Michael Chavez, Vice President of Marketing, ClickFox

ClickFox software maps customer interactions to a business-relevant model of the system, offering insight into how your customers are interacting with your business, and a roadmap to ROI. ClickFox models and analyzes behavior in any interactive system, including: interactive voice response (IVR), speech recognition, Internet/intranet sites and Web applications, kiosks, and CRM and enterprise applications. ClickFox is privately held with investors that include Cedar Fund, Delta Ventures and Veritas Venture Partners, and is headquartered in Atlanta, Georgia. ClickFox's customer base includes companies such as Bank of America, BellSouth, Coca Cola, Countrywide Financial, Sprint, and The Weather Channel.

Speech Technology Magazine sat down with Michael Chavez, vice president of marketing at ClickFox, to talk about aligning user goals with the business goals of the enterprise.

Q. Hi Michael, what is your role at ClickFox?
A.   I'm currently the Vice President of Marketing.  Previously, I was Vice President of our Professional Services group and was a member of the founding executive team at ClickFox.

Q. You presented at SpeechTEK West on Customer Behavior Intelligence. What is this and how does it impact enterprise clients and their customers?
A.   Customer Behavior Intelligence (CBI) is essentially about turning existing customer interaction data into a meaningful story.  This previously unknown story is about what customer's goals are, what they actually do when they interact with enterprises and, most importantly, why.

The benefit to enterprise clients is that for the first time, they can actually see, investigate and continuously measure and monitor step-by-step, end-to-end behavior for all of their customers within any interactive system or across multiple touchpoints.  This capability—making visitor behavior obvious and measurable—allows enterprises to align their service goals, and specifically, their goals for their speech applications, with what users want to do and how they want to do it.  

Q. How do you suggest enterprise clients begin to align their business goals with the goals of their users?
A. First of all, the industry is just now realizing that the best and most financially successful self-service systems are those that customers want to use, as opposed to those  they feel they are forced to use.  There's a lot of good data out there suggesting that consumers don't dislike self-service, they just dislike bad self-service.  So, the best you can do, to balance financial results and customer satisfaction, is to get customers to adopt self-service systems willingly.

To do this, the first step is discovering and defining the customers' current reality.  Most of this misalignment of service goals and customer goals stems from lack of understanding about what users are willing to do and what they aren't when they hit, say, a speech-enabled IVR.  Without this knowledge, companies are left to rely on their assumptions and beliefs, which are usually incomplete. 

By implementing a CBI capability, companies can now see in an intuitive, visual map of interactions, where customers are having difficulty, what led to this difficulty, and where they are tending to abandon self-service and why.  Once they know these things, companies can develop very reliable hypotheses about what users really want to do and what's causing them to behave the way they are. In most cases, we can uncover how they respond to the prompts, structure and logic, giving us insight into what would happen if we changed those things.

Once we've fixed those things, the next big step is deciding what we should pay attention to over time.  Companies need to make sure that they can measure the impact of their changes—at the behavioral level—in order to understand how well their changes worked at the macro level.  Then, they need to continuously monitor the system to find other improvement opportunities, gain more insight into customers' needs and wants as they respond to a changing marketplace and work to integrate behavioral understanding into both their service delivery as well as their customer relationship management and marketing initiatives.

In essence, companies want their customers to use their speech systems and be happy about their interactions.  The only way to can get there is to understand exactly what to do to make it better and the only way to do that is to test assumptions against actual user experiences.

Q. What if the enterprise client doesn't know its customer's goals/experiences, then what?
A.   Few companies do know their customer's interaction goals and fewer really understand their experiences.  Many have sample or anecdotal insight, which is important, but again, is incomplete in terms of helping you to prioritize your initiatives. 

To complement this, you have to look at what customers actually do.  This is much more reliable than asking them what they think.  And, you must observe the behavior of as many customers as you can, simultaneously and continuously, to get the full picture.

Q. How do analytics play a role in this process and what are your suggestions for improvement?
A. With truly behavioral analytics, companies can learn about customer goals by observing continuously how they respond, in aggregate, say, to changes in the IVR system.   This process delivers amazing insight into what customers' really care about and how they want to achieve what they want to do.

To improve analytics, companies have to honestly assess what they currently have in terms of customer behavioral knowledge and insight.  Next, at a minimum, they should conduct an initial assessment or pilot using CBI software to understand what their customers are doing in their speech systems.  Then, they should compare the real story that customers tell through their behavior with what they originally thought was happening.  This gap analysis helps adjust thinking to reality and to generating and prioritizing improvement initiatives.  Finally, companies should be sure that, once they've acted on these initiatives, they measure the impact, not just at a high level, but at the level of actual behavior so that they know definitively that it worked or did not.  Then, they can decide whether to keep this process going by synchronizing it with their existing release process.

Q. During your presentation, you mentioned a case study on a Fortune 100 telco. Who is this client and how did you apply these theories to their deployment of speech?
A.  This client was a major long-distance provider. They're a household name.  What we found were essentially three things.  First, the company had lots of beliefs about what customers did and did not want from a speech interaction.  Second, we found that the customer's behavior told a very different story.  And, this story was essentially that speech was not the problem.  The assumptions that the company made about what prompt should go where, how prompts were worded, how they communicated to customers, the level of customers' understanding and the overall flow and logic, accounted for over 80% of the problems that led to customers bailing out to an agent or hanging up altogether.  The system did not really need much more speech tuning; what it needed was behavioral tuning.  Third, we found that iteration was critical to really understanding what would work.  It's not until you see how customers react to your changes that you really get insight into how to improve your speech results.

Q. Do you have any additional comments?
A. One last thing.  I encourage any company wanting to embark on this journey to take a non-judgmental, learning-focused attitude to improvement.  You can't fix a complex system with complex interactions in one go-around.  You have to do it over time, acquiring insight along the way.  The only way that will happen is to make a few mistakes, notice them quickly, integrate that understanding into your thinking and readjust.  What we have seen is that this is the fastest road to self-service success.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues