Production data is separating digital plants

# Production data is separating digital plants Two plants can own the same press. The sheets from both can look good on a tour. One of them can tell you what an order cost, where the waste went, and what the click was for. The other owns a fast machine and a mess. This is

Production data is separating digital plants

Two plants can own the same press. The sheets from both can look good on a tour. One of them can tell you what an order cost, where the waste went, and what the click was for. The other owns a fast machine and a mess. This is about the second plant. The machine isn't the difference. The difference is that nothing in the building is tied to the ticket, so the ticket can't explain the month.

They bought well. The device images cleanly when the file is clean and the stock is the stock they think it is. Visitors get the demonstration, and the demonstration is honest as far as it goes. A sheet comes off. The colour is decent. The minder's pride is real. Then you ask what yesterday's orders cost, and the answer is the price list. Substrate, a quantity band, a finishing line, a number that was typed into the estimate before anyone had opened the file. The list doesn't know that one of those orders spent the morning in preflight. It doesn't know that another was imaged twice. It doesn't know that the folder spoiled the first attempt because the lay cracked on the fold. All of that happened. None of it has a home. At month end the accounts show a margin that isn't the margin the quotes promised, and the gap is a single lump. Nobody can point to the orders that made it. They can point to the press. They do.

The quote and the order never meet

The estimating habit is the root, and it feels like discipline. There's a rate for the stock. There's a rate that is supposed to cover the click. There's a line for cutting or folding that assumes a normal make-ready and a normal spoil. A normal job is an idea the list needs in order to be a list. The plant's actual jobs aren't normal in the same way twice. One file sails through. The next is a rescue. One fold is a single score on an uncoated sheet the folder already knows. The next is a heavy lay on a stock that cracks unless someone changes the score, and the first sheets teach that lesson into the bin. The quote was the same shape for both. The plant discovers the difference only as a feeling that "folding wasn't great this week".

Because the quote doesn't get revised by what happened, the people on the floor stop believing it, quietly. They'll still run the job. They won't tell the estimator that this class of work always reprints, because there isn't a place to put that sentence where it would change the next quote. The estimator keeps using the list, because the list is what they have, and the floor keeps absorbing the difference, because the floor is where the difference shows up. Over a month the easy jobs subsidise the rescues, or they don't, and if they don't the lump in the accounts is the only signal. A signal that arrives as a lump can't be used. You can't fix a lump. You can fix an order, if the order is allowed to be wrong in a way you can see.

The buyer doesn't see this at all. They see a price and a delivery. When the delivery slips, or the colour takes two rounds, they don't think "their data model is thin". They think the plant is unreliable. The plant, which has been unreliable in a specific, recordable way, answers by working later. Later is not a record either. It's a mood the building gets into, and then a wage bill.

Waste that never meets the ticket

The waste is visible. That's the insult. There's a bin at the press and a pallet bin by the delivery, and both of them are doing a job the management accounts aren't. Start-up sheets. A reprint because the imposition was wrong and the pages ran in the sequence the file had, which wasn't a signature. A reprint because the logo on the approved proof was the old logo, and the file that was imaged was whichever file was on the operator's desktop. A stack that blocked because it was packed warm, or because a coating hadn't settled. Spoils from the folder. All of it is product that was paid for, in stock and in imaging, and that nobody can sell.

The press counter still climbed. If you trust the counter, the shift was busy and successful. The counter counts images. It doesn't count copies you can invoice. A shift can be full of images and poor in sellable work, and the second plant will describe that shift as a good one because the device was running. Running is the easiest thing to see. Sellable is a comparison between what was imaged and what left, and that comparison requires someone to have written both numbers against the same job. They don't. The bin is emptied. The counter is admired. The argument about productivity is an argument about the counter, which means it's an argument the mess will always win.

Nobody is hiding the bin. There's no theatre of deceit. There's a habit of treating waste as a housekeeping issue, a thing you clear so the next job can stand at the machine. Housekeeping doesn't ask why this job contributed more than that one. Without the why, the same imposition gets rebuilt badly next week, the same warm stack gets packed the same way, and the same desktop file gets imaged because that's where the last person saved it. The waste is a lesson. The plant throws the lesson out with the sheets and keeps the method.

The click arrives as a bill

The device has a counter, and the supplier has an invoice. In the second plant the invoice is overhead. It gets spread across the month in a way nobody could reconstruct, or it gets turned into a flat figure per sheet that was copied into the price list and then forgotten. It isn't attached to the order. So the order that was imaged, rejected, and imaged again doesn't appear as two clicks. It appears, if at all, as a late job. The click paid for both, including the one in the bin. The price list charged for one.

Coverage doesn't enter the conversation. A leaflet with light text and a poster that's nearly solid don't pull on the device in the same way, and on many digital processes they don't cost the same way either. The second plant's list treats a sheet as a sheet. The heavy sheet is the one the minders groan about and the estimators can't see, because groaning isn't a field on the ticket. Over time the plant becomes strangely keen on the light work and strangely surprised by the heavy work, without ever having decided to be. The invoice explains the surprise once a month, in a total, after the quotes have all gone out.

There's a version of this that's almost careful and still blind. They do record clicks, in a spreadsheet the supervisor keeps, a total per shift. The shift total can't be argued back into orders. A bad order and an easy order share the shift, and the total makes both of them look like the average. The average is what the price list was already using. You've measured your way back to the lump. Measurement that doesn't touch the ticket is a nicer lump, not a different system.

The mess is the ordinary day

The information that would have explained the lump is lying around the building in forms that don't meet. Files arrive by email and by a link that expires. The file that was actually imaged is on a desktop, or in a folder named in a way that made sense on the day. The job bag has an earlier file, or a proof, or nothing. Proofs go out from a device that isn't the press, and the approval comes back as a sentence in a thread: "fine, please run". Which file "fine" referred to is a matter of memory. Imposition gets rebuilt when the finishing lead says the cut is wrong, and the rebuild lives in that person's head, or in a spare preset, not as the standard for the next time this format appears. Colour targets live with one operator. When she's away, the shift matches to memory and to the screen. Tickets say "as before". Before isn't a specification. Before is a person who may be off.

Scheduling is a whiteboard and someone who walks the floor. A job that's easy to image jumps the queue, because an easy job makes the press look occupied and makes the counter move. A job that's stuck in a file argument doesn't appear as stuck. It isn't on the board as a preflight problem with an age. It's a pile of messages, and then it's a phone call from the buyer, late. The plant reads the phone call as evidence that it needs more hours. It adds overtime. Often the press was waiting. Waiting doesn't have a column, so it doesn't get mentioned when the overtime is justified. The justification is the counter and the complaints, side by side, and nobody lays the waiting in between them.

This is the mess. It isn't a scandal and it isn't a chaotic personality. It's a set of ordinary shortcuts that never got closed: the desktop file, the verbal approval, the colour in one person's eye, the whiteboard, the bin that isn't on the ticket, the click that isn't on the ticket. Each shortcut saves a minute. Together they make the plant unable to say what an order did. The first plant, the one with the same press, can say it because the order has a record. Time before the file was printable. Stock issued. Clicks. Whether it was imaged again, and why. What finishing spoiled. What left. That plant's arguments are about those lines. This plant's arguments are about effort and bad luck.

What the mess does to the next machine

The second plant will buy another press, or a newer one, before it will count a reprint. The reasoning is sincere. The current device is "flat out", the counter says so, and the late jobs say the capacity isn't enough. Another device is a purchase you can point at. A record is a change in how a job bag works, and it doesn't photograph. So the mess gets a bigger stage. Two devices, both fed by desktop files and verbal approvals, both with bins that don't talk to the estimate. The lump gets larger and no easier to open.

Meanwhile the quotes stay blunt. Awkward jobs are underpriced, because the list doesn't know they're awkward, so the plant wins them and regrets them. Easy jobs are priced like the awkward ones, so a buyer who only has easy jobs takes them to a plant that can tell the difference and charge for the difference. Or every job is priced with a pad of fear, and the plant loses the straightforward work that would have run cleanly and keeps the rescues. You can't see which of these is happening from the monthly lump. You can see it from orders. They don't have orders in that sense. They have a pile, a counter, and a feeling that digital is harder than the demonstration suggested.

The press was never the part that failed. It's fast, in the plain sense that it can image a ready file without a long start. What they own, besides that, is a system that doesn't know a ready file from a rescue, a sellable copy from a click, or a wasted sheet from a productive hour. Two plants, one device class. The separation isn't the hardware on the tour. It's whether last Tuesday's orders can still be explained on Wednesday. In the second plant they can't. Wednesday is already busy producing another pile the ticket won't remember.

Written by

Helen Ward

Sheetfed printer

Runs sheetfed offset and writes up the job that went wrong on her press.

Related articles

Process

Flexo still fits packaging

# Flexo still fits packaging The buyer had been to a presentation, and he arrived wanting the range moved. Film, board, a family of labels, all of it, onto a di...

Process

Print on demand moves the stock and the cost

# Print on demand moves the stock and the cost The publisher showed me an empty bay and he was proud of it. A year ago the bay had held a title that sold steadi...

Process

Versioned packaging is a production problem

# Versioned packaging is a production problem Marketing called it a celebration. A commemorative sleeve, one brand, a dozen pictures, a few lines of type that c...

Process

Print that can be scanned, used and recovered

# Print that can be scanned, used and recovered The code scanned on the flat sheet. I watched the buyer's phone pick it up from the press table, quick and pleas...