Who are you? A Leading Health Insurance Plan Wants to Know

Who are you?  A Leading Health Insurance Plan Wants to Know

One plan’s experience improving their IVR                                                                      

A large health insurance plan and AVOKE customer recently used AVOKE Analytics to improve the identification process in their IVR.  They knew that members struggled to provide their ID, but needed to understand the specific reasons why – and what changes would improve the process.  The company wanted to both reduce the amount of effort required to use the IVR, and increase the number of members that successfully identified in the IVR.

The IVR prompted members to speak the alpha-numeric identifier, or to get help by saying “instructions”. The help function instructed members to use the keypad to enter their unique identifier and to press the star key in place of each letter in their ID.

Using AVOKE Analytics, the company discovered that the help instructions were not clear to their members, despite the fact that they sounded good to the people that had written and approved them.  Seventeen percent of members asked for the help, but only half of them succeeded in following the instructions to enter their ID.  Many of the failures were due to members incorrectly using the star key between every digit, instead of just in place of letters.  And 25% of the failures were due to members pressing the pound key at the end of their ID, which the IVR interpreted as an incorrect extra character in the ID string.

AVOKE also revealed that members were frustrated by the error-handling process, which forced over 20% of members through three or more re-tries before routing them to a default agent.

The bottom line is that 37% of members tried and failed to identify in the IVR.  This frustrated members and added handle time by increasing the number of calls where agents had to perform a full authentication process instead of a shorter ID confirmation.

AVOKE Analytics not only helped discover the issues, it also provided the ability to drill-down and listen to actual caller behavior inside the IVR.  Listening to real caller behavior is far superior to log data or test calls, since testers know the system too well to interact with it like real callers.  And log data doesn’t capture the caller’s behavior in enough detail to learn why the IVR isn’t working as it was intended.

In this case, the company used insights from AVOKE Analytics to develop the following improvement strategies:

(a) Change the initial prompt to “Say your member ID or key in just the numbers”.  This simplifies the design by eliminating the separate help function.  It also allows members who immediately enter touch-tone numbers to be successful.

(b) Modify the IVR application logic to interpret touch-tone “2” as either the letter “A” or “B” when entered in the correct position for a letter in the ID string.  For this company, the letters “A” and “B” make up greater than 80% of the letters used in their IDs.  And “ABC” are the letters associated with the “2” key on a telephone keypad.

(c) Allow an optional “#” at the end of the ID string.

(d) Discontinue multiple ID requests. After the second failed attempt, send the member to an agent for authentication.

Results:

AVOKE Analytics provided informative data to the company, who was able to employ a quantitative, data-driven approach, knowing the ROI before committing to make any changes.  Using insights from AVOKE to drive their improvement efforts, the ID success rate increased from 49% to 55% of callers.  Members experienced a better interaction, requiring less effort to resolve their call (average time spent in the IVR was reduced by 10 seconds).  And the contact center benefited from reduced handle time since more calls were offered to agents with authenticated customer information in the screen pop.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s