Managing Daily Performance: An Intra-Day Strategy to Manage Staffing and Service

By Penny Reynolds

The workforce management process involves a great deal of planning and strategizing to establish a reasonably accurate forecast of workload and an acceptable set of schedules for the staff to work to meet the desired service goals. However, within the actual workday, there are likely to be variations in either workload or staffing. Identifying when there is a change from the plan and determining how to adjust to it is the role of the intraday management processes.

Tracking the Plan vs. Actual

The first step in managing daily performance is to track whether the workload or the staffing levels are going
as planned so that corrective actions may be taken.

Because workload is a function of volume and handle time, if either volume or handle time is more or less than you staffed to accomplish, then there will be issues with daily performance. There may be more contacts than forecast or fewer.Handle time can be longer or shorter than the historical data would suggest.More workload than planned without a staff adjustment will result in missed service goals and stressed staff.On the other hand, less workload than planned may result in overstaffing and low occupancy of the agents.

In addition to deviations from planned workload, the staffing may be different from what was planned.Remember that even as the workday arrives there is an assumption of some shrinkage percentage to cover for absences, tardies, and unexpected situations.But shrinkage is an across-the-board percentage of loss assumption and the reality is that the number of people who call in sick, or the cancellation of a training class leaving several extra agents available for an hour is generally not known until it happens.It is even more challenging when skill-based routing and multimedia contacts are involved since all agents are not interchangeable – it matters which agents are not in place as the schedule assumed.

There are key metrics that should be compared and monitored to determine whether there are potential issues with workload or staffing. Some of the key comparisons that allow you to identify potential issues with workload include:

Forecast AHT vs Actual AHT
Forecast Volume vs Actual Volume
Forecast Workload vs Actual Workload

Some of the key comparisons that allow you to identify potential issues with the staffing plan include:

Forecast Required Staff vs. Scheduled Staff
Actual Required Staff vs. Scheduled Staff
Scheduled Staff vs. Average Positions Staffed (APS)

As variations from the plan are identified, it may be appropriate to reforecast for the remainder of the day.The situation may only affect a brief time or may last most of the day or all day depending on the cause, so noticing that a change has occurred is the initial flag but determining the likely length of the impact is also important to identify the need for a revised forecast and the best adjustment strategy.

The final comparison between scheduled staff and average positions staffed is particularly useful in identifying issues with fulfillment of the staffing plan. Average positions staffed indicates the number of staff logged in and doing work, and when compared with the scheduled staff indicates how much of the scheduled staff were actually working. Various factors that can compromise the accuracy of scheduled staff include:

Adherence to schedule
Unentered schedule exceptions
Undocumented schedule changes (e.g., trades)

Schedule Exception Management

Managing the entry of schedule exceptions into the WFM process/system is critical so that the real-time counts of scheduled staff can reflect the actual situation. Then adjustment strategies can be based on accurate data.By definition, a schedule exception is any deviation from the originally planned schedule assigned to an agent.The closer the timing of the data entry is to the first knowledge of the exception, the better planning can be supported.Entering absences or other changes the day after they happen may be helpful for historical reporting or payroll accuracy but does nothing to support intraday decision processes.

There are three basic types of exceptions:

Planned – These are known well ahead of time and should always be factored into scheduled staffing assessment. These are planned into the long- or near-term schedule, and might include plans for training classes, coaching sessions, or vacations. All of these should be documented into the plan as soon as they become known.
Unplanned + known – These may be a result of a last-minute management decision or an agent action such as an ad hoc coaching session or cancellation of a team meeting that leaves 15 agents on the phone. These exceptions should be documented as soon as possible to improve the daily staffing assessment.
Unplanned + unknown – These may not be known in time to be considered in planning intraday adjustment strategies, but they should be documented after the fact to improve long-term staffing plans. Failure of agents to adhere to their assigned schedules would fall in this category.

As the exception handling process is established there are some definitions that will be helpful.The first is the minimum and maximum length of an exception.For example, the plan and the adherence percentage goal assume a certain amount of loss due to agents being tied up on a call when their break or lunch period arrives.If an agent is 3 minutes late in going to break, is it worthwhile to have the agent request an exception, have the supervisor review and approve it, and WFM to enter it into the system?Many centers have determined that a reasonable minimum length is an efficient way to manage the volume of exceptions and the adherence goal is set to match.

The second definition is reason. If the exception is caused by a supervisory decision (such as a team meeting that runs longer than planned), the agent should not be penalized on their adherence goal. However, if the agent is late coming back from lunch, that is an agent adherence issue. The first should be entered into the system, but the second normally is not entered as it will show up on the reports as “out of adherence.”

Approval process is also something that needs to be defined. Who asks for the exception and who needs to be involved in the approval process? A clear step-by-step process can better ensure that everything is reported properly but an overly complex process can result in more lost time than necessary.

Timeliness of the information is also critical. The sooner the WFM team can be aware of the variance and build it into the intraday information, the better the adjustment strategies will be. In some cases, an advance notice may be required for approval and determining who needs to be informed is part of the process.

Data entry of the exceptions is handled in a variety of ways. The agent may enter it with a built-in approval process required. The supervisor may need to do the data entry for their team members. In some centers, all exceptions are entered by the WFM team. There is no “best practice,” but there are trade-offs. For example, if supervisors do the data entry and approval for their team and are incented to achieve high adherence results,
it may be tempting to approve exceptions that might otherwise be rejected.

Overall, the exception handling process needs to consider the value of the information versus the volume of work required. Too little information makes decisions difficult and strategies may be based on inaccurate assessments. Too many exceptions may be burdensome and expensive compared to the value they provide. Finding the balance for your center is the key. Some of the values of accurate exception tracking include:

Support for long-term staff planning such as hiring and shift bids
Improved shrinkage calculation and forecasting accuracy
Identification of reasons for variations and opportunities for improvement
Providing data to support coaching and discipline processes
Improving the accuracy of adherence scoring
Ensuring accurate time reporting (especially if linked to payroll)

Daily Performance Metrics

There are several real-time measures that need to be monitored to determine if they are out of normal boundaries. Reacting to these situations quickly can generally improve the consistency of the service results as the day progresses. These are:

Agent status – Has an agent been in the after-call work state much longer than is normal? Should a supervisor be notified to provide assistance?
Agent adherence – Are the staff where they are scheduled to be? Is an agent supposed to be logged in but is missing?
Number of contacts in queue – How many contacts are waiting to be served? Comparing the number of agents logged in to the number of waiting contacts can suggest whether the wait will be longer than the goal.
Age of the oldest contact in queue – This is sometimes called “longest delay in queue” and indicates the longest wait that a caller has experienced before being answered (also longest delay to abandonment is available). But that is an after-the-fact measure and monitoring the developing situation in real-time can help to prevent unacceptable delays.

Longer term measures are also useful in managing the intraday situation. While these metrics are not real-time, they can give some insight into problems that have occurred that may be addressed in future periods. Some metrics to watch include:

Service levels – This is the measure of the percentage of callers that were answered within the goal over a specific time period. It could be over a half-hour, hour, or even the entire day. It is only reported after the interval is completed so the opportunity to make a small adjustment to prevent a missed goal is past.
Average speed of answer (ASA) – This metric also measures the average wait for callers over a period of time, just like service level.
Abandon rate – This measures the percentage of callers who entered an agent queue but hung up prior to being answered by an agent during a specific interval. This is especially important in revenue-generating centers but provides useful information about caller tolerance for wait.
Delayed response – For work that is not related to inbound calls (or web chats), the service level and ASA do not apply. It is still important to know if the time to respond to the contact is within acceptable boundaries.

It is important to consider the interval over which these longer-term metrics are measured and how averaging may conceal significant fluctuations. In some centers, when some intervals have been below the goal, overstaffing is planned to exceed the goal in other intervals to improve the average results for the day, week, or longer period. However, customers served during different periods experience different waits. Those who are served slowly during busy understaffed periods wait longer than is acceptable and may decide to take their business elsewhere. Those who call during light overstaffed periods are served quickly and come to expect that level of service in the future and may be disappointed on the next contact. Focus on minor adjustments in real-time may provide for a more consistent result in all periods. It is becoming more common to see centers using goals to track consistency such as 85% of calls answered in 30 seconds in 80% of the periods.

In addition to measurable items that may be out of normal boundaries, there are often internal business events that can put the daily performance plan at risk if not factored into the workload or staffing plans. For example, has there been a flash marketing campaign that generated higher workload than the forecast? Perhaps there was an invoicing error that caused a lot of calls. System downtime such as telephony or IT maintenance can cause both fluctuations in workload but may also result in inaccurate reporting data. Weather disruptions can have a significant effect on some centers. Tracking the impact of such events and factoring them into the daily plan ahead of time whenever possible can help to improve daily performance.

Reaction Strategies

Once the situation is understood and something needs to be done to fix it, it is time for reaction strategies. It is best to have a set of strategies already thought out so that they can be implemented quickly with minimal communication required. This applies to both understaffing and overstaffing situations. Here are some questions to consider when building a set of reaction strategies:

What is the cause of the situation?
How severe is the problem?
What is the impact on service?
How long will it last?
What will it cost to fix it?

Understaffing Strategies

There are some staffing strategies that may be useful for understaffing situations. A relatively simple process involves optimizing the placement of breaks and/or lunch periods. Shifting just a few agents to a somewhat earlier or later break can bring an interval the extra help that it needs. Overtime can be offered for periods that are at risk and part-time shifts can be extended to cover the intervals needed.

Some centers ask supervisors to take calls, but this is a strategy that should be used with care. Not all supervisors are trained to handle the contacts their team handles, and it is important to ensure that help is available to agents if they have a difficult contact during these busy periods. Another strategy to consider carefully is postponing meetings or training. These off-phone periods are needed to communicate effectively and ensure agents have the skills they need to handle contacts quickly and effectively.

Another source of staff might be the answer needed. For example, agents who normally don’t handle calls can be asked to log into the phone queue for a short period. In some centers, agents who have left the center for other jobs in the company may be called upon for short periods of assistance or for overtime opportunities. Of course, routing contacts off-site to an outsource partner is another option to be considered.

It may also be possible to address an understaffing situation by reducing the workload to a level better handled by the number of staff on hand. Perhaps your center asks customers to validate address/phone information on every contact, or maybe marketing is asking for some survey questions to be included in each call. If there is any contact content that is optional, there is an opportunity to leave it out during understaffed periods to reduce the overall workload.

After–call work may be deferable. What is normally done right after the call might be put off to a slower time, but there are risks to this strategy. Error rates tend to increase and some may be forgotten all together.

Where customers can be forced to a self-service option, or where it can be recommended during the queue, it may reduce the load for agents to handle. Shifting contacts to multi-media interactions may also provide a bit more time to react, but it may increase the overall handle time for these contacts, so care should be exercised in encouraging that option.

There are some technological strategies that may help in understaffing as well. Delay announcements that inform the customer of the potential wait can encourage some to call back at a better time. Many centers have implemented systems that offer the caller an option to enter a number for a call back when their turn arrives rather than continuing to wait on the phone. While this can provide an attractive option for the customers, it does not reduce the workload in the current period but ensures that every call will need to be handled (even if the caller might have been willing to hang up and call back at another time or choose another medium).

Routing calls to other staff may also be an option. Where there are multiple centers or outsource providers available, the workload may be overwhelming in one area but not in another. Routing calls to the shortest wait can be helpful for customers but also evens out the workload of the staff. The same is true of skill reassignments. While it may be best to handle some calls in the best skill group where possible, a backup team may be tapped in an overload situation.

Overstaffing Strategies

While it may seem like a gift when there are more staff available than needed, the situation can result in unnecessary cost and low agent occupancy. If the situation is likely to last longer than a short time, redeploying staff to other activities may be appropriate. This may be great opportunity for an impromptu training session, additional coaching, or team meeting. If the situation is expected to last for several hours, it may be appropriate to offer some staff time off (with or without pay).

There may be ways to adjust the workload as well by having agents do non-call work, research, and call backs. That optional content eliminated/delayed in the understaffing situation can now be exploited. Perhaps there is an opportunity to attempt an upsell or do that data verification. Even proactive outbound calling (perhaps to welcome new customers) may be scheduled. Chat support can be more freely offered than normal. If the situation will last several intervals, it may be appropriate to loan staff to other departments – perhaps in exchange for help in other periods. Routing calls to the outsourcer may be reduced as well.

It is good to have a plan in place for both the under- and over-staffing situations. A graduated plan with defined triggers and the process that will be implemented at that time can help everyone to know what to do and when to do it with minimal communication. Some find a color code to be helpful, such as code yellow means a minor issue with some small adjustments while code red means an all-out effort.

Daily Planning Meetings

For many centers, a planning meeting is part of the daily process. There may be one in the early part of the day or several at each shift change. These are designed to be brief and to provide information to the appropriate parties on the “state of the business.” The timing is typically planned when the WFM team will know who is absent and has seen a couple of intervals of actual workload to compare to the forecast.

This is meant to be a “stand-up” meeting if held in person or a brief conference call for dispersed personnel. A few minutes to review any changes that have been seen and what might need to be done to react to them is the primary purpose. If there is information on something that is coming up later in the day that may need adjustments, that should also be included.

The most important communication from such a meeting is the call to action. If there is an adjustment that is needed, the action should be clearly stated and those responsible identified. Follow up communications can provide additional details to those who want them, but this meeting should not be longer than a few minutes.

Analysis for Long-Term Success

Once the day is over, the information gained should be gathered to aid the longer-term processes. These include metrics to improve forecast accuracy by identifying patterns where forecasts are consistently off, and to improve scheduling efficiency by identifying repeatable periods of time where staff levels are not meeting demand. Historical data analysis can also help to improve intraday effectiveness by identifying issues with:

Communication between WFM and Operations
Effectiveness of action plans
Schedule staff vs. average positions staffed
Shrinkage
Adherence
Outsourcer utilization

Summary

Managing the intraday processes in the contact center takes constant monitoring and communication between WFM and Operations personnel. Think of it as learning how to avoid being hit by a bus!

Look – identify any bus about to hit you
React – don’t just stand there
Report – identify and report repeated dangers to authorities to avoid future mishaps

It is a rare forecast and schedule that is exactly right for the workload and staffing that show up on the actual workday. Adjustments are almost always required. The sooner the variances are noticed and action plans communicated, the better chance the center has to provide a consistent customer experience and a reasonable work environment for the staff.

Penny Reynolds was Co-Founder of The Call Center School and is a popular speaker and writer in the area of call center operations. Recently retired, she serves as an Educational Advisor to SWPP, continuing to provide thought leadership and training to the workforce management community. She can be reached at pennyreynolds00@gmail.com or at 615-812-8410.