Someone sends you a drawing on a Tuesday. A bracket, a bit of welding, a couple of bends, powder coat. You have done a hundred like it.
You look at it for about a minute and a number arrives. Four hours, maybe four and a half. You write it down, put your rate on it, add material, and the quote is out before lunch.
That number came from memory. And memory is a generous estimator.
Memory rounds down
Ask yourself how long a job like that takes and your brain does something reasonable. It goes looking for the last few times you did it and picks the one that felt typical.
The one that felt typical is almost never the one where the fixture went missing, or the first part had to be remade, or the customer called halfway through with a small change. Those jobs feel like exceptions. So they get left out, and the number you remember is the one where everything went right.
Over a year, the exceptions are not exceptions. They are the job. A quote built from the clean version is short every single time, by just enough that nobody notices on any one job.
The numbers you need already exist
Here is the frustrating part. If your shop tracks time at all, the real answer is already sitting somewhere. You have quoted that bracket before. Your people clocked the hours. The job shipped.
The problem is that the hours live in one place and the quote lives in another. To use them you would have to dig out three old jobs, find the time on each, line it up against what you quoted, and do the maths. Nobody does that at eleven in the morning with an RFQ waiting.
So the memory number wins. Not because it is better, but because it is right there.
Quote from the last few jobs, not the best one
The fix is simple to say. When you price a line, look at what that same kind of work actually took on your last few jobs, and let that move the number.
A few rules make it work:
- Use your own jobs only. Another shop's hours tell you nothing about your people, your machines or your fixtures.
- Wait for a few jobs before you trust it. One bad job should not reshape your pricing. Three or more finished jobs is a fair start.
- Look at the step, not the whole job. "Welding ran 18% over" tells you what to change. "The job ran long" does not.
- Show it, do not apply it. Sometimes the last three jobs were slow for a reason that will not happen again. The estimator should decide.
Do that and something useful happens. Every job you finish makes the next quote a little more honest. The shop that does this for a year quotes very differently from the shop that does not, even if they started at the same rate.
What this looks like in Quotara
This is the reason we built Quinn, the AI estimator inside Quotara.
When you add a labour line, Quinn looks at the jobs your shop has already shipped that used the same part or operation. If there are at least three, it shows you what they were quoted at and what they actually took, and suggests an adjusted number. It never changes the quote by itself. You decide.
The hours come from the floor, where your team clocks in and out of each step. So nobody has to dig anything out. The moment a job ships, its real hours start counting toward the next quote.
And Quinn only ever learns from your shop. Your rates and your hours are never mixed with another shop's or used to price anyone else's work.
Try it on one line
You do not have to change how you quote to test this. Next time a familiar job comes in, write down your memory number first. Then look at what the last three of those actually took.
If the two numbers match, great. Your memory is better than most.
If they do not, you have just found money that was walking out the door on every one of those jobs.
Try Quotara free for 14 days. No credit card required.