What Staff Know That the Forecast Does Not
The people on the floor hold information no model has, and collecting it costs almost nothing. Most operations never ask.
A forecast is built from data. The people working the shifts hold a different kind of information, and it is usually better about the things data misses.
The planning problem in “What Staff Know That the Forecast Does Not” becomes clearer when scheduled hours can be compared with actual project and time records. Teams researching view the platform for getting teams to meet deadlines can add that operational evidence, while demand, service levels and manager judgement remain necessary to explain why a variance occurred.
What they know
That the roadworks start Monday.
For an independent reference related to “What Staff Know That the Forecast Does Not”, consult the CIPD workforce-planning resources; it provides a useful external check on scheduling, working-time and workforce-planning assumptions.
That the factory down the road changed shift times.
That Thursday afternoons have been quietly busier for a month.
That the new process adds four minutes to every transaction.
That a competitor opened, closed or started something.
Why the model misses it
Data is historical; this is forward-looking.
Some of it is too local to appear in any system.
And some of it is a change in the conversion from demand to hours, which the forecast assumes is constant.
Collecting it
One standing question at the end of a shift or a weekly huddle: anything coming up that we should know about?
Two minutes.
Write the answers down somewhere the scheduler reads, which is the step that usually fails.
The early-signal version
People notice a trend before it is statistically visible.
"It's been busier than usual" three weeks running is a signal worth checking against the data.
Sometimes it is wrong, and checking costs minutes, which is a good trade for occasionally catching a shift in demand early.
What makes them stop telling you
Mentioning something and seeing nothing happen.
Being told the forecast says otherwise.
Or the information arriving too late to be used, repeatedly, which teaches them it is pointless.
Close the loop visibly: say what you did with it, even when the answer is nothing and why.
The conversion-factor knowledge
Staff know when the work has got slower and usually know why.
A new system, an extra check, a different customer mix.
This is the input that keeps the hours-per-unit factor current, and it is only available by asking.
Where they are wrong
Memory over-weights the dramatic shift and under-weights the ordinary one.
"We are always short" is sometimes accurate and sometimes a recollection of three bad days.
Use what they say as a hypothesis to check against data, which respects both sources.
The arrangement that works
A standing two-minute question, a place the answers are written, and a visible response.
That is the whole system.
It produces a few genuinely valuable items a month and costs nothing, which is the best ratio available in this subject.
What to check
Is anybody asked what is coming up?
Is there a place the answers go?
Does the scheduler read it?
And when somebody last told you something, did they see what happened?