Who Should Log Actuals: The Operator or the Planner?

EDGEBIC by User Solutions gives you two doors into the same actuals record: the kiosk at the machine and the Log Actuals dialog on the planner's desk. Choose one as primary per work center and use the other as a fallback, because both write to the same day grid and the last save wins. The question of who logs production actuals looks like an org chart question. It is really a data integrity question, and the shops that never answer it end up with numbers nobody trusts.

This is not a question about capability. Both doors work, both are supported, and both feed the same reschedule. It is a question about ownership. When two people can write the same cell and neither has been told they own it, the cell eventually holds whichever value was typed last, and no one can say which one was true.

Two Doors, One Record

An actual in EDGEBIC is stored against an operation: a work center, on a job, on a calendar day. That record holds an actual start, an actual end, and a per day figure for hours and pieces. It does not remember which application wrote it.

The kiosk builds that record out of punches. Every tap opens or closes one uninterrupted stretch of setup, run, pause, or down time, and the timestamp is taken when the tap lands. The first Start Run stamps the operation's actual start; Complete Operation stamps the actual end and rolls the run time up into daily hours. Good and scrap taps become piece counts, and every pause and every scrap piece carries a reason from four categories. See how the kiosk captures a production punch for the mechanics.

The planner builds the same record by typing. Edit Actual Dates sets the start and end, with Start = Now and Use Sched Start as one-click options. The Log Actuals grid then shows one row per calendar day of the operation's window, with Sched Hrs beside Actual Hrs and Sched Pcs beside Actual Pcs, so the planner types the real figure next to what the plan expected. See logging actuals from the planner.

Both roads end at the same row. That is the design's strength and the reason ownership matters.

What Each Door Is Good At

The asymmetry worth noticing is in the detail row. Typed totals cannot be decomposed. If an operator types eight hours for Tuesday, nothing in that number tells you that ninety minutes of it was a changeover and forty minutes was a coolant alarm. Punches tell you, which is why setup variance and pause analysis only exist for the machines that punch. See how setup phases become setup variance.

The Case for the Kiosk as Primary

Put the kiosk first where the answer to "what actually happened" is contested or expensive.

That usually means the bottleneck, the machine with the long changeovers, and the machine everyone blames when a job is late. Those are exactly the places where a day total settles nothing and a phase breakdown settles everything. They are also the places where the reschedule is most sensitive, because the plant's dates flow through them.

The kiosk's other advantage is timing. Rescheduling plans the remaining work forward from your last logged position, so the value of an actual decays with age. A punch is current by construction. A typed entry is current only if somebody remembers to type it, and the honest answer in most shops is that they remember on Friday. See the daily rhythm of a kiosk-run shop.

The Case for the Planner as Primary

Planner entry is the right primary more often than kiosk evangelists admit.

A work center that runs one job for three days does not benefit much from tap-level detail; one grid row per day carries almost the same information. A work center with no power outlet, no wall space, or an operator whose hands are never free is a bad terminal site whatever the theory says. And a shop that is still establishing the habit of reporting anything at all is better served by one planner typing a short list every morning than by five terminals nobody touches.

Planner entry also owns a set of actions the kiosk deliberately does not have: correcting a previously reported value with a reason, reviewing days that carry the auto-filled badge, and completing or reopening a whole job. Those are judgment calls, and they belong at a desk.

Why "Both, Whenever" Is the Wrong Answer

The failure mode is not dramatic. It is quiet.

Concurrency in EDGEBIC is last-write-wins. There is no merge dialog and no conflict prompt on a daily actuals value: the most recent save is what the record holds. So if an operator punches Monday as 6.0 hours and a planner, working from a shift report that rounded, types 5.5 for the same Monday, Monday is 5.5. Nobody is warned, because from the system's point of view a planner correcting a value is a completely normal thing to do.

There is one guardrail, and it is narrow. An actual start is written once: if a planner has already recorded a start, a later kiosk punch will not overwrite it, and the reverse holds too. See why an operation's actual start is written once. Daily hours have no such protection, and they should not, because correcting them is the whole point of the Log Actuals dialog.

The other cost of an undecided owner is duplicated effort that looks like diligence. An operator punches the job, then a supervisor types the same day into a spreadsheet, then a planner types the spreadsheet into the grid. Three records of one shift, two of them redundant, and the one that survives is whichever was entered last.

Choosing By Work Center, Not By Plant

The decision does not have to be uniform, and it is better if it is not. Assign a primary door per work center and record it somewhere the whole team can read: a column in your work center list, a line in your planning standard, a note on the terminal itself.

A realistic split for a twelve-machine shop looks like this: kiosk primary on the three cells with heavy changeovers and the assembly bottleneck, planner primary on the eight machines that run long jobs, and the paint booth on planner primary because the operator wears gloves all shift. Nine of twelve never see a terminal, and the data is better than it would be with twelve terminals half-used.

Then write the fallback rule, which matters as much as the primary. On a kiosk-primary machine, the fallback is Manual Entry at the terminal, or the planner's grid the next morning if the shift ended without it. On a planner-primary machine, there is no fallback to write, only a deadline: today's hours are in before tomorrow's schedule run.

A Worked Decision

CNC-Mill-1 is the shop's contested machine: six changeovers a week, and every late job argument ends up pointing at it. It gets a terminal, and the kiosk is primary. The planner's role there is review only: each morning she opens the Job View, checks for auto-filled days, stamps any actual end the floor forgot, and looks at the punch history for pauses worth a maintenance conversation.

Paint-Booth-1 runs one color per day and the operator's hands are wet. No terminal. The planner is primary, and the booth supervisor drops a note on her desk at shift end with hours and good pieces. She types two grid rows a day, which takes ninety seconds.

Saw-1 is somewhere in between: short jobs, quick setups, low stakes. It shares CNC-Mill-1's terminal through the Switch WC button when the operator is willing, and falls back to planner entry when nobody bothers. That is an honest description of a low-stakes machine, and writing it down beats pretending it is kiosk primary.

Three machines, three answers, one document. Nobody wonders who owns Monday.

The takeaway

The two doors into EDGEBIC's actuals record are equally real and equally supported, which is why the choice of primary has to be made by you rather than by the software. Decide per work center, write it down with its fallback, and remember that daily hours are last-write-wins: an undecided owner does not produce two records, it produces one record of uncertain origin. Explore the platform at EDGEBIC, see the upgrade path in RMDB to EDGEBIC, and read on in closing the loop from shop floor to plan and who may correct a kiosk entry.