Implementation Engagement · One Week of Reactive Overtime

How 320 overtime hours get bought one Friday at a time

A current-state exhibit from a capacity planning engagement, told as the GM's week: the schedule built Sunday night from last year's gut feel, the client promo that landed Wednesday with no warning, and the premium hours bought Friday afternoon to protect ship SLAs.

~320
overtime hours bought per week, average across the prior quarter
~10%
of all hours worked are overtime
~15%
of that overtime is planned; the rest is reactive
~27%
weekly volume forecast error, when a forecast exists at all
Day-before
schedule visibility for warehouse staff, sometimes same day

Overtime is not the problem here. It is the shock absorber for a missing forecast. Client order files arrive every week with real volumes, and nobody converts them into labor hours. So the schedule is built from last year plus instinct, the week does whatever it wants, and the gap gets closed with premium hours on Friday. The GM eats the blame; the finance lead sees the premium only after it is spent.

340 staffed hours a day meets 505 hours of Friday demand

Both series are labor hours per day on the warehouse floor. Demand is what the client order files implied, converted after the fact; staffed hours are what the Sunday-night schedule actually put on the floor. Capacity stays flat all week. Demand does not. The red overage is the gap, and the gap is what gets bought back at 1.5x.

Order volume, in labor hours Staffed hours Hours over capacity (bought as OT)
Day Order volume (labor hrs) Staffed hrs Hours over capacity What happened
Monday3253400Schedule from Sunday night holds. 15 spare hours that cannot be banked for Friday.
Tuesday3303400Still fine. Nobody has opened this week's client order files as labor hours.
Wednesday38534045A brand client's promo goes live. First anyone on the floor hears of it is the order feed.
Thursday450340110Pick lines fall behind. Backlog rolls into Friday.
Friday505340165The scramble. Premium hours bought at 1.5x to protect ship SLAs.
Saturday1501700Skeleton crew clears the residue.
Sunday000Closed. The next schedule gets built tonight, from last year plus instinct.
Week2,1451,870320The over-capacity total matches the prior quarter's weekly overtime average.

Illustrative week, constructed to match the client's prior-quarter operating pattern: about 320 overtime hours per week, roughly 85% of it reactive. Hover or tab across the days for exact values; the table carries the same numbers. The point reads in one glance: capacity is flat because the schedule was set before the week was known. Demand is not flat, and the difference is paid for at a premium.

13.5 hours from backlog discovered to lines clear

This is the process that currently decides 85% of overtime. It has no owner, no price check, and no memory. It runs on whoever says yes.

Manual improvisation Pain point
6:00 am

Supervisors see the backlog

First shift walks in to Thursday's unpicked lines. Wednesday's promo volume is now Friday's emergency.

Found 2 days late
9:00 am

The floor walk

Supervisors go line to line asking who can stay tonight. No rotation, no list. Just whoever says yes.

Volunteer roulette
2:00 pm

Temp agency called

Too late for same-day coverage. The agency needs 48 hours of notice the operation never gives it.

48 hrs too late
4:00 pm

Premium OT bought at 1.5x

About 165 hours approved in a hallway to protect ship SLAs. Finance finds out on the payroll run.

165 hrs at 1.5x
7:30 pm

Lines clear

Trucks make the cutoff. The SLA holds. The crew that opened the building at 6:00 closes it at 7:30.

13.5-hour day
After close

The same 15 people carried it

The yes-sayers are always the same 15 names. Two of their coworkers quit last year citing Friday surprises.

Concentrated burnout

$137K of overtime last quarter, and the costs that never hit a ledger

~$137K

OT spend, prior quarter

About 4,160 overtime hours at an illustrative fully-loaded OT rate of $33/hour. Only ~15% of it was chosen in advance; the rest was bought under Friday pressure, at the worst possible negotiating position: 4:00 pm, trucks at 8:00.

Departures last year citing Friday surprises

Exit interviews said the same thing: "every Friday is a surprise." Both came from the group of 15 that always says yes. Recruiting, hiring, and retraining costs sit on top of the OT line, in a different budget, invisible to this conversation until someone adds them up.

Quiet costs nobody invoices

  • Supervisor time: most of Friday morning goes to recruiting stay-overs instead of running lines.
  • Error rates: mispicks climb on hour 11 of a 13.5-hour day, and rework lands on Saturday's skeleton crew.
  • Schedule trust: staff see schedules the day before, sometimes the same day. Nobody can plan a life around that, so fewer people say yes each quarter.

The fix is not banning overtime. It is deciding it on Monday instead of begging for it on Friday. The same 165 hours, chosen four days earlier at chosen crews, is a plan with a price tag. Chosen at 4:00 pm under a truck deadline, it is a tax. The next exhibit shows the model that moves the decision.

Next exhibit: the capacity model

A volume forecast the GM can read, measured labor standards for the six core activities, and a one-page weekly plan: forecast units to required hours to scheduled hours to planned overtime, set in a 30-minute Monday meeting and reviewed every Friday.