Print shop quotation automation: how does the veteran estimator's know-how get into the system?

Look at the paper, the quantity, the finishing, how urgent it is, and the veteran estimator names a price in one sentence. The young salesperson spends half a day on the same job and still carries it over to ask him. When the master takes a day off, the estimating desk stops. This post breaks down the four completely different kinds of things mixed into one estimate: which two should go to the system now, which one goes in but needs an override, which one should not be a formula but a checkpoint, and why the rush-job line is the easiest one to get wrong. Not which software to buy. Just how to take it apart.

Johnny Yang12 min read
Print shop quotation automation: how does the veteran estimator's know-how get into the system?

“I look at a job like this once and I know what to quote. Ask me to write it as a formula and I can't.”

That day, beside the estimating desk at a print shop, the old master finished saying this, pulled up the file the customer had sent, scrolled through it twice, took down a paper sample book worn fuzzy at the edges, tapped a few keys on a calculator, and the quotation was done. The young salesperson next to him had spent the whole morning on the same job. When he finished, he didn't dare send it, and carried it over for the master to look at anyway. The owner said jobs like this come in every day, and every one of them has to go past the master. When the master takes a day off, the estimating desk stops.

I'm Johnny Yang, Brightstream Technology, Taiwan. I've led implementation work with my team for seven years, 252+ projects across 32+ industries. The print-shop situation where “quotations are stuck on one person” is common in the projects I've seen, and the owner has usually put up with it for years. Here is the conclusion in one go: quotation automation for a print shop is not “press one button and a quotation comes out.” It is splitting the estimate into two halves — what can be calculated goes to the system, what a person has to look at stays with a person, and a visible checkpoint connects the two. What can be calculated: paper usage, sheet layout and imposition, color count and press labor, standard finishing steps. Each has one correct answer, and these should go into the system now. Lookup items, like paper cost by stock and weight, price tiers for outsourced processes, and terms already agreed with regular customers, should also go into the system, but with one named person responsible for maintaining them. An outdated table is more dangerous than no table. File quality, whether this artwork will actually print, whether this customer will keep revising until you want to drop the job, whose press time gets bumped when a rush job comes in — don't force these into formulas. Have the system stop and flag them, and have a person tick to confirm and leave a reason. The first half isn't about speed. It's so two salespeople quoting the same job get the same answer. The second half isn't about replacing the master. It's about recording his reasoning every time, so the next person has something to learn from. If nearly every job you take is fully custom-built, or nobody is willing to maintain even the paper price table, don't touch a system yet.

Where the owner's instinct is right, and where it's wrong

First, the owner's side. The owner feels “quotations can't be fully automated.” That instinct is right, and more right than most software salespeople will admit. Print cost is not pure arithmetic. Same quantity, same paper, and two jobs can cost very different amounts, because the artwork is different, the file quality is different, the press schedule that day is different, and the finishing shop is busy or not busy this week. The master is valuable precisely because he folds all of that into those few minutes.

But the instinct goes wrong in two places, and both hurt.

The first mistake is reading “can't be fully automated” as “so the whole thing has to run on experience.” So even things with one answer that you can look up in a reference book — how many pieces per sheet, how many overs to allow, how to count reams — stay in the master's head too. That ties the parts that need judgment to the pure arithmetic. The result: the master is chained to the estimating desk, and more than half of what he does there is a calculator's job.

The second mistake is setting the goal as “press one button and a quotation comes out.” The owner goes to look at systems with that expectation, decides none of them can do it, gives up on the whole thing, and keeps letting the master carry it. In reality, just taking over the half that can be calculated drops the pressure on the estimating desk a lot, and it brings a more valuable side effect: for the same job, salesperson A and salesperson B finally get the same number.

One estimate mixes four completely different things

On site I do something very low-tech. I pull out recent estimates and go line by line: “Was this line calculated, or did the master just call it?” When you're done, you find that a neat-looking estimate mixes four completely different kinds of things.

The first is pure arithmetic. Sheet layout, reams of paper, press labor by color count, labor hours for standard processes. Given the inputs, there's one answer. If two people get different answers, someone made a mistake. Keeping this in someone's head has no benefit.

The second is lookup. Cost by paper stock and weight, price tiers for outsourced processes, terms already agreed with regular customers. The answer isn't unique, but the options are limited and can be listed. It doesn't need judgment. It needs someone to update the table regularly. The barrier to putting this in a system isn't technical. It's whether anyone is willing to be the person who maintains it.

The third is rule-based judgment. How much spoilage to allow, what kind of artwork must be proofed first, how bad a cross-page alignment problem has to be before you send it back to the customer. These can be written as conditions, but someone has to sign off on the conditions themselves. They can go into the system, as long as there's an override and every change records a reason.

The fourth is pure judgment. Will the halftone dots in this artwork blur on this paper. Is the customer's file once again exported from a word processor. Will this customer, after signing, keep revising until you want to drop the job. Don't touch this one. Forcing it into a formula just produces a quotation nobody trusts.

TypeWhat it looks like on the estimateHow to handle it
Pure arithmeticSheet layout, paper, press labor by color, standard processesStraight into the system, no manual room
LookupCost by paper stock and weight, outsourced process tiersInto the system, but assign someone to maintain it
Rule-based judgmentHow much spoilage, whether to proofInto the system, but must allow override and record the reason
Pure judgmentFile quality, customer habits, rush-job conflictsNot a formula; make it a mandatory confirmation checkpoint

The half you can calculate: the goal is to kill version differences

Many owners think handing pure arithmetic to a system is about saving time. Saving time is a side benefit. The real gain is consistency.

The most common leak on site isn't quoting too low. It's the same kind of job, quoted by different people, coming out at different numbers, and nobody knows which one is right. Salesperson A uses the Excel file from a few years ago. Salesperson B uses his own saved copy. The master uses the one in his head. The customer takes A's price to B, B doesn't dare say it's wrong, and absorbs the difference. All of this happens under the table. At year-end the margin looks off, and you can no longer trace which jobs did it. I break down this kind of invisible bleeding in more detail in “Blind spots in quotation margins”.

So the work for this half is clear. How the sheet is cut, how imposition is calculated, how overs and spoilage are carried, how much press labor each color count means, how each finishing step stacks — all of it goes into one set of logic, and whoever runs it gets the same result. After that, the argument at the estimating desk changes from “your number is different from mine” to “how much spoilage should this job carry.” Only the second kind of argument is worth having.

One more thing that often gets missed: keep the calculation, not just the total. Which paper price list this job used, which set of processes, how much spoilage was allowed. When a customer comes back with “that's not what you quoted last time,” you have something to check. Without this layer, a quotation system is just a better-looking calculator.

The half a person has to look at: make judgment a checkpoint, not a rule

This half is where the whole thing most easily goes wrong, and it almost always goes wrong the same way: trying to write the master's judgment as rules.

One type of failed implementation I've seen is very typical. Bring the master in for a few days, make him spell out every judgment condition, write them into a long chain of conditions, and stuff that into the system. At launch everyone follows it. Then someone finds a condition gives a skewed result on a certain paper and changes it by hand. The person who changed it leaves no record. The next person doesn't know why it was changed, and changes it too. A while later, nobody trusts the numbers the system puts out, and every job goes back to the estimating desk.

What works better is giving the system a different role: the system doesn't make the judgment. The system makes sure nobody skips looking.

Concretely, put a mandatory checkpoint in the quotation flow. The system lays out the warning signs on this job — new customer, converted file, spot colors, this paper stock has been hard to get lately, the delivery date falls on days that are already fully booked — and then stops and waits for a person to tick. The person can say “I've looked, it's fine,” or “this one needs a proof first.” The system doesn't guess. It only stops things.

The real value of this checkpoint is the second layer: at the moment of ticking, write one sentence on why. Over time those sentences become a judgment checklist grown from your own shop, more accurate than any training material. It's also key material for succession. What a second-generation owner or a new estimator lacks most is never data. It's “why the previous generation decided that way.” I covered this in “Digitizing for succession: what to do first”. If only the amount goes into the system, the reasoning still walks out the door with the person.

The rush-job line is the easiest to get wrong

Rush jobs deserve their own section. They're the most misunderstood line in a print shop's quotation.

The owner's first instinct is to set a markup formula for rush jobs, put it in the system, and let salespeople apply it themselves. This usually runs smoothly at launch, then stops working. Salespeople, to close the deal, find reasons to call a job less urgent than it is. Customers learn too: the markup is fixed anyway, so every job is a rush. Once rush jobs become normal, the markup is a number with no deterrent.

The deeper problem is that the cost of a rush job was never “fast.” It's displacement. When a rush job goes in, the real cost is which already-scheduled jobs it pushes out, whether those customers get upset, whether someone has to work overtime, whether it has to be sent out to another shop. Those answers are different every day and tied to the schedule at that moment. No single formula calculates them.

So rush jobs shouldn't go into a formula. They should go into approval. The system does one thing: calculates the conflict and shows it to a person — which jobs this one pushes out, which day overflows, whether that press still has the slack. The person looks, then decides whether to take it and whether to charge extra. The benefit is that responsibility lands where it belongs. Who made the call, and what the schedule looked like at the time, stays in the system, not in some LINE chat nobody remembers.

When you should not do this yet

This section may matter more than everything above. A quotation system is an easy thing to buy wrong.

The first case is jobs that are too custom. Some shops take small runs where every job has different specs and nearly every one needs a person to look at it. In that structure, the share you can automate is low to begin with, and the build and maintenance time won't pay back. What that shop should do is record the master's reasoning, not put in a system first.

The second case is nobody maintaining the base data. Paper prices move, outsourced process terms change, customer agreements get renewed. If no clear person is responsible for updating these tables, soon after launch the system starts putting out outdated numbers, and because it looks official, people are more likely to accept them without question. If you can't even name who will maintain them, you're not ready.

The third case is that your real problem isn't the quotation. Some shops quote reasonably well. What's actually leaking is that nobody follows up after a quotation goes out, nobody reconciles the order once it's won, and nobody chases the payment once it's reconciled. Those bleed every month and are far more serious than a slightly slow quotation. Get the order wrong and neither gets done well.

The fourth is more awkward: the person holding all the judgment is the owner, and he has no plan to let go. In that case, however good the system is, salespeople will go around it and ask the owner, because that's faster.

The first thing you can do tomorrow

Don't go find a vendor first. Don't hold a meeting about whether to do an implementation. Do one small thing.

Pull up your recent estimates, grab a stack at random, and with two pens of different colors, mark next to each line: was this calculated, or did someone call it. When you're done, lay the “someone called it” lines out on the table and see which kinds repeat most. They're usually these: how much spoilage to allow, whether this file can go straight to press, whether to take a deposit from this customer, whether to accept the rush job. Those lines are the real judgment points in your shop. Everything else is a calculator's job.

After that, do only two things. For the calculated lines, have one person consolidate the logic into one set. Your existing tools are fine for now. The point is that the whole shop keeps only one version. For the lines someone called, starting this week, every time someone makes the call, write one sentence next to it on why. In a month you'll have two things: one consistent set of calculation logic, and a stack of real judgment records. Then talk about which system to use, and your requirements will be clear enough that vendors will have a hard time bluffing you. If you're not sure which side to start on, we offer a free 30-minute diagnostic.

Found this useful? Pass it on
30 minutes · free

Bring the one thing you are stuck on — a free 30-minute diagnostic

You do not need to know what system you want. Tell Johnny Yang about the one thing that keeps blocking you, and in 30 minutes you will know whether to skip it, build it with coaching, or pay for a custom build. No pitch — if we cannot help, you will hear that straight.

Johnny Yang replies himself · limited spots each month · or just book a free 30-minute diagnostic

Johnny Yang (熊董)
Johnny Yang (熊董)
Founder, Brightstream Technology

Serial founder leading a small team that builds systems into an owner's everyday work.

As a founder
7 yrs
Projects delivered
252+
Industries served
32+
About Johnny