Ask the Workforce Wizard

Question:

I have a few questions about Agent Availability that I was hoping you could answer.   What is the common way to calculate Agent Availability? Should we use agent “log in” time or “scheduled time” as part of that calculation?  And should we exclude breaks, lunches, etc.? Thanks!

Answer:

Agent availability is calculated differently in centers that use the statistic.  It is often used along with or instead of agent occupancy. While occupancy typically measures the amount of time the agent is directly involved in handling contacts including the after call work state, availability may focus on that time plus any time the agent was idle waiting for work.

Some use availability to measure the time the agents are logged in and either serving callers or involved in some other customer-facing activity (such as emails, chats, etc.) when they would be logged out of the phone queue but in a defined work state.  Still others are focused on whether the center “got a day’s work for a day’s pay.” The result is typically a percentage of the total time that the agent is scheduled (if non-call work is involved) or logged in time (if only call handling time is considered).  Time that is scheduled for breaks and lunches would generally not be included in any availability statistic as the agent will not do any work for the customers during those periods.

The question is what do you want to know?  Are you looking for idle time that might be filled with some other work?  Do you want to measure the total amount of time the agent is serving customers in any medium? Are you trying to determine if the agent put in a full day’s work?   Each center does its own calculation to focus on the information desired.

Question:

Can you please share your thoughts on service level definitions and the treatment of abandoned calls?   In a prior environment, we removed short abandons from the denominator of the service level calculation.  In my new job, I am finding alternate calculations in use and am working on a best practice recommendation.

Answer:

In general, the treatment of abandons in the service level calculation is open to interpretation.  The SWPP StaffSmart Calculator assumes that all abandoned calls are included in the offered calls. Of course, if some or all abandoned calls are deducted, the service level percentage will rise accordingly.  But is it really telling the center what it needs to know to manage service and customer expectations? If only 1% or 2% of calls abandon, it won’t make much difference to the final numbers, but if the abandon rate is say 10%, the adjustment will make the service look much better than the customers probably perceive.

An analysis of the abandons to determine how long they held before abandoning would be very useful in determining how they should be handled in the calculation.  (Most ACDs offer this type of report.) If the abandons have held on past the service level goal before abandoning, the customers probably got tired of waiting and to ignore this in the calculation of service level could be deceptive.  However, if the abandons occur in the first 5 seconds, for example, then they might be more safely ignored.

The real question is “what do you want to know”?  If you are trying to find ways to make the service level result look as good as possible to senior management (especially if the center is habitually understaffed), then ignoring the abandoned calls can help do that.  But if you want to know the percentage of callers who actually got served in the goal time, then including the abandons might make more sense. It gets hard to convince management that more staff are needed if the service level goals seem to be met because the calculation has not included customers who were not served at all.

There is also the in-between option, which would be to ignore abandons that occur in some specific number of seconds (typically below the goal or at just a few seconds).  This would clear out calls that the center really had little chance to answer but keep the ones who gave up after waiting too long.

There is also the issue of repeat callers who may abandon one time and then call back and be served another.  There is no easy way to determine how many do that, but if the abandon rate is low, it can probably be ignored.  When it is high, the repeat callers may be significant.

There really is no easy answer on this and there are many ways of looking at how the calculation should be done.  It is, however, important to look at the calculations done in each of the systems in use in the center to make sure they are synchronized.  It can be challenging to deal with a WFM system that uses one method while the ACD uses a different one, for example. Some reporting systems can be adjusted so that all calculations match, but not all vendors offer that flexibility.  It may be necessary to adjust all the others to the one that cannot be changed. Alternatively, the data might be exported to another tool and perhaps adjusted there.

Ask the Workforce Wizard

Question:

I have a few questions about Agent Availability that I was hoping you could answer.   What is the common way to calculate Agent Availability? Should we use agent “log in” time or “scheduled time” as part of that calculation?  And should we exclude breaks, lunches, etc.? Thanks!

Answer:

Agent availability is calculated differently in centers that use the statistic.  It is often used along with or instead of agent occupancy. While occupancy typically measures the amount of time the agent is directly involved in handling contacts including the after call work state, availability may focus on that time plus any time the agent was idle waiting for work.

Some use availability to measure the time the agents are logged in and either serving callers or involved in some other customer-facing activity (such as emails, chats, etc.) when they would be logged out of the phone queue but in a defined work state.  Still others are focused on whether the center “got a day’s work for a day’s pay.” The result is typically a percentage of the total time that the agent is scheduled (if non-call work is involved) or logged in time (if only call handling time is considered).  Time that is scheduled for breaks and lunches would generally not be included in any availability statistic as the agent will not do any work for the customers during those periods.

The question is what do you want to know?  Are you looking for idle time that might be filled with some other work?  Do you want to measure the total amount of time the agent is serving customers in any medium? Are you trying to determine if the agent put in a full day’s work?   Each center does its own calculation to focus on the information desired.

Question:

Can you please share your thoughts on service level definitions and the treatment of abandoned calls?   In a prior environment, we removed short abandons from the denominator of the service level calculation.  In my new job, I am finding alternate calculations in use and am working on a best practice recommendation.

Answer:

In general, the treatment of abandons in the service level calculation is open to interpretation.  The SWPP StaffSmart Calculator assumes that all abandoned calls are included in the offered calls. Of course, if some or all abandoned calls are deducted, the service level percentage will rise accordingly.  But is it really telling the center what it needs to know to manage service and customer expectations? If only 1% or 2% of calls abandon, it won’t make much difference to the final numbers, but if the abandon rate is say 10%, the adjustment will make the service look much better than the customers probably perceive.

An analysis of the abandons to determine how long they held before abandoning would be very useful in determining how they should be handled in the calculation.  (Most ACDs offer this type of report.) If the abandons have held on past the service level goal before abandoning, the customers probably got tired of waiting and to ignore this in the calculation of service level could be deceptive.  However, if the abandons occur in the first 5 seconds, for example, then they might be more safely ignored.

The real question is “what do you want to know”?  If you are trying to find ways to make the service level result look as good as possible to senior management (especially if the center is habitually understaffed), then ignoring the abandoned calls can help do that.  But if you want to know the percentage of callers who actually got served in the goal time, then including the abandons might make more sense. It gets hard to convince management that more staff are needed if the service level goals seem to be met because the calculation has not included customers who were not served at all.

There is also the in-between option, which would be to ignore abandons that occur in some specific number of seconds (typically below the goal or at just a few seconds).  This would clear out calls that the center really had little chance to answer but keep the ones who gave up after waiting too long.

There is also the issue of repeat callers who may abandon one time and then call back and be served another.  There is no easy way to determine how many do that, but if the abandon rate is low, it can probably be ignored.  When it is high, the repeat callers may be significant.

There really is no easy answer on this and there are many ways of looking at how the calculation should be done.  It is, however, important to look at the calculations done in each of the systems in use in the center to make sure they are synchronized.  It can be challenging to deal with a WFM system that uses one method while the ACD uses a different one, for example. Some reporting systems can be adjusted so that all calculations match, but not all vendors offer that flexibility.  It may be necessary to adjust all the others to the one that cannot be changed. Alternatively, the data might be exported to another tool and perhaps adjusted there.