Tools
Things you can push on.
An argument you can only read is easy to nod at. These are the same arguments built as instruments, so the mechanism can be handled directly. Every one of them runs on an illustrative model: the mechanics are real, the numbers are not benchmarks and not client data.
Built
Open one and start moving values.
Forecast Reality Check
The demand was not what the plan was told to expect: volume, mix, or handle time arrived different. The miss was born before the first schedule was cut, and no amount of day-of heroics was going to close it.
Name the Gap
A staffing simulator in four parts: run the same well-planned day twice and watch it land differently, take a missed day apart into forecasting, planning, absence, and adherence gaps, watch a managed metric improve while its paired truth degrades, and run the occupancy loop out to sixteen weeks.
Cost calculators
Three calculators for seeing how the bill compounds: the people who left, linked avoidable repeat contacts, and capacity lost to unplanned shrinkage. Put in low, central, and high assumptions (nothing leaves the browser), and the combined view stays withheld until you reconcile both overlap and wage bases.
Just add one person
Service level plotted against agents on an Erlang C queue. The marginal gain from the next agent, and the occupancy the current one is carrying, are both read off the same curve.
The interval wall
The same nine-contact day as the inventory essay's figure, read against a ruler you can move: the model books each contact into the interval where it begins, while the floor holds it until it ends. Slide the interval length and the two readings drift apart, until the ruler is shorter than the work itself and not one contact fits inside a single interval. A companion to "Just add one person": that one slides the agents, this one slides the ruler.
The drift board
The snapshot ladder from the adherence essay, made playable: one simulated day captured five times, from the forecast the schedules were built against to the copy that survives the late corrections. Inject drift between the captures and read what the comparisons say: a shift traded by consent washes out by the day grain, while the meeting that appeared unplanned and the coaching block that dissolved into the queue get named.
Under review
Working drafts.
One figure here has been decided: the same day was drawn two ways, and neither ships. A chart that pairs two unrelated readings invites the join it claims to refuse, however it is drawn, so the groundwork page now makes the point in prose and points at the operator sheet on the WFM page, where the readings share an interval row without being plotted against each other. This preview keeps both drawings for the record. The numbers are simulated, generated by a committed script from an illustrative demand day, not real operational data.
Adherence and investment on one grain
Two figures of the same day, both carrying off-phone investment and adherence to the same interval grain: a dual-axis chart shading the three lowest adherence readings, and a panel build with each series on its own declared scale.
Why these exist
The forecast is a model. The floor is real.
For a few minutes, a simulator closes the gap between a number and the day it claims to describe: you set the plan, you set what the floor did, and the consequence arrives without anyone defending a position. What the instruments keep showing is the same thing the essays argue. The model is not the reality. The floor is the reality, and the plan exists to serve it.