IVR Tuning: You Get What You Pay For

Ah, the dreaded tuning exercise.  Once the province of skilled professionals paid large sums to keep an IVR application performing and relevant, now in the hands of every business user with a cloud solution.  Raise your hand if you love tuning your IVR!  No one?  What about those fancy analytic tools you have that show you exactly where callers are abandoning the application?  Doesn't your reporting tool allow you to drill-down to individual calls and even listen to them?  Doesn't all that information make the tuning exercise easier?  More fun?

The truth is, people procrastinate or completely avoid work they do not enjoy (which is why I am writing this instead of preparing a slide deck).  Despite the powerful tools available to IVR administrators, poorly tuned applications, even ones with loops, dead-ends, and bad error handling, are easy to find.  There is a surreal clarity that comes from working in the customer experience field.  We tend to participate more with automation and are more keen at noticing errors and defects.  More typical users abandon much more quickly.  They rely on instinct much more than we do.  I think for the average customer, the minute they notice the IVR sounding like an IVR ("please listen carefully as these menu options have recently changed"), they start looking for the exit.  Their complaints about the automation tend to be less specific.  They will not tell the representative that there is an error on the third menu or that option 2 always takes you back to the beginning.  Instead they just "hate your automated system".  Now you know you have a problem.  You should have tuned your application sooner but there were five other projects that were more important, ones that you hated less.

This is an example of where the promise of cloud has let us down.  Your Vice President of Support bought the idea that IVR tuning was something that could be handled internally.  There is a great set of tools designed to identify problems and an easy design interface to make small changes. And without that line item for service and maintenance, the ROI looked even better.  One year later, after the rubber has met the road and the wheels have nearly fallen off, after the Service Director who recommended this system has moved-on to another job, no one remembers that self-management  was a big part of the justification for buying that cloud platform.  Time to call your vendor and buy some training so you can begin to get a handle on that participation rate.  Maybe in three months you can finally begin making those changes.

The cloud IVR space is saturated with companies promising ease-of-use and total self-management.  They have deceptively low setup fees (because all they are doing is creating a user account) and even templates to get you started.  If your needs are simple and your call volume is low then so is your risk.  Most of these vendors employ only a handful of Voice User Interface (VUI) designers and rarely have a Speech Scientist on staff.  They offer no one to turn to for regular tuning and do not have the expertise to diagnose performance problems.  Technology is important and a slick user interface makes for a very nice demo but they are no substitute for expertise, experience, and a culture of service.

The market is also proving this out.  As we see more and more consolidation these slick start-ups are being gobbled-up by the veterans who know how to support enterprise customers.  Multi-tenant is giving way to multi-instance (which is a nice way to say hosted) and executives are seeing that the improved customer experience of live agents dedicated to a specific channel (rather than elusive multi-tasking, multi-channel, super-agent) is preferable to the fuzzy ROI of self-management.

Large enterprise customers have always chosen vendors who stay engaged and provide application tuning support for their entire life cycle.  IVR tuning is an iterative process.  It never ends.  Reporting and analytics will alert you to when your application is falling below performance thresholds but only after it has begun to cost you money.  And unless you are skilled and experienced at tuning, you will be learning on the job, compounding your losses because the tuning exercise takes weeks or months when an expert could have done it in days.  Mid-market companies buy in to the idea that self-management is more affordable or is the only affordable option, when nothing could be further from the truth.