Managing Daily Staffing and Service

The pieces are in place. You gathered and analyzed mounds of historical information to
arrive at a call forecast. Then you calculated the number of staff needed by half-hour to
meet your center’s speed of answer goals. Finally, you juggled schedules until you
arrived at a reasonable mix of efficiency and acceptability. All this planning means that
you should have the “just right” staff in place every half-hour to match the workforce to
the workload. But you know what they say about the best laid plans…

This article will outline the steps of tracking call center performance within the day to
ensure that the plan is actually working based on the realities of the day.

First Things First

The last four articles in this series have outlined how to get to a schedule requirement by
half-hour. Once the forecast is done and schedules have been assigned, the next step of
the process is to manage the daily schedules to ensure service levels are being met half-
hour by half-hour.

The Real-Time Tracking Process

There are three steps in the daily performance tracking process:

Step 1: Tracking
Step 2: Communications
Step 3: Reaction

Tracking and Analysis

Tracking performance within the day means tracking the three elements that affect
service: call volume, average handle time (AHT), and staffing levels.

Let’s take a look at how a variation in any one of these components might affect net
staffing and service. For example, what if staff adhered to schedule and the call volume
forecast was right on target, but calls took 30 seconds longer to handle than planned?

Time Call Forecast Forecast Actual Actual Required Net

Volume AHT Staff Calls AHT Staff Staff
6:00 280 320 56 280 350 61 -5
6:30 310 320 62 310 350 68 -6
7:00 350 320 69 350 350 76 -7
7:30 380 320 75 380 350 82 -7
8:00 420 320 82 420 350 90 -8
8:30 450 320 88 450 350 96 -8
It’s easy to imagine a scenario where calls simply take longer to handle than planned.
Perhaps the computer system is slow today, or a different format for the billing statement
causes an extra question per call. Or perhaps this is the week a new hire class comes on
the phones and their longer handle times drive up the average from 320 seconds to 350
seconds.

With this longer handle time as the only variation, staff requirements are affected
significantly. In the above example, staffing at 8:00 would need to be adjusted from 82
to 90 staff in order to meet an 80% in 20 seconds service goal. If staffing is not adjusted,
the service level will drop to only 6% of calls answered in 20 seconds.

Of course, things might go the opposite direction and workload could be less than
forecast. Consider the effect on staffing requirements and service level if the marketing
campaign doesn’t go as well as expected and call volume is 10% lower than planned.

Time Call Forecast Forecast Actual Actual Required Net

Volume AHT Staff Calls AHT Staff Staff
6:00 280 320 56 252 320 51 +5
6:30 310 320 62 279 320 56 +6
7:00 350 320 69 315 320 63 +6
7:30 380 320 75 342 320 68 +7
8:00 420 320 82 378 320 75 +7
8:30 450 320 88 405 320 80 +8

This time at 8:00, only 75 staff will be needed instead of the 82 that are scheduled to
work. With everyone on the phones, service level will be 98% in 20 seconds. And while
that’s great from a service perspective, the overstaffing represents an unnecessary
expense.

Tracking Schedule Adherence

Since one of the three factors to track by half-hour is staffing, a critical piece of the daily
management plan is tracking and managing schedule adherence. You’ll want to match up
real-time status and work state information from the ACD against planned schedules and
daily schedule exceptions to effectively track whether staff members are doing what
they’re supposed to be doing.

Historical information about adherence and compliance is useful, but what’s really
needed in an intra-day environment is real-time adherence information. This can be
achieved by walking around and checking filled seats or can be done quite effectively by
a workforce management system. These real-time adherence systems can take real-time
status messages from the ACD and compare to an agent’s work schedule for the day.
Any variations that exceed user-set thresholds are reported immediately.
For example, suppose Rachel’s supervisor has set a five-minute “grace period” for
getting back from breaks and lunches. Rachel is scheduled for a break from 10:30 –
10:45. As long as Rachel is back by 10:50, nothing is reported. But at 10:51, if Rachel
isn’t logged back in and available, her name pops up on screen as out of adherence.
While these tools may be viewed as “Big Brother” or “Watchdog Technologies” by
some, they can be an effective way to make sure the third component of the half-hourly
puzzle (staffing) stays on track.

The Real-Time Picture

It’s important to track the three components of call volume, handle time, and staffing
levels throughout the day to see what the impact will be on net staffing. You’ll want to
start early in the day and watch these numbers to see if any significant patterns are
developing. As soon as you spot a trend that seems to be consistent, it’s time to
reforecast based on the new numbers and predict what net staffing will be as a result for
the remainder of the day. If you catch it in time, you can fix it before it’s too late.

While all three numbers should be tracked against one another and used to calculate net
staffing, there are single numbers of half-hourly performance that may provide a quicker
look at whether or not you’re in trouble. These measures of call center performance can
indicate if all is well or about to get out of hand.

For this snapshot of performance, some call centers use a measure of the number of calls
in queue. They may set a threshold so that as soon as a certain number of waiting calls
exist in the queue, a reactive process occurs. Others use the age or length of the oldest
call in queue and when that wait time exceeds a defined limit, a reaction is imminent.
Both of these are true real-time measures and represent a snapshot in time.

The question at hand is whether these are the measures that should be used to actually
direct the real-time reaction process. It is our opinion that while these two measures are
useful as warning signals, using them to initiate adjustments may be ill advised.

Think instead of setting and meeting service objectives as one half-hour race rather than
six five-minute sprints. While an exceedingly high number of calls in queue at any given
time might be alarming, it is likely that this number may go back to normal within five or
ten minutes as breaks overlap or the natural ebb and flow of random calls happens. You
should evaluate carefully before using these real-time measures to drive changes and
adjustments.

So what is the single best indicator of performance and service if not those real-time
snapshots? Actually, service level or average speed of answer (ASA) for the half-hour is
not only the number that should represent the day’s service picture, but should also be the
measure used in each half-hour to drive adjustments and reaction strategies.

Communications Strategies
Once the numbers indicate that overstaffing or understaffing is happening, the next step is
to communicate the problem. Developing a communications strategy involves deciding
what to communicate and who should be in the communications loop.

It is likely that the workforce planning specialist or team will be the first to know there is
a problem. This group may communicate directly with staff for the necessary
adjustments, or may relay the information by team or by supervisor where changes are
needed. This communication may begin as soon as there are warning indicators in order
to provide a “heads up” about coming changes, or may not occur until it is time to
actually make the adjustments.

Reaction Options

It’s important to have a reaction plan in your back pocket for situations that require just a
little tweaking as well as major adjustments. You should always know your options
whether you need to add or subtract two staff or twenty.

Selecting a strategy depends upon answers to the following questions:

• How severe is the problem?

• What is the effect on service level to customers?
• Can you fix the cause rather than react?
• How long will the problem last?
• What are the options from easiest to hardest?
• What will each adjustment cost in terms of dollars, resources, and effort?

Common staffing reaction strategies for understaffing include:

• Have supervisors or other staff take calls

• Delay meetings or training
• Eliminate optional call content
• Engage other qualified staff or outsource

Likewise, understaffing can also be addressed by making technology changes such as:

• Re-routing calls to other sites or groups

• Adjusting delay announcements
• Changing ring delay settings
• Adjusting trunks
• Invoking other technologies such as IVR or callback messaging

As seen in an earlier example, sometimes the adjustments need to go the other way. In
cases where overstaffing is apparent, the following reaction strategies might be used:

• Do impromptu training
 • Schedule team meetings
• Catch up on paperwork
• Make proactive outbound calls
• Offer time off without pay

Conclusion

Once the forecast is in place, staff calculations are complete, and staff schedules are
assigned, there’s daily work to be done in the workforce management process. Tracking
the half-hourly components of call volume, handle time, and staffing levels will provide
you the information you need to communicate status and make necessary changes to
ensure service goals are consistently met. The key is having a systematic process in place
to track the information so there’s sufficient time to react to make a difference for the
day.

Next Steps

In this article series, we’ve covered the entire workforce management process. These
steps are:

1. Forecast workload
2. Calculate staff requirements
3. Create staff schedules
4. Track call center performance

The articles have outlined the “how-to” steps of performing each phase of the process.
And while all these steps can be done manually, there’s a point at which automating the
process with a workforce management system is a must. In the final article in this series,
we’ll discuss workforce management automation, including qualification criteria,
questions to ask the workforce management vendors, and how to budget for the process
in terms of dollars and resources. Stay tuned!

About the Author….

Penny Reynolds is a Founding Partner of The Call Center School, a company that provides a wide range of
educational offerings for call center professionals. Penny is a popular industry speaker and is the author of numerous
call center management books, including Call Center Staffing: The Complete, Practical Guide to Workforce
Management and Call Center Supervision: The Complete Guide for Managing Frontline Staff. She can be reached at
615-812-8410 or by email at: penny.reynolds@thecallcenterschool.com.