A Day in the Life of a Planner: Morning Review to End-of-Day Commit

A planner's day has four fixed points: a 20 minute morning review, targeted responses to interruptions, an afternoon capacity check, and one end-of-day commit the shop runs on. This walkthrough follows a single Tuesday in EDGEBIC by User Solutions from 07:40 to 16:20, with real jobs, real interruptions, and the decisions each one forced. It is one of the worked examples we publish so the rhythm is visible rather than described. A planner who has not built that rhythm yet should start with a new planner's first week, from first sign-in to a committed schedule.

The shape matters more than any individual click. A planner who reschedules continuously produces a plan nobody can work from. A planner with a rhythm produces a document the floor trusts.

07:40, Read Yesterday Before Touching Today

Ridgeline Metalworks runs 41 open jobs across 14 work centers. The planner arrives before the day shift settles and opens the schedule.

First question: what did the shop actually do yesterday?

Operators log hours on the kiosk against each operation as they run it, so yesterday's reality is already in the system. Open the affected jobs and compare planned against logged. Three things to look for:

• Operations that logged more hours than planned. These will push downstream work.
• Operations that were supposed to start yesterday and show nothing. These are either behind or unlogged, and those are very different problems.
• Operations marked complete early. These free capacity that other jobs can now claim.

Today's read: J-4471 logged 9.5 hours against a planned 8 on the mill (behind), J-4455 shows no punches at all on an operation that was due to start (a supervisor confirms it did run, the punch was missed), and J-4438 finished a day early.

The missed punch matters because the engine treats logged hours as the truth. An operation with no hours reads as not started and the next reschedule will move all of its work. Getting yesterday's punch corrected before the morning run is 90 seconds now or an hour of confusion later, and the recipe is how to enter actual hours for yesterday.

07:55, The Morning Run

With actuals correct, run the scheduler across the plant.

This run does one thing that matters above everything else: it reconciles the plan with reality. Completed work is never moved, so every operation with logged hours keeps its slot exactly as the operators recorded it. Only unfinished work is placed into new time slots, and half-finished operations schedule only their remaining hours, computed as planned minus logged.

J-4471's overrun ripples: its grind step moves out half a day, and the job behind it on the grinder pulls in to fill the gap. J-4438 finishing early releases 6 hours on the weld cell, and the next weld job takes them.

Total elapsed: about four minutes.

08:10, The Late Jobs List

Run the Late Jobs report. Today it shows four jobs whose planned end exceeds their due date.

Sort by days late:

Two of these are new consequences of yesterday's mill overrun. That is the number that decides whether the morning needs action: two jobs one day late, both caused by a single operation running long, is a normal Tuesday and not an emergency. If eight jobs had appeared, the cause would be structural rather than incidental.

Note them and move on. The 20 minute review is over at 08:12.

09:20, Interruption One: Sales Wants a Date

A salesperson wants a delivery date for a possible 120 piece order. Nothing exists yet: no order, no commitment.

This is a quote simulation, not a schedule change. Run a what-if promise date against current capacity without committing anything to the plan. The simulation reads the same finite calendars the real scheduler does, so the answer accounts for the 41 jobs already committed rather than pretending the plant is empty.

Answer back in six minutes: earliest honest ship is three weeks out, or two weeks if the weld step routes to the pool instead of the single cell. Nothing in the live plan moved. The mechanics are in how to simulate a delivery date for a quote.

10:45, Interruption Two: A Machine Goes Down

The mill throws an alarm. Maintenance says six hours minimum, probably the rest of the shift.

This is the one interruption that earns an immediate run, because the plan is now actively wrong and supervisors will be standing around.

1. Block the day on that one work center's calendar. That zeroes the mill's capacity for today without touching any other station.
2. Scope the reschedule to the jobs whose remaining work sits on the mill. Every other job's reservations stay in place.
3. Read the result.

Three jobs are affected. Their logged hours stay frozen exactly as recorded, and their remaining mill work moves to tomorrow. Two of the three still make their due dates. The third goes one day late and joins the list.

Elapsed: eleven minutes, including walking to maintenance. The full version of this journey is a machine breaks at 10 AM.

12:30, Interruption Three: A Customer Pull-In

A customer wants J-4425 a week earlier.

This one does not get an immediate plant run. Type the new date into the job's Due Date cell, click Save Changes, and tick the job for the end-of-day run. An inline edit changes the promise, not the plan: no bar moves until the scheduler runs.

That is the important discipline. The answer to the customer can wait four hours, and batching this change into the evening run means it gets evaluated against the day's other changes rather than against a plant state that is about to be superseded anyway.

If the answer genuinely cannot wait, scope a single-job run and read Days Late, as in handling a customer pull-in request end to end.

14:00, Interruption Four: A Priority Argument

Two department managers both want their job first on the weld cell.

Open the Drive Schedule grid and look at the float. One job has four days against its due date, the other has one. Set the priorities accordingly (bands of 10 leave room to insert later), click Save Changes, and stage it for the evening run.

The argument took two minutes because the grid answered it. Neither manager was arguing about who deserved the machine; they were both arguing without knowing which job could absorb the wait.

15:10, The Afternoon Capacity Check

Once a day, look forward rather than backward. Run the Work Center Utilization report for the next four weeks and read three numbers from the KPI strip: plant utilization, overloaded count, underloaded count.

Today: 74 percent plant, one overloaded station, four underloaded. One overloaded station in a four week window is worth 10 minutes of attention, not an alarm. Drill into its backlog, confirm nothing anomalous, note it for Thursday.

This is the habit that stops next month's crisis. A station that reads CRITICAL three weeks out is cheap to fix. The same station reading CRITICAL on the day is overtime.

16:00, The Commit

Everything staged during the day now gets evaluated together in one run:

• The pull-in date change on J-4425
• The two priority edits from the weld argument
• The mill's blocked day, already applied
• Yesterday's corrected punch

Run the plant. Read the late list one more time: five jobs, two of them new today, all with known causes and named owners.

Then issue the paperwork by exporting the committed plan to Excel and printing the operator dispatch list. From this moment until tomorrow morning's run, today's committed plan is not rewritten. That single rule is what makes the printed sheet worth reading, and it is why supervisors stop working from Monday's paper.

The Day, in Time

One hour of screen time across a nine hour day. The rest of the day is where a planner is actually useful: on the floor, on the phone, and in the conversations the numbers made possible.

What This Walkthrough Proves

1. Actuals come before the plan. A run against yesterday's incorrect data produces today's incorrect plan.
2. Not every interruption is a scheduling run. Three of four today were staged as edits and resolved in one evening run.
3. Targeted runs beat plant-wide churn. Scoping the breakdown response to three jobs left 38 others untouched.
4. The forward check is the cheap one. Ten minutes looking four weeks out prevents the overtime call three weeks from now.
5. A fixed commit time is what earns trust. The plan is stable for the day it covers, which is the only reason anyone works from it.

For the weekly and monthly versions of this rhythm, month-end capacity review and scheduling a five job week with mixed priorities are the companion journeys.