Speech Analytics: Have You Ever Heard a Smile?
Speech analytics is making sure that the loudest sound on a call is a smile.
In many spheres of human endeavor, from science to business to education to economic policy, good decisions depend on good measurement.
-- Ben Bernanke
"If you can't measure it, you can't manage it."
I am sure this isn't the first time you've read those words. They have been drummed into my head since the first day I encountered automatic call distribution (ACD) technology sometime in the early 1980s. Contact center (or "call center" as it was called back then) supervisors and business unit leaders lived and breathed time in queue, average hold time, and call abandon rate. Their jobs depended on keeping those and similar metrics in line and under control. A spike in one number would often lead to spikes in others, and ultimately, to disgruntled callers and lost revenue.
Like everything else, a lot has changed since I first entered the workforce and learned how to spell PBX. The number of metrics that a contact center gathers and measures has not only grown, but the emphasis has changed. Rather than putting all their eggs into the numbers games, customer care centers are concerned with identifying at-risk callers, agent retention, automation, trend analysis, and script compliance. It's not enough to know that a call was dropped; IT directors and business unit leaders want to know where the call dropped, why the call dropped, who the caller was, who the agent was, and the full context of the caller. In addition to measuring the technical data about a call (start time, stop time, hold time, etc.), a greater importance is now placed on the customer and agent interaction.
For example, speech analytics can be used to gauge the tone of the customer and agent's voice. Is the customer angry? Is the agent acting rudely? Is the agent using the prescribed words and phrases at the appropriate times? Is the customer using words that the agent isn't responding to?
By listening to the audio stream, a contact center can improve the customer's experience while driving down costs and driving up revenues. This can occur historically by analyzing recorded conversations or in real-time as the transactions are occurring.
The many benefits include:
- Improved sales
- Customer retention
- Improved marketing
- Increased brand awareness
- Compliance and regulatory adherence
- Enhanced agent training
- Streamlined escalations
- Improved agent behavior
Clearly, there's gold in them thar voices.Structuring Unstructured Data
As powerful as speech analytics might be, making sense out of hundreds, thousands, or even millions of conversations can be a daunting task. Voice can be extremely unstructured, and extracting the relevant data from the unimportant chatter is not a trivial task. It's easy to become lost in a sea of words.
Additionally, a speaker's tone can turn a statement upside down. For example, "Really?" can indicate happiness, anger, sarcasm, confusion, or disgust depending on how it's said and in what context. There are even cases where the sex or cultural background of the speaker can make a difference in how words and phrases should be interpreted.
Despite the challenges, many organizations have accepted speech analytics as an important business tool. According to the 2015 PWC State of Analytics Report:
- 90% of high performers say analytics is absolutely critical or very important to driving the company's overall business strategy and improving operational outcomes.
- High performers are 3x more likely than underperformers to be heavy analytics users, gleaning value via analytics in 10 or more disciplines. On average, high performers analyze more than 17 different kinds of data -- almost double the number analyzed by underperformers.
- By 2017, business leaders will invest more resources in data and analytics in these areas:
- Tools and technology (51%)
- People (35%)
- Training (35%)
While there are many ways to apply speech analytics, one of the most obvious places is within an organization's call recording platform, and vendors such as Calabrio, Verint, Nice, inContact, CSI, and TelStrat offer comprehensive solutions. Since call recorders are typically engaged with a call from cradle to grave, they are an ideal place to analyze voice dynamically or historically.
Outside the call recorder, companies such as Avaya provide solutions that perform analytics as calls are in progress. For example, the Avaya Real-Time Speech Snap-In can listen in on active calls and dynamically influence that call's behavior.To Infinity and Beyond
While using speech analytics to make present moment decisions is how most contact centers are operating today, a great deal of thought and effort are now being applied to predicting the future. By employing statistical and other mathematical models to in-progress and recorded conversations, businesses can predict everything from customer attrition to a looming spike in contact center traffic.
For instance, a food distributor can use a series of calls about tainted products to foresee a large-scale recall and proactively begin damage control measures. This may not have been possible if conversations were treated as singular events and not holistically. In other words, multiple calls about the same thing equates to a trend.
Speaking of trends, if you have been following my articles here on No Jitter, you will know that I am a firm believer that voice is being supplanted by text as the preferred method of customer-to-business interaction. This means that in addition to applying analytics to spoken conversations, the same technology must be applied to SMS text, Web chat, and email. If evaluating human speech for present moment and future decisions is important, it's just as important when those conversations are typed on a keyboard or thumbed on a touch screen.Have You Ever Heard a Smile?
I have seen a lot of motivational posters over my many years in the corporate world. While most were completely forgettable, some managed to stay with me. One of the most memorable was extremely simple, yet highly effective. Above the words, "Have you ever heard a smile?" was a very happy woman speaking on the telephone. I remember it because I have heard smiles. I also hope that people have heard my smiles, too.
In many ways, that sums up speech analytics. Contact centers want to know if their customers are hearing smiles and if those customers are smiling in return. They also want to hear when customers and agents are grimacing. They want to hear those things (and so much more) to understand the difference between success and failure. They want a tool that can help them streamline interactions and maximize profits.
I began this article saying that if you can't measure it, you can't manage it. However, that's only the beginning. You need to measure, evaluate, decipher, comprehend, predict, and take action. That's what speech analytics brings to the table. It's making sure that the loudest sound on a call is a smile.
Learn more about contact center trends and technologies at Enterprise Connect 2016, March 7 to 10, in Orlando, Fla. View the Contact Center track sessions; register now using the code NJPOST to receive $200 off the current conference price.
Andrew Prokop writes about all things unified communications on his popular blog, SIP Adventures.