The Old QA Methodology is Dead

“This call may be monitored for quality assurance purposes.” When I was young, I thought this meant that any prank calls would result in a callback and a stern warning. The truth is, if a company is still using QA methods from my childhood chances are that nobody will be listening.  This is starting to change, however, as more and more companies leave sampling behind and begin to adopt speech analytics technology as a way to automatically monitor and score 100% of their calls.

CallMiner recently worked with a large financial services client who had been scoring calls in a very traditional fashion; by outsourcing manual agent quality monitoring to an outside vendor.  Let’s take a look at some of the reasons why this was proving ineffective, costly, and why it should serve as the eulogy for traditional call center QA practices:

1. The number of calls monitored each month was statistically insignificant and resulted in a lack of confidence in the scores. 

Traditional sampling only covers about 5% of all calls that come into a call center, making it virtually impossible for managers to get a true sense of why customers are calling or how each individual agent is actually performing. If the sampled calls included customer escalations, how could managers evaluate the seriousness of the problem?  In addition, because each agent is being evaluated on such a small data set, it becomes difficult to truly measure their skills.  Performance feedback may be misguided or completely off the mark, which defeats the entire purpose and deteriorates morale.

2. Manual reviews of calls were inconsistent and prone to error, causing additional oversight. 

 With human review, there are always bound to be errors. Perhaps the QA analyst evaluates the call too quickly, misunderstands sarcasm on the part of the caller, or overlooks a critical insight. Different QA analysts may have different opinions on what really happened on a call. As a result, agents receive inconsistent feedback and coaching opportunities may be missed.

3. Manual review of calls was expensive and not scalable.

For a contact center that has hundreds or thousands of agents, manual call reviews are simply not a cost-effective or scalable QA methodology. As your agent pool grows, the costs of QA quickly spiral out of control. As a result, you’re left with gaps in coverage or less than perfect monitoring. 
 

4. Monthly reporting was time consuming.

It’s always interesting to see how companies aggregate reports with a manual QA system. Most of the time, it consists of a cumbersome spreadsheet that requires days of effort to update.  Some analyst groups end up spending the majority of their time wrestling with reports. This isn’t the best use of critical QA resources.

Our financial services client realized that they needed a new approach to QA monitoring. By moving to a speech analytics solution from CallMiner, they are now automatically monitoring and scoring 100% of agent interactions. The results:

• Automatic categorization of calls based on the tagging of contacts that contain certain language patterns, keywords, phrases, or other characteristics has shown to be more accurate than manual contact categorization. The data is now being used with confidence for decision-making across the entire organization.
  
• An expanded monitoring program that now covers all agent teams at no additional costs.
  
• Improved call selection, making it easier for managers to find calls where agents use, or don’t use, the appropriate language. This information is then used to ensure compliance and focus coaching sessions.
  
• Identified instances when agents were not speaking clearly; something that manual QA monitoring had overlooked. This provided a valuable coaching opportunity to help improve customer experience.
  
• Immediate analysis of results from coaching. Call center managers could see improvements in agent performances’ directly in the system within days of coaching sessions.
  
• Automated, customized reporting that significantly reduced monthly report generation efforts.

With the continued development of speech analytics software, it’s clear that traditional QA methods like outsourcing are costly and outdated.  Leading companies are now using speech analytics to automatically monitor and score 100% of their contacts, providing agents and managers with much more timely and accurate performance feedback.