Staffing sensitivity — the what-if tool
A forecast is a single number; reality arrives as a range. This tool takes one interval’s assumptions and shows how the agents you need move when volume, AHT, your service target or shrinkage come in wrong — so you can see which error hurts most, and size your buffer against the one that does.
Tornado — which error moves the roster most
Each bar is the swing in gross agents if that one input is wrong by the amount shown, holding the rest at base. The widest bar is your biggest planning risk.
Sensitivity grid — volume × AHT
Gross agents for every combination of a volume miss (columns) and an AHT miss (rows). The two compound: a busy day with a heavy mix is the top-right corner, and it is a long way from the centre.
Why this matters
Planners are taught to chase forecast accuracy, but not every error costs the same. Volume and handling time usually have roughly equal leverage — a 10% miss on either moves the roster about the same amount — because both feed straight into offered load. Shrinkage is often the quiet third lever: a five-point miss can cost as many heads as a sizeable volume error, and it is the one most likely to be set from a stale policy figure rather than measured. The service-level target, by contrast, usually moves staffing far less than people expect, which is why arguing about 80/20 versus 80/30 is rarely where the real money is.
The practical use is buffering. Once you can see which input your plan is most exposed to, you size your contingency — and your forecasting effort — against that input, instead of spreading worry evenly across all four. That is the difference between planning with numbers and planning with judgement about the numbers.
Maths: Erlang C, the engine behind the Erlang C calculator. Gross = net ÷ (1 − shrinkage). See As Much Art as Science for the thinking behind it.