How to Spot Bad CX in Your Contact Center

by Alissa Pagels-Minor

Speech Analytics tools can be used to evaluate all of the calls going in and out of a call center for important customer experience (CX) insights. If you’ve ever called a 1-800 number, you probably know how a great experience during that call can affect your opinion of the brand and whether you will continue doing business with them. Better CX is proven to be linked to overall profit, so modern organizations look for ways to track and optimize CX across the entire enterprise. Research shows that increasing customer retention by a mere 5% can result in a profit gain of 25%-95%!

55% of shoppers in the US said that one bad experience would keep them from returning to a brand, meaning aligning customer experience with operational goals is essential to a brand’s survival.  So what key metrics can executives be looking for in calls to measure customer experience?


When examining call recordings, it’s important to note significant changes in sentiment, rather than just one overall sentiment score. Is the change on the caller’s side? Use this metric to track overall customer emotion, and events within the call. Change in sentiment on the agent’s side from positive to negative? This might indicate a challenging interaction that could require coaching.

Speech analytics solutions, like VoiceBase, can also pinpoint 100% of calls in which customer sentiment was under a certain score, providing critical data to call center management, and key stakeholders.


Too much silence on the agent’s side could mean the agent hasn’t been trained enough on that particular topic or situation, leading to a less than amazing customer experience.

If an agent needs some extra coaching to reach a certain knowledge competency, you can spot those opportunities with silence-aware speech analytics technologies, like VoiceBase. Tracking silence ratios is an important metric to add context to calls.


Customer calls are a great source for organic product feedback. Is there a glitch with one of your products? Is something not working as well as expected? By surfacing correlations between your products and negative statements, you can uncover issues at scale. For instance, if 50% of your callers are complaining about a, keyword spotting can surface that trend and alert you to maintenance needs. 

Mining recorded call data is a great way to get authentic, unfiltered product feedback before your surveys alert you to a widespread issue. Using a BI tool that has word cloud ability such as Tableau, you can group topics of conversation and find themes across all calls. 


Call centers with great customer experience will have better First Call Resolution Rates, or FCRs. Generally, the more a customer has to call, the higher the operational cost and the worse the customer experience. This is a great indication of how successful your agents are at solving customer problems, but the insights usually end there. 

With Speech Analytics like VoiceBase, you can monitor the reason that the issue was resolved, leading to greater insight about your customers and the challenges they face, and how the best agents address them. You can then use that data to coach your team, leading to improved CX across the board.


Post-call surveys are a common tactic to determine customer satisfaction scores, but getting survey participation is a battle. Responses that are captured tend to be very polarized or unreliable. Pro Tip: use a speech analytics AI model to extrapolate NPS scores from 100% of your calls.  Building a model to track overall scores can provide a baseline and ongoing indications of overall customer experience. You can even set up triggers to notify you of any significant changes – which could indicate a major issue and alert you before it is too late.

Continually improving your customers’ experience is a challenging task, but worth the effort! If you’d like a free demo of how VoiceBase can help you optimize your CX with speech analytics, get in touch.

More From the Voice analytics blog

Predictive Analytics for Strategic Insights

Predictive Analytics for Strategic Insights

Predictive analytics is an advanced form of data mining that leverages machine learning to identify patterns in voice recordings, intuit a speaker’s intent, and predict a future outcome — be it a sale, account cancellation, or one of many customized “X” signals your clients might request.

read more