measuring service levels

[scriptscience]

We have several call centers, one of which has a touch tone menu for the caller. The Application service level (percent of calls answered within target answer which is 20 seconds) is below 50% whereas our non menu areas are in the 80's. Should I measure an area with a menu at the skillset? Is there a different target answer if the phone center has a menu versus not? Help!

 

[Milesprower]

I'm assuming the menu is built in the primary (not master) script, so that the time callers take to make their menu choice is counted against the application service level. If you want to be consistent across your sites, you should build the menu in the voice mail system before the call is transferred (viz CDN) to the SCCS.

If you do not want to change scripting, you can estimate the average time callers are taking to navigate through the menu and add that to your service level threshold. So if you service level threshold is 20 seconds and it takes (on average) a caller 15 seconds to navigate the menu choices, you would adjust your service level to 35 seconds.

The first option is preferred as you will then have exact conformity across your call centers and you will not be making a "guess" as to the average time it takes to navigate the menu.

 

[captaingadget]

What Miles says is absolutely spot on, other things to consider are: putting the menu in the master script will only be successful if you are using CC6 as the application threshold timer in SCCS5 still starts when the the call enters the call centre. With CC6 the application threshold timer starts when the execute script command is executed in the master script.

If you still want to keep the menu selection in Symposium you could make the menu selection route the call to another CDN which will execute another script from the master so that the threshold timers are cleansed as it will become a new call as far as Symposium is concerned.

 

[scriptscience]

Thank you both for your timely response. I am going to restate my question as you are both veterans of this forum and I highly respect both of your opinions. If there are two identical call centers and one staffs up so that you go straight to an agent and the other cuts costs by having a menu (menus generally save the call center money)shouldn't the time the customer spends navigating the menu be factored into the equation? Generally customers would rather go straight to agent than a menu, they don't like menus. If you don't count the time in the menu, how can you put a check on poorly designed menus? I think a center that answers the call in the target answer time (20 seconds in our case) without a menu is providing a higher servce than one where the customer is in the menu for 5 minutes and then when presented to agent they answer within 20 seconds.
Looking forward to your input.

 

[captaingadget]

It's good to know that you're actually interested in the "customers experience". Here's my view, if you want to measure the customer experience then you must include the time spent in the menu. If you want to measure the call centre performance then you should exclude the menu time. This is where us as designers fall out with the call centre managers who are only interested in the service level figures. So it's rather up to you or your management to decide which ethic is going to prevail and then choose a solution which will meet this.

 

[DancingDave]

A well designed menu would, however reduce handling time and hand-offs as the customer is more likely to be put through to the right person, first time.

It might not be true if you call centre agents are multi-skilled, but if this is the case you would have to question whey there is a menu there at all?

If the menu is to gather MI, then you could always consider putting it in the cloud.

DD

 

[Milesprower]

You are asking the right question: what is the best process for our business? There are many other factors to consider outside of what SCCS can tell you.

What is the total cost (and revenue) per call? We know we can reduce (and measure) cost, but that does not always translate into what is best (bottom line) for the business. Too often, this question is not even asked as it raises many core business issues that many are not ready to address.

 

[Milesprower]

One follow-up for Captain: I know many customers who have put menus into the master script (going back to Rls 1.5) to keep that time out of the primary script service level calculation. I know of no limitation in Release 5 that lumps time in the master script into time in the primary script. Can you elaborate?

 

[scriptscience]

Thanks everyone. This is what I have concluded from your valuable input.

1.I believe are bad menu design is the major culprit so we are going to work on that. The most common menu selection is currently embedded in the middle of the menu versus at the beginning.

2. The touchtone IVR is located in a department within our brokerage firm providing home office support to our financial advisors so the customer experience is very important as they generate the revenue and are the ones calling in. From your input I believe we should measure at application versus skillset. We do not gather account number or the like from the IVR; it is to get them to the right associate specializing in their request.

3. When I pull the "Application Delay Before Abandon" would all of you agree that it is the summation of time navigating and waiting in the queue? Therefore a call that spends 30 seconds navigating the menu and 10 seconds waiting in the queue looks the same as one that is 10 seconds in the menu and 30 seconds waiting in the queue.
I am trying to distinguish between calls that abandon after threshold (20 seconds) while navigating menu and were never presented to a skillset and those that for example abandoned after threshold because they were in the menu for 15 seconds and then in the queue for 5 seconds. What report ,if any, can I pull to tell me how much time they actually waited in the queue versus the time they were navigating the menu?

4. For Dancing Dave, Can you explain MI and cloud??"If the menu is to gather MI, then you could always consider putting it in the cloud."

Thanks everyone.

 

[DancingDave]

ScriptScience, we have several of our key numbers give a voice menu that is on the telco's platform ("cloud") i.e. not based on CC6/Call Pilot.

This allows us to make better use of CP channels, doesn't distort Management Information in terms of not having to factor in menu navigation into Grade of Service. The telco provides us with the number of callers that pressed 1, 2, 3 etc.

The only drawback is that you don't have as much control of the message as when it's on yoy own kit.

DD

 

[captaingadget]

One for Miles. Abandon time, answer time etc are all pegged in the master script in SCCS5 and not in the primary application. Below is an extract from "Whats New In CC6"

Application call answer delay and call abandoned delay pegging is more
accurate: statistics are calculated from the time the call enters the primary
application rather than from the Master script.