In UCCE, reporting information is split into two parts: before delivery to the agent, and after delivery to the agent.
For the first, Call Types give you everything you need. But they need to be set in the script at various points, and you need to understand how they work.
If all you have is the Call Type attached to a dialed number on a scheduled script, then you are going to lack detail as the call flows through to the "Queue To Skill Group" node, and you will not be measuring the true abandon rate - although ASA will be accurate.
Once the call is delivered to the agent through a skill group, average hold time and call duration will be pegged against the agent (grouped by agent, skill group, team or peripheral) and against the skill group. This is more straight-forward.
Going back to the first reporting paradigm, the Call Type, the best way to get what you need is to use extra call types wherever the call divides and/or has the possibility of being released or dropped.
Finally, right in front of the "Queue to Skill Group" node, set a Call Type to give accurate abandoned statistics.
Consider the following: calls start the main routing script. These have pegged the Call Type attached to the scheduled script.
Now you check for holidays and if it's a holiday, you want to release the call - so have a Call Type to count these.
Then check for Opening Hours - again, you want a Call Type to count the number of after-hours calls you get.
Now you want to give the call to the IVR, and in there, customers can abandon. You want to know about these; so set a call type going in. Otherwise you will pollute your final Call Type measuring agent performance with abandon at VRU.
Say the IVR is a simple - "for English press 1, for Spanish press 2" - you will probably have Call Types to count the English calls and the Spanish calls.
Now the last part - right before the "Queue to Skill Group" (or a combination of these in any complex overflow). I call this Call Type the "queuing call type" and have a naming convention - ends in "_Q". This one will give accurate abandon stats, because we have cleared calls that abandoned in closed messages, holiday messages, and in the IVR.
Only two types of abandons can peg against this Call Type - when the caller drops in queue, and "abandon ringing" if you don't use auto-answer.
You need absolute clarity in the design of the "Call Type Tree" and must know which ones typically "Overflow Out" - the intermediate ones with no escape path, just for reporting.
When the customer has fairly open reporting requirements (they want stats on everything), you will end up with a lot of Call Types. I don't have a problem with that, but some do.
WebView reports can be very misleading if there are ways the calls can "bleed" out of the Call Type and you cannot account for them.
Good script design means that the scripts must be well instrumented with a good Call Type tree. Then WebView will be accurate.
Regards,
Geoff
