Turn Speech Analytics into Action

It’s easy to get excited about the promise of . The ability to gain insight into customer interactions and potentially enhance internal quality efforts is a compelling prospect when faced with fewer resources and the need to focus on elevating the customer experience. But can speech analytics live up to the hype? Does it offer enough of an opportunity to turn analytics into action? While all indications are that analytics solutions can be powerful tools, the value of the solution cannot be derived simply from the technology itself.

Let’s first examine the potential value of recorded content. Say you manage a center of 100 agents, each talking to 12 customers per hour for eight hours a day—each day approximately 10,000 call recordings will be generated. If your quality management staff reviews (on average) 10 calls per agent per month, then you are effectively reviewing the content of 34 calls per day of the 10,000 calls you recorded. This means that over 99% of the content—and the inherent value—is never realized. There’s no question of the potential value that exists in the content of these recordings, but how do you extract the intelligence from these recordings that will allow you to identify actionable opportunities?

So why is it worth pursuing that remaining 99%?

Performance Management

Analytics solutions augment quality assurance efforts in identifying agent improvement and performance maintenance opportunities. Without such a solution, quality assurance efforts are typically characterized by:

  • Manual conversion of unstructured data into structured data.
  • Difficulty in targeting the skills and knowledge that matter the most.
  • Uncertainty around agent behaviors that drive desired results.

Combining quality assurance efforts with an analytics solution will allow for:

Selection of high-value call types, new or infrequent scenarios, or specific skills for evaluation.

Identification of calls with markers relevant to your important metrics—handle time, first-call resolution, customer satisfaction and sales.

  • The ability to shape training and motivation with greater insight into collective agent performance.
  • Easy creation of pools of sample calls to demonstrate desired performance.
  • Proactive identification of trending terms to heed of potential quality issues.
  • Automate scripting and compliance adherence.

Analytics does not equate to performance management magic; there’s no “wizard behind the curtain” that will drive performance up. Or down. Or sideways. With an analytics solution, you are still going to listen to calls. You will still need the skills to interpret and derive insight from the data collection. And you must be cognizant that use of analytics can feel more intrusive to your agents.

Business Intelligence

This capability allows for data provisioning and analysis capabilities that transcends the needs of the contact center to potentially add value throughout the organization by:

  • Gauging the frequency of specific contact types.
  • Gauging the effectiveness of communication efforts.
  • Increasing the speed of analysis and ability to react to customer and environment changes/needs.

Customer Relationship Engagement

Using customer-identifiable information to anticipate and track behaviors and needs, and perform targeted service recovery efforts, are high-value outcomes of analytics usage. Analytics solutions will allow you to:

  • Target specific interactions tied to a customer or focused group of customers.
  • Identify and quickly respond to issues surfacing through social media channels, averting the potential need for perception recovery.
  • Pinpoint identification of risk to the organization.
  • Target customer perception and branding efforts at the individual customer level.

Implementation Success
Success with the use of analytics solutions does not just happen; what a solution identifies for your “inspection” is defined by you and your perception of the value of identification. And if you do not take the time prior to implementation to define a strategy for the use of your solution, then you may end up capturing a lot of data/information that—while interesting—may not provide anything insightful, actionable or of ultimate value to your organization or your customers. Refrain from being swayed by “cool” technology by identifying project goals before engaging in vendor selection.

Potential deployment goals may include the ability to identify:

  • Unhappy customers
  • Customers at risk of churn
  • Customers experiencing service resolution issues
  • Marketing and product enhancement opportunities
  • Agents using inappropriate wording/responses
  • Risks to organizational reputation/customer perception

As part of the implementation process, a keyword library will be built that will facilitate the identification of situations similar to those above. Typical keywords and phraseology may include something like:

  • “…frustrated”
  • “you people…”
  • “…cancel my service/subscription”
  • “…my money back”
  • “Every time I call…”

And remember—when you build your keyword library you are not just identifying words and phrases used by your customer— you are also identifying those used by your agents.
Validating Effective Use
While the most valiant of efforts prior to the implementation process are paramount in positioning your organization for success in using an analytics solution, the ultimate validation of the value of both your solution and your planning are the results derived by your efforts. When crafting your strategy for analytics deployment, consider the following as potential gauges of success; those which correlate to your organization’s metrics and comparison measurements must be included in your planning and implementation process:

  • Improve the customer experience
  • Increased revenue
  • Increase efficiency and cost-effectiveness of resources and processes
  • Increase the amount of actionable intelligence available to the organization
  • Elevate agent performance (and performance improvement in other areas benefiting for analytics data/intelligence)

Pitfalls to Avoid
Analytics solutions quickly provide targeted information that you have indicated is important to both your contact center and your organization. However, there are pitfalls to avoid when faced with receiving this magnitude of information so quickly. Those include:

…and spending time identifying activities that are inevitable to the business you operate. For example, most organizations express an interest in knowing when a customer makes the decision to discontinue their business relationship. However, customer churn is inevitable for most organizations so the value of such information is not rooted in the ability to identify every instance of churn, but rather in identifying those instances of “preventable” churn.

Do not allow the quick presentation of data to affect your need to trend and analyze the information being received, or you could find yourself making performance assumptions that are flawed.

When you identify information you feel you want to collect, validate the value of that information with others (both within and outside of contact center operations). Ask yourself, “How will I use the data to make things better?” and “How will I measure success?” If you cannot answer both of those questions (and neither can your business partners) then it is probably not data worth collecting.

There are certain aspects of your business that are recurring and have limited ability to drive customer satisfaction, and which fall within acceptable “norms.” Capturing data/information tied to operational norms is typically a low-value activity. Create high-value data/information by identifying situations outside of operational norms that you can use to either improve the customer experience and/or emphasize behaviors exhibited in high-quality customer interactions.

Jim Grace is a Contact Center Solutions Consultant at Interactive Intelligence, with 19 years’ experience in contact center management and consulting.

– Reprinted with permission from Contact Center Pipeline,

Related Articles

Back to top button