The two methods above are generally acceptable for relatively near-term planning: the first method is roughly fine for budgeting and plans up to a year out while the second has real strength over periods of up to 3 months. Now we’ll turn our attention to a system that is robust for very long-term planning and allows for some strategic change to be embedded in the calculations.
But first: If I employ 1,000 advisors but only need 100 to run my call centre, how many desks do I need? 1,000? 100?
Strictly speaking, the answer is probably less than 100. If you don’t need all those other people to deliver the level of service you’re aiming for, why go to the expense of giving them a desk? You might as well send them home. (You might also wish to consider sending some other people home too, such as the guy who recruited them or the Marketeer who said you’d need them.)
This suggests that the number of desks needed might be driven less directly by the number of people we employ.
We start by making one simple observation: in any given time period (day, week, month), the maximum number of desks required equals the maximum number of agents required to deliver the required service level in the busiest half hour.
Assuming we have long-range call and AHT forecasts, we can use our current analysis of peak resource requirement to find the busiest half hour in every month into the future for as far as we have forecasts for. All we need to do is use our spreadsheet add-in of
agents(80%, 30, busiest_hour_volume, AHT_in_that_hour)
and we will have the number of desks required. (Strictly speaking, rather than choose the hour with the highest call volume, it is better to choose the hour for which calls x AHT is at a maximum.)
Given (say) a strategic desire to have customer service staff cross-selling products, we can look not only at the increase to operational expenditure but also consider the capital implications of increasing our infrastructure.
Equally, the impacts of introducing new systems or even changing the target service levels can now be measured in capital, as well as operating, cost terms.
If your call centre has a number of call types, each with a different intra-day shape to call flow and AHT, there are two ways to apply this method. First, you can apply it to each call type separately and simply add up all of the maximum seating requirements; this is likely to over-state your needs. Alternatively, you might be able to merge the call types (assuming no desktop system constraints and operational flexibility) to build an aggregate intra-day pattern and AHT from which you can build up a more comprehensive and accurate figure. This is particularly important for outsourcers.
There are, however, a few amendments we need to make to our basic model before we can put it into production:
- First, an allowance needs to be included for advisors on paid breaks (or any planned shrinkage that ties up a desk). You will have scheduled more people to work than are required to deliver the service level because you know that some of them will not be taking calls even when occupying a desk.
- It is also to be recognised that at the peak requirement times there is often an over-lapping of rotas; this means that you generally do have to seat more people than you strictly need. An allowance equal to the shift efficiency as defined elsewhere on this site usually works out about right.
- Thirdly, don’t forget to allow your first line management a desk, unless it is your policy to require them to stand during the busiest periods.
If your business ever has to contemplate expansion involving the potential occupation of a new site, this is the method to use. I have used exactly this method to demonstrate that a new building was not required, even though previous analysis had suggested otherwise. Take it from me, you have to be just as sure when you recommend no expansion as you do when you say it’s time to grow.