Original Publication: Customer Management Insight - September 2008
View Online
It seems that far too many customers have experienced the following scenario: They call a company’s call center and discover that it’s been outsourced overseas. While there’s nothing inherently wrong with this — and the agent is certainly professional and helpful — the customer simply cannot decipher the agent’s accent. The language barrier proves too high, and the conversation starts going in circles. Frustration builds to the point where the customer hangs up and redials, hoping for a “better” agent. As they wait on hold, fuming, they wonder if it would be easier to just take their business elsewhere.
As more and more businesses race overseas to capitalize on lower salaries and operating costs, an unfortunate consequence is that these types of situations are becoming alarmingly frequent. And the negative results are typically undiagnosed and not well understood. While the economics of outsourcing contact centers is undoubtedly appealing, many businesses later discover the downsides. As agent turnover rates continue to soar higher and higher, businesses find themselves scrambling to recruit and hire the best agents. In short, they soon discover that finding the best offshore candidates is a time-intensive, costly and imprecise process. What’s more, those tasked with recruiting and hiring decisions are often ill equipped or too burdened by their workloads to accurately assess even the most basic of qualifications. One such basic — yet highly sensitive — qualification is an agent’s ability to fluently and articulately speak the language of their customers (see sidebar).
To address these challenges, leading-edge contact centers are turning to speech recognition and analytics technology to make smarter, faster and more consistent hiring decisions. While speech analytics and recognition tools in the call center are typically associated with generating business intelligence, the same underlying technology is now being applied to help reverse the hidden costs of manual, inefficient hiring processes: longer call times, increased transfers and repeat callers, lower service levels and lost customers.
Automating the Agent Language Evaluation Process
Forward-looking centers are turning to speech analytics tools to make smarter and faster hiring decisions by automating the most labor-intensive and imprecise parts of the process. At first glance, automating the hiring process seems like a radical notion. But rather than taking human judgment completely out of the equation, speech analytic applications can help contact centers quickly boil down and sift through a large pool of applicants by first objectively measuring and scoring a candidate’s language fluency and articulateness. Leveraging the same underlying speech analytic technology used in contact centers for mining customer and business intelligence, these new hiring assessment tools enable hiring managers to automate a large chunk of the upfront assessment processes by comparing a candidate’s recorded reading of a script to the phonetic standard of the specified language, measuring the quality of pronunciation and fluency in a matter of minutes. Using these tools to do the heavy lifting of analyzing, scoring and ranking the recorded audio against a predefined standard, managers can save countless man-hours each month by automating preliminary evaluations of hundreds or even thousands of applicants.
So how does something like this actually work in practice? In the case of one speech analytic vendor, contact center hiring managers first develop a script to include common language as well as words and phrases that are known to be problematic based on the applicant’s native language, including key words and phrases specific to a business vocabulary. Next, applicants register for a language evaluation through a web browser and complete the assessment on the computer; applicants read the scripted paragraphs of text and their responses are recorded. In a matter of minutes, the tool scores each applicant (or existing agent when used for training purposes) on pronunciation abilities, fluency of speech and pace of conversation.
A score is then produced for each individual based on a combination of measured skills, and results are consolidated and reported to users who can use the rankings to quickly narrow down a small pool of candidates with the highest language aptitude for the next round of evaluation.
In addition to leveraging the innovative solution to the hiring process, contact centers can also leverage the tool to better align an agent’s language skills with customer needs. For example, agents with less fluency can be assigned to more routine and easier customer calls, while more complicated customer issues could be routed to agents with higher language proficiency.
Case Study: Streamlining and Accelerating Hiring
Global ISP, a global Internet service provider, is a great example of how companies can leverage speech analytics in the real world to address the persistent challenges of quickly, cost-effectively and consistently hiring the top contact center agents in the age of global outsourcing. The company outsourced 100% of its customer care functions through a partner network of outsourced call centers. Their ability to provide a fast, consistent language skills assessment for onboarding new agents and evaluating existing ones was critical to maintaining required agent staffing levels without sacrificing quality of service. The company dedicated four full-time employees to manually assess candidates’ language skills, in addition to relying on the judgments of their various outsourced partners’ own hiring managers — a laborious and subjective process resulting in inconsistent hiring decisions and evaluations.
The company had to answer a key question: How can we keep up with high agent turnover rates and achieve global consistency in recruiting, screening and training across all of our outsourcing partners scattered throughout the globe?
Global ISP was already using Nexidia’s speech analytic solution to analyze and understand what was driving certain calls, such as transfers escalations and credits. After an evaluation process, the company decided to invest in the vendor’s new Language Assessor tool to automate and streamline its global agent hiring processes to ensure service consistency and quality worldwide. An intuitive online system was set up for the language evaluation process, instructing applicants to register and complete assessments online. The customized assessment script included product references specific to the company, as well as some frequently mispronounced and misunderstood words commonly encountered in the course of their business.
Using the tool to assess proficiency against North American English, the solution analyzed the recorded script readings and assigned various scores — for pronunciation, fluency and speed — which were consolidated to a total score for each applicant. The assessment results were broken down into individual phonemes so reviewers could pinpoint a candidate’s specific problem areas. Applicants scoring below 60 were not invited to the next round, while scores in the 61-70 range were reviewed manually; scores above 70 were automatically passed on to the next round for in-depth interviews. After deploying the speech analytic hiring assessment tool, the company was able to dramatically increase the number of applicants they could pre-screen, not to mention boost the overall quality of the applicant pool in subsequent evaluation rounds. The staff of four FTE’s previously dedicated to manually reviewing applicant recordings was slashed to a single employee — who now spends only a small amount of time reviewing applicants in the gray area (score of 61 to 70), focusing instead on other intangibles, such as a candidate’s skills relating to etiquette or empathy.
In the end, the service provider significantly streamlined and standardized the agent hiring process, saving the more expensive and time-consuming skills testing for candidates who passed the initial hurdle by demonstrating the appropriate language skills. With a consistent, objective assessment system in place across all applicants, the company and its network of outsourcing partners were able to focus on other agent training initiatives, such as identifying current agents requiring additional language or accent training.
The Dangers of Communication Breakdown
According to recent findings from the CFI Group’s Call Satisfaction Index, poor agent-customer communication is a primary contributor to poor customer experience in offshore call centers; their research concluded that agents “with poor communications skills resolved problems 45% of the time, while representatives who speak clearly are able to solve issues 88% of the time.”
In addition to the obvious negative customer experience implications, these scenarios can also significantly add to a company’s cost to serve. Worse, these interactions can and do result in lost customers and lost revenue.
The study reported that the two biggest customer complaints were (1) not getting their problem solved on the first call, and (2) communication problems with offshore call centers. By leveraging the same underlying technologies powering existing speech analytics solutions, global organizations can ensure that outsourced service agents possess the most basic abilities to communicate effectively with customers and thereby lay the foundation for building positive customer experiences and strong customer loyalty.
Chris Jeffs is vice president of product management for Nexidia. www.nexidia.com
