No exposed tuning method. More brute force / trial and error would be to lower performance goals for the planning groups then generate a new schedule to see how predicted staff numbers look.
You could also open a Care ticket and then have it escalated to R&D so we could look under the covers to see if anything stands out. Barring some of those items I mentioned, we expect our predicted staffing to be within a reasonable range from actual agents if service goals are met.
One other item I didn't mention is if you have very fragmented planning groups and agents. E.g., many many low volume planning groups with agents spread all over. The more homogenous the load and resources the more accurate and precise forecasting and staffing prediction will be. The more heterogeneous and segmented/fragmented the harder it is to be accurate and precise.
------------------------------
Jay Langsford
Senior Director, Workforce Optimization Engineering
------------------------------
Original Message:
Sent: 11-13-2020 11:26
From: Greg Barrett
Subject: Forecasted Agents overstated
Good info but now it gets tougher.
Items #1, 2 and 3 – metrics are dead on. No obvious issues.
Item #4 – RTA runs about 97%.
Other things:
#1 and #2 – same BU
#3 always a challenge. However we see correct staffing numbers in lower volume periods of our work day. Once we get into high volume intervals the system is asking for too many agents.
#4 our emails are not overwhelming in off hours. Just a handful that shouldn't impact overall staffing.
So if we don't see any obvious issues, what is the best way to adjust the output of the system to be more realistic. I've worked in systems in the past that allowed me to set a Reporting SLA (our true KPI) and also a Forecast SLA that we could adjust up/down to impact staffing levels. Any thoughts?
Original Message:
Sent: 11/13/2020 11:08:00 AM
From: Jay Langsford
Subject: RE: Forecasted Agents overstated
It's always good to utilize the intraday monitoring view. Here are things to check when you think forecast staffing is higher than you need:
- Lower actual Offered than forecast
- Lower actual AHT than forecast
- Higher actual abandonment rate than goal specified or no goal at all specified for abandonment rate (I've seen cases where customers have no goal and routinely abandon 30% or more per interval)
- Lower actual agents than scheduled (should also look for low schedule adherence for the dates in question) - poor historical adherence is modeled into staffing requirement predictions
Other things to be considered:
- Are there agents not in the BU that can handle interactions in the BU?
- Do you flow-out interactions from the BU to non-BU agents routinely and at a significant volume?
- We are attempting to meet all service performance goals per interval so if you have hourly or daily business goals, know that we strive to hit more consistent levels of performance.
- If you have a large volume and backlog of deferred work (i.e., emails), we don't batch them up, but rather forecast staffing to keep the backlog at a steady state.
------------------------------
Jay Langsford
Senior Director, Workforce Optimization Engineering
Original Message:
Sent: 11-13-2020 08:26
From: Greg Barrett
Subject: Forecasted Agents overstated
We are seeing the Forecasted agent count being overstated in the Schedules view. Our transaction volume is very accurate +/- 5% and we are actually doing better than our SLA goal on all channels. However, the forecasted agents shows we are understaffed by a good amount.
I want to use this forecasted number for future budget planning but it must be accurate.
How can I impact that number? I'm thinking of modifying our SLA goal templates in WFM but not sure how that impacts our Performance Views and other metrics.
#Ask Me Anything (AMA)
#Implementation
#PlatformAdministration
#Unsure/Other
------------------------------
Greg Barrett
Outdoor Network, LLC
------------------------------