Using Analytics to Improve IVR/Speech Rec Effectiveness: Nexidia

ATLANTA, GA – February 9, 2010 – Automated self-service voice systems, both DTMF IVR and speech recognition, are excellent cost reduction and quality service enhancement tools if deployed and managed right.

One of the best ways to find out if they are performing to expectations is by using speech analytics, which Nexidia has done. TMCnet contacted Jeff Schlueter, Nexidia’s vice president of marketing and business development, and asked him about this. Here are his insights, the study and results achieved and recommendations:

1. Improving self-service optimization is one of the more common things that Nexidia gets involved with, especially with new customers as it’s a good bit of “low hanging fruit” that can provide immediate economic benefit to the client.
2. Nexidia studied the self-service improvement components of more than 20 projects conducted in 2008 and found that 23 percent of all calls into a typical IVR system resulted in callers requesting to be transferred out to a customer service agent. This is a significant percentage of calls that aren’t being handled correctly by the IVR system, so the challenge was to determine just why callers were being transferred to a live agent, and how the systems could be improved to keep more calls in the IVR and handle customer issues more effectively.
3. Using Nexidia’s Enterprise Speech Intelligence ESI application to systematically review the bulk of these transferred calls, we identified three main categories companies can focus on to improve the effectiveness of their IVR systems. Speech analytics also provides the diagnoses and action plans that can be used to improve the systems to keep and effectively serve more of these calls:
   1. The most common type of error was what we call “Automation Opportunities”. These are what most folks expect to find, where the IVR system was poorly worded or menu choices were not intuitive to understand. Another segment of these calls is simply adding new capabilities into the IVR to reflect new service choices that customers had, not reflected in the IVR design. This category represented about 50% of all the calls that were transferred out of the IVR.
   2. The second category of transfers deals with routing improvements to the system itself. This manifests itself usually as simple error in how the IVR system is set up to transfer calls, where the call center intends a call to be routed to “X” and in fact it is routed to “Y” instead. This can often occur when calls are routed to an outsourcer or “off-hours” center during non-busy times. This category represents about 30% of all transferred calls.
   3. The third area for improvement is not actually in the IVR itself, but in the standard scripts and practices that companies should use to steer customers to the automated solution more effectively. By analyzing calls coming into the contact center by default (e.g. those that came directly into the contact center and not first through the IVR system) we see an opportunity for contact center agents to engage more proactively in “Self Service Advocacy,” i.e. train the agents to inform and educate customers on the availability of self-service solutions so customers will start to use them more regularly.
4. A word about the projects studied for this analysis. They spanned a variety of industries, from health care and health insurance to retail companies and financial services firms. As noted above, this is often one of the first types of analysis that Nexidia engages in with a customer as it can provide immediate benefit to the bottom line by identifying inefficiencies in the self-service system and point to clear ways to improve it.
5. The use of speech analytics in these types of projects follows Nexidia’s familiar flow. Nexidia’s patented phonetic analytics approach allows us to analyze large volumes of recorded conversations and pick out the specific topics and issues that are being discussed in each call. This provides a detailed categorization of “call drivers” and points to the specific action plans that can be implemented to address these call drivers and improve the systems. Contact centers usually have a very detailed understanding of the costs to handle each call into their system, compared to handling customer issues through self-service, so being able to retain more calls in the self-service system can have significant positive impacts to the bottom line.