human language in a smart and useful way. By utilizing NLP, developers can organize and structure knowledge to perform tasks such as automatic summarization, translation, named entity recognition, relationship extraction, sentiment analysis, speech recognition, and topic segmentation.”
Machine learning maintains and advances NLP
As CallRail Product Manager Adam Hofman explains, the machine learning element of conversation analytics means that the technology is constantly acquiring knowledge: “NLP and machine learning do a good job of recognizing patterns that appear in audio very often. Agents that continually take phone calls are helping train the conversation technology to become more accurate and precise.”
Tools like CallRail’s Conversation Analytics use this advanced NLP technology to ‘listen’ to calls, transcribe the conversation, determine who’s speaking (the caller vs the agent), spot and tag specific phrases along with frequently noticed keywords, qualify a call as a lead or not, and provide generally greater insights into calls.
Mann suggests that we’ve arrived at an inflection point, where companies like CallRail can take raw technology and adapt it to solve the business problems that digital marketers are facing every day. For the call analytics industry, he expects this will bring big changes to the way companies do business: “Conversation analytics allows people to analyze phone calls in a way they never could before. It used to require a user to take a human to come in and figure out the subject of phone calls and hence the outcomes.”
Conversation analytics and call tracking software
With the latest call tracking technology in your marketing stack, tools like conversation intelligence and NLP can give your businesses a quick idea of lead quality, sales team performance, customer loyalty and experience, and more. The AI behind call intelligence analyzes the cues of a call and automatically determines the quality of a lead automatically — even when dealing with hundreds of thousands of calls.
This tech can process multiple elements of a call, all in real-time: The length of a call, the number of turns taken by each speaker, the presence of important industry-specific keywords (such as the words and phrases that are spoken most often by good leads), and more.
Using calls to improve marketing campaigns
In the CallRail App, our conversation analytics tools include automatic lead classification along with a transcript to help users understand customer issues, sentiments, and experience. If you notice that a lot of your calls have the same theme, you’ll be able to address it and improve the customer experience.
With conversation intelligence, you can also determine what marketing campaigns are bringing in the most qualified leads. By associating each campaign with a tracking number, you can get detailed information on how well each campaign is performing across all channels (online form fills, offline calls, etc.).
Phone calls are a minefield of data, and often the missing point in attribution for many businesses. With the detailed information that calls provide, companies can make their marketing even more effective.
The future of conversation analytics
When asked how he thinks the technology will continue to evolve, Mann believes that sentiment and context analysis are the next frontiers for NLP and machine learning engines.
“I think the next steps for this technology will mean it’ll be able to determine if callers are satisfied or not, discover based on context whether it’s a sales or support call, and converge data points on individual calls to tell a comprehensive picture of the caller and business journey.”
Hofman agrees: “As the industry grows and people discover new ways to tweak algorithms, technology advances to better analyze the audio signals. We might be able to tell a computer, ‘These signals indicate that a person is satisfied or dissatisfied.’”
Tools like Conversation Analytics in CallRail’s call tracking software allow marketers to truly understand the data and analytics behind the human voice.