Joining the press to the sales desk is the hard part

# Joining the press to the sales desk is the hard part The press knew the job had started. The counter was moving, the minder had signed the colour, and the delivery belt was building a pallet that matched the ticket on the console. Sales, two floors up, had a spreadsheet open on a job

Joining the press to the sales desk is the hard part

The press knew the job had started. The counter was moving, the minder had signed the colour, and the delivery belt was building a pallet that matched the ticket on the console. Sales, two floors up, had a spreadsheet open on a job of the same name and a different quantity, and they told the buyer the job would start tomorrow. I heard both versions within ten minutes. Nobody was lying. They were each looking at the record they trusted, and the two records had stopped being the same job sometime after the estimate and before the feeder. Joining the factory to the selling desk is the hard part of running a plant. The machines were never the mystery.

Two clocks, two quantities, one angry phone

The buyer's call came to sales, which is proper, and sales came to me, which is also proper, and then we stood at the console and discovered the pallet. Sales had promised a Monday start because Monday was the date in the spreadsheet, typed in when the order was taken and never revised when planning pulled the job forward to fill a gap. The factory had done the right thing with the time it had. The buyer had been told the wrong thing about a job that was, by then, halfway to the wrapper. He wasn't furious about the print. He was furious that his own customer had been given a date we had already beaten, and his warehouse wasn't ready. An early pallet can be as rude as a late one when the other person has booked a dock.

Quantity was the second wound. The spreadsheet said the ordered amount. The console said impressions, including a restart after a feeder fault. The delivery note, which the warehouse was about to write from the pallet labels, said a third figure, the good sheets they were actually going to wrap. Three honest numbers, and not one of them was the number the others were using. The invoice, if we had let the week run on, would have been a fourth, because accounts works from the delivery note if they remember and from the order if they don't. This is how a simple job grows a dispute. Not from bad faith. From four doors into the same room, and no agreement about which door is the job.

I stopped the argument about who was right and I made them write the three figures on the ticket, with a word beside each. Ordered. Run. Good. The buyer could be told the truth in one call: we started early, the good quantity will be the ordered quantity, the extra impressions were ours and will not appear on the invoice. That call took five minutes. Finding a place to write the figures so the next shift would see them took the rest of the afternoon, because the spreadsheet doesn't live on the press and the console doesn't live in sales. We were, for one job, the human link. A plant cannot be a human link for every job. That is the whole problem dressed up as a busy day.

Why one system is harder than buying one

Everyone agrees, in the meeting, that there should be a single job record. Agreement at that altitude is cheap. The difficulty starts when you ask whose numbers get overwritten. Sales lives on price, promise, and the story told to the buyer. The factory lives on time, spoilage, and the story told by the machine. Those stories are both legitimate, and they threaten each other. If the press record shows a heavy makeready, the estimate looks optimistic and the estimator feels accused. If the sales record shows a price that barely covers the board, the minder starts to wonder why he is sweating a job the building doesn't rate. People hide figures when the figures can be used against them. A shared record without a shared truce is a surveillance system, and the staff will keep a second book. They already do. It is the spreadsheet.

There is also the dull mechanics, and I won't pretend it is only politics. The order is taken in one place, planned in another, plated or ripped in a third, run on a machine that counts in its own dialect, finished by people who tick a paper traveller, and invoiced by someone who needs a quantity and a price and very little poetry. Each of those steps grew a tool, or a tab, or a book, because the step had a problem the others didn't share. Replacing the lot with one named product, on a Monday, is how you stop the factory for a fortnight. I have no interest in that kind of heroism, and I won't recommend a supplier here. The hard part is not the logo on the software. The hard part is deciding which event in the life of a job is allowed to change the date the buyer was given, and who is allowed to type it.

Identity is where it usually breaks. Sales calls it the client's purchase order. Planning calls it the works order. The press calls it whatever was on the plate ticket. A reprint adds a letter. A version adds a child. A split delivery adds a row. Within a month a "single record" has triplets, and the buyer is asking why his order appears twice. If you don't decide, before anyone shares a screen, what a job is, the system will decide for you in the most irritating way. We spent longer on that argument than on any function list. A job is the thing a buyer ordered, at a quantity, for a delivery. A version is a child with its own quantity. A reprint against spoilage is not a new job and it is not billable. Write that down in language a minder and a handler will both use, or don't bother connecting anything.

What the plant gains when the record is actually shared

When it works, even partly, the gain is not a dashboard. I don't need a dashboard. I need the handler to see that the job has started, and the minder to see the quantity he is allowed to wrap, and accounts to invoice the good sheets rather than the impressions. Those three sights remove a class of phone call that currently passes for communication. The buyer who rings "just to check" is often ringing because the last time he checked, we didn't know. A record both sides trust is a quieter office. Quiet is worth money in a plant, though I won't invent how much. It is the hours sales doesn't spend apologising for a status the factory had already achieved.

The deeper gain is that estimating starts to meet reality. If makeready time, restarts, and actual good quantity land on the same record as the quote, the next quote can be written by someone who has seen the last one land. That is uncomfortable the first month, because the gaps become visible. A job that was sold as a quick plate and ran as a long colour-match will sit there in public. Good. The alternative is that the minder knows and the estimator never does, so the plant keeps selling the imaginary version of itself. Sharing the record is how the imaginary version retires. It requires that a bad estimate is treated as information, not as a hanging. If it is a hanging, the numbers will go back into the private spreadsheet before the quarter is out.

Planning gains a sequence it can defend. A promised date that only sales can edit, and that the factory can only miss, is a stick. A date that moves when a job is pulled forward, or when a proof is still unsigned, and that moves in a place sales will see before they speak to the buyer, is a tool. The early pallet on that Monday would have been a phone call on Friday if the pulled-forward start had shown up on the handler's screen. The buyer could have moved his own dock. We didn't lack kindness. We lacked a single place where "we started" was a fact rather than a rumour.

The fight about whose number is real

We tried this badly the first time, and the failure is more useful than the brochure I didn't read. We asked the press to be master. Every impression flowed to the office, and the office drowned. The buyer does not want impressions. He wants good copies, and the machine doesn't know good. It knows movement. Minders started ignoring the screen because correcting it took longer than running, and the corrections they did type were doubted by sales, who could see their order being edited by someone they don't manage. Trust collapsed in a fortnight. The second book came back.

The second attempt made sales the master, and the factory was told to follow the spreadsheet. That lasted until a job was late and the spreadsheet still said tomorrow. The minder, who had the late job in his hands, couldn't change the date the buyer was hearing. He stopped believing the screen, which was fair. A record that the people at the point of work are forbidden to update is a poster. Posters don't run presses.

What held, the third time, was dull enough to survive. One job number, born when the order is accepted, used on the ticket, the pallet, and the invoice. Sales owns the promise to the buyer and can see the factory status. The factory owns the status and can see the promise. Either side can change the field it owns. Neither can quietly change the other's. A move in the promise sends a prompt to the planner before anyone rings the buyer. A move in the status, from planned to running to wrapped, is visible upstairs without a phone call. Spoilage is recorded as spoilage, against the job, by the shift that made it, and it is not an excuse for a second invoice. We didn't buy a revolution. We stopped keeping two clocks.

It still breaks. A night shift will run a job early and forget to change the status, and the morning handler will tell a lie in good faith. A salesperson will agree a new date in a corridor and write it in a notebook. The shared record doesn't cure those habits. It makes them visible the next morning, which is the condition for curing them. I still walk down when the two floors disagree. I walk less often, and the ticket usually has the same quantity as the screen when I arrive. That is the gain I am willing to claim. Not a transformation. Fewer days when the press and the desk describe two different lives for one pallet.

Leave the second book nowhere to hide

If you are trying this, start with the job the two sides already fight about. Quantity, date, and whether it has started. Get those three onto one record that both rooms look at before they speak. Leave price visible to the people who price, and don't wave it at the console as a moral lesson. Leave the minder's spoilage visible to estimating, and don't wave it at the buyer as a confession. The point of one record is that the building tells itself the truth in time to change a promise. It is not that every number becomes a weapon.

The spreadsheet on the sales desk is still there. I didn't ban it in a speech. I made it useless for the three facts that cause the angry phone calls, by making the shared record faster to check than the tab. People abandon a second book when it is the slow one, not when they are told they are disloyal. The console still counts impressions. The delivery note still counts good sheets. The difference is that both numbers now have a name, and the buyer is only ever told the one that will arrive on his dock. The press knew the job had started. Sales knows it too, now, before he does. That is the join. It was never going to be the easy part, and it is the part the plant actually feels.

Written by

Priya Shah

Prepress

Makes plates and proofs, and sends the file back when the press cannot fix it.

Related articles

Industry

Buyers want shorter dates and clearer proof

# Buyers want shorter dates and clearer proof Her email had three questions and a date that had moved in by a week. Could we deliver sooner. Could she see where...

Industry

Recycled consumables and taking them back

# Recycled consumables and taking them back The form asked if we take material back. Plates, blankets, ink tins, solvent, cores. Sales had already ticked yes, b...

Industry

Carbon paperwork in cross-border print

# Carbon paperwork in cross-border print The cartons were printed, creased and packed. The colour had been signed. The truck was in the yard in Italy, waiting f...