More information around different SL % intervals

[marosenthal]

Hello,

I was wondering if anyone has experience with creating different Service Level % intervals based off the interactions level data? We have it set to calls within 30 seconds currently but we also want data around 60/90/etc

We're having trouble getting interaction level data to properly match up to what we see when viewing the queue level performance - if anyone has feedback it would be much appreciated!

[Eos_Rios]

I create several different service levels and ASA's due to differing regulatory requirements by states being handled by the same queues, what are you doing and what are you comparing it to that is giving unexpected results?

[marosenthal]

Appreciate the response here.

My understanding is that there's queue level data, and then individual interaction level data. We can only see SL % at the queue level, but our initial thinking was if we can replicate at the interaction level then we could create whatever service levels we'd want to look at ultimately. When we extract data from the queue level and then from the interaction level, we're having issues getting metric counts like handle volumes to match up exactly.

Would you be willing to share how exactly you were able to create the different service levels/ASAs on your end?

[Eos_Rios]

@marosenthal

My understanding is that there's queue level data, and then individual interaction level data

Every event pegs at it's own moment and in it's own place, but that doesn't necessarily mean all events can be measured at all levels. Also how you use your platform can greatly change what things mean.

For our implementation we base everything on the provided Metrics from Conversation Detail.

tAnswered      The amount of time an interaction waited to be connected to an agent.
tAbandon       The amount of time before an end user abandoned an interaction in a queue.

Answers only occur at Agents
Abandons only occur at ACDs.

We calculate our service levels by the number of times those metrics occur within a given Queue within the defined ranges for that Queue.

So Queue 1 may have a threshold of 30 seconds (30000) and queue 2 may have a threshold of 15 seconds (60000);

Which summarizes to;

We choose to hold abandons and answers to the same threshold but it's also possible to apply different thresholds to each this way.

The simplest version of service level there is ( Answered In Threshold + Abandoned in Threshold) / Offered
But we have a few variations due to different requirements over how abandons are counted in different regulatory entities. The Genesys native version of the formula is dependent on configurations involving short abandons and what your short abandon threshold is.

For more casual reporting we use those two metrics to create waterfall tables showing count of calls that fall into varying degree buckets from 10 seconds to 120 minutes by queue by interval;