WFM stands for Workforce Management, but what does that mean?
In a nutshell, it’s the job of getting the right agents, with the right skill in the right place at the right time. Workforce Optimisation is another term, which just takes the planning of your resources (staff) to another level, but in essence it is all Resource Planning.
If you are asking the question, then you are probably working in a call centre that hasn’t invested six figure sums in purchasing software to forecast how many calls are going to arrive each 15-minute interval of every day, automatically schedule flexible shifts for agents to work to meet demand, and track what those agents are doing during their shift. That said, you are probably doing all these tasks in some shape or form, but perhaps it is time for you to take a breath and think about how you approach the job of Resource Planning, and whether, before long, you may benefit from acquiring a cloud-based solution to make the job easier, such as u-wfm.
The WFM function can initially be divided into 3 functions:
- Real Time
It is the glue that binds together the 3 main perspectives of call centre management:
- the customer
- the staff
- the bottom line
Good resource planning will ensure that there are enough agents on the phone to deliver a good customer experience. This in turn balances the workload of the agents, so they should neither be too quiet or too busy. The result being an efficient operation, with costs managed effectively, good customer retention/satisfaction and a happy workforce.
It is often said that WFM is both and art and a science, mathematics and algorithms play a major part in a sophisticated forecasting model, but without the common sense and business knowledge of someone who understands your company and how it operates, this will fall short of the accuracy you need to run an efficient call centre. The human element is as important as fancy footwork in Excel or WFM software, and that is where the art comes into it.
In its simplest form, Forecasting is analysing historic call volume and Average Handle Time (AHT) data to predict the call arrival pattern in the future. You can produce a forecast with simple school maths, which is a good start. If you have a degree in statistics or a black belt in MS Excel, then you can really go to town on your analysis and predictions, or better still purchase a WFM software application, which will do all the hard work for you, leaving you time to apply your business knowledge to the outputs.
Once the call volumes and AHT have been predicted by 15 or 30-minute interval for the period in question, then you need to calculate the agent requirements using Erlang C in a simple environment, where all agents answer any call that arrives, then you need to apply Shrinkage to ensure you have an overall agent requirement figure against which to schedule. Erlang C? Shrinkage? I hear you ask.
Erlang C is an algorithm, which you can have as an add-in in Excel, or is inbuilt into WFM software. Google away if you want all the technicalities, but in essence it is a formula used to factor in the random arrival of calls during the period being forecast, which is key to an accurate agent requirement figure. The formula makes certain assumptions, such as the caller will be answered on a first come, first served basis by the longest available agent, that there is no cap on the number of calls and that there are no dramatic variations in calls offered during the interval being forecast. This is why it is used extensively in forecasting and has been for nearly 60 years. But …. there’s always a but …. it also assumes “infinite patience”, that is to say the caller holds on until they get answered. In the 1940’s people didn’t abandon calls willy-nilly like they do today, so the algorithm has been modified by most software developers to factor in an abandoned call rate. This is perhaps where your Excel workbooks may fall a bit short of what you need. If you have a multi-skilled and/or multi-media environment, then you do need to look at options for simulation, as Erlang C won’t cut it.
I haven’t forgotten I used the term Shrinkage earlier. In simple terms, this is the percentage of paid agent hours you lose to offline activity including training, communication, absence etc. Any unplanned offline activity needs to be factored into agent requirements so that on the day you end up with the number of agents taking calls that you need to meet service levels. Again, this topic can be explored in depth, but suffice to say you need to be tracking the amount of non-phone hours there are in your centre every day, to ensure you are calculating the correct agent requirement figures to meet service level, otherwise you are always going to be on the back foot.
A good forecast will produce a smooth intraday (15 or 30-minute interval) agent requirements, ie a graph without major peaks and troughs.
Once the forecaster has produced an intraday agent requirement, in an ideal world, the scheduler will create shifts that ensure there are sufficient agents to answer the calls throughout the schedule period. They will also plan in the offline activity during periods where there is some overstaffing, if you have that luxury, but at least avoiding periods of understaffing.
If the graph has peaks and troughs, then they are in danger of scheduling a lot of breaks or other offline activity during a half hour period that looks as though it’s going to be really quiet, that could just end up being particularly busy. A smooth profile on the graph will ensure that offline activities are balanced throughout the day.
The schedule fit (match to the agent requirement graph) is easier to achieve if your agents have flexible contracts, so they can be scheduled to start at different times and possibly have different shift lengths and rest days.
You have the forecast and the schedules, so now it’s time to deliver that customer service. Real Time is all about managing the day in real time, monitoring the accuracy of the forecast, adjusting the scheduled activities in the event of unexpected busy/quiet periods and monitoring the Team Leader management of the agent activities – are they adhering to their schedules? are their handle times, in particular wrap (ACW, after call, post call, not ready – there are lots of different terms for this) within acceptable tolerances?
The outputs of WFM should ensure that the customer experience is what your organisation aspires to. Customers need to be satisfied with the speed and quality of response. This is down to having the right number of agents with the right skills available when they call.
The outputs of WFM should ensure that your agents have schedules that give them a work-life balance, with no shifts they dread because they are going to be mad busy or bored to death. Balancing the workload through the week is important to avoid any perceived awful shifts to work. If agents are bored or over worked, this will ultimately impact your sickness and attrition figures.
The outputs of WFM should ensure that Senior Management have an efficient operation, with happy staff and good customer retention/revenue/satisfaction figures.
As previously mentioned, it’s an art not a science. Balancing the 3 perspectives is difficult and getting an accurate forecast and good schedule fit is harder still. In a small single skilled centre you can get by with MS Excel and good communication, but ultimately your operation will grow in complexity and there will come a point where you will need to consider a software solution to free up your planner’s time to analyse the data and advise the operation on where changes may be needed, rather than simply producing more of the same.
Additional Challenges & Useful Tools
Long Term Holiday Planning
Planning for vacation over long periods of time can be a difficult task to complete. To achieve a quick and easy calculation of how and where to allow agents to ‘spend’ their vacation time, U-WFM are introducing a calculator that will allow this operation to be completed in record time.