This article originally appeared in the November 2012 issue of INTERNET TELEPHONY.

This is the first in a series of articles that breaks down popular myths associated with IVR systems and offers tips for increasing utilization. The other two pieces of this series can be found at www.tmcnet.com. The specific URLs are http://tmcnet.com/59204.1

and http://tmcnet.com/59205.1.

Do you know what percentage of calls your IVR system is successfully processing? This data is key to making improvements in IVR system utilization and increasing your return on investment. But few IVR system vendors want you to know your true IVR utilization rate. But before we can improve IVR system utilization we need to establish a utilization baseline. 

Businesses are aware that increasing IVR utilization will increase their ROI. After all, a call directed to a live customer service agent costs 100 times more than a call handled through an IVR system. However, it is a commonly held belief that no standard exists for measuring IVR utilization. In the absence of such a standard it is impossible to compare the rates of different vendor offerings or know when you’ve improved your existing system’s utilization rate.

Why it’s a myth

Increases in IVR utilization are not easy to achieve: It takes significant time and effort to improve usage, and not every vendor is up to the challenge. Many IVR system vendors would have their customers believe that there are multiple ways to calculate IVR system utilization. Using different calculations or changing them over time gives IVR system vendors an advantage. That’s why vendors don’t want customers to use a standard form of measurement. As long as customers don’t have a way to compare today’s usage with tomorrow’s – or their system against a competitors’ – then the numbers can be tweaked in the vendor’s favor.

Busting the myth

It’s easy to show an improvement in IVR system utilization when the math is manipulated. However, playing with statistics doesn’t change the number of calls going to live customer service agents. It simply changes the way the calls are being counted.

For example, counting customers who hang up during the “please hold” message as “satisfied” rather than “transferred” increases the IVR utilization rate. If a call center is closed (for a holiday, due to weather conditions, etc.), counting callers during this time frame as “satisfied” rather than “transferred” increases the utilization rate during the call center’s open hours.

The math can be manipulated to show an improvement in utilization even when the number of calls to customer service agents has stayed the same. The mythical utilization number is just that. It’s not real because the math isn’t correct.

How to calculate IVR system utilization

The following calculation is a standard means of measuring IVR system utilization regardless of the system you are using:

So, if an IVR system received 200 calls and 50 of them were transferred to the call center, the transfer rate rate would be:

50 / 200 x 100 = 25%

And the utilization rate would be:

100 – 25 = 75%

When using this calculation, it is important to remember that not all calls transferred to the call center are the result of a failure in the IVR system. Almost all IVR call flows have points in them that, for business reasons, direct callers to the call center. For example, callers indentified as sales opportunities may be sent out of the IVR system to a live sales agent who is better capable of pitching a product to that individual’s needs. Other examples include collection accounts, and fraud flagged accounts. 

In general there are three types of transfers:

Business Initiated– These are cases above where specific business rules mandate the caller be transferred from the IVR to an agent regardless. This includes cases where the caller explicitly selects an IVR option that is not implemented in the IVR and requires an agent.

Caller Initiated– These are cases where the caller is explicitly requesting a transfer to an agent by pressing zero.

Application Initiated– These are cases where the application is forcing the caller to transfer to an agent due to too many user input errors, speech recognition errors, caller timeouts, or overall input attempts.

The business initiated transfers have nothing to do with IVR system utilization and should not be counted against the IVR in the utilization calculation. For this reason, you should subtract these types of calls from the number of callers transferred, but keep them in the total number of callers. Because the IVR system performed as requested, these calls should count toward IVR utilization, not against it. So the updated formula will now look like this:

How to determine if your utilization numbers are correct

Unless you have very detailed call flow reports that show exactly where every caller either hangs up or is transferred, it is difficult to determine whether the utilization rate your IT department or IVR system vendor is providing is correct. Request from your IVR provider or internal IT shop exactly that: a detailed call flow report showing all prompts/menus/steps where callers are hanging up or being transferred.

These reports should be very detailed and precise. Don’t be fooled by cumulative totals, such as 100 callers tried to transfer funds in the IVR; 50 succeeded, 25 hung up, and 25 were transferred. You need very detailed reports of where the call failed. Was it date selection, amount selection, or from/to selection? If your IVR vendor is not showing you this level of detail, it’s time to switch IVR providers. You’re not getting an honest utilization rate, which would enable you to increase utilization and your ROI.

Donny Jackson is senior vice president of software development at USAN (News - Alert) (www.usan.com).