The first one always gets priced off the last one that rhymed

A man walks into the shop with a problem and no print. He needs a 10-inch auger at a length nobody sells, because the new bin sits four feet higher than the old one. Or he wants a squeeze chute that works left-handed because that is the side his alley comes in on. Or he has a grain cart with a cracked auger housing and wants to know what it will run to fix it before he decides whether to fix it.

Work out your own shop rate → Free, no account, five minutes.

In every one of those conversations you have to produce a number, usually before he leaves, and you have never built the thing. There is no routing, no history, no lot to spread anything over. So the number comes from the last job that rhymed with this one — and that number came from the one before it, which came from a figure somebody wrote on a shop traveler in 2014 and nobody has questioned since.

That is the first-unit problem, and it is not a knowledge problem. Any shop that has been building short-line equipment for three generations knows what steel costs, what a hydraulic cylinder costs and how long it takes their own people to weld out a frame. What is missing is the arithmetic that turns those known things into a number for an unknown machine. That arithmetic is not hard. Almost nobody writes it down.

You cannot estimate a machine you have never built. You can estimate steel, bought parts, machine time and hours on the floor, because you do all four every week. So stop looking at the machine and start looking at the four piles it comes apart into.

There is a fifth bucket most shops leave out, which is the hours spent figuring out what the customer actually wants: the two phone calls, the sketch, the trip out to look at the bin. On a repeat production run that time is nothing. On a one-off it is real, and if you never put it in a quote you are giving it away on every custom job you take.

Here is the whole thing on one machine. A cattleman wants a hydraulic squeeze chute — left-hand headgate, hydraulic squeeze, hydraulic tailgate, sheeted sides, on a self-contained 12-volt power unit. You have built manual chutes. You have never built this one. Shop rates for the example: $95 an hour on the machining side, $85 in the fab bay, $70 for assembly and paint.

Steel first. You take off the frame from your own sketch: 140 feet of 2x2x3/16 square tube at 4.32 lb/ft is 605 lb; 60 feet of 2x3x3/16 rectangular tube at 5.59 lb/ft is 335 lb; 48 square feet of 3/16 plate at 7.65 lb/sq ft is 367 lb; 6 square feet of 1/2 plate at 20.4 lb/sq ft is 122 lb; 30 feet of 2-inch schedule 40 pipe at 3.65 lb/ft is 110 lb. That is 1,539 lb on the truck. At $0.88 a pound for the tube, $0.74 for the plate and $1.05 for the pipe, the steel is $1,304.56. About 140 lb of that leaves as drop, which comes back at $0.11 in the scrap bin — call it a $15.40 credit.

Bought parts next, and on this machine they are the heavy end: two 2-inch bore by 8-inch stroke cylinders at $185, one 2.5 by 16 at $265, a 1.5 hp 12-volt power unit with a 3-gallon reservoir at $1,150, a four-spool valve at $310, hose and fittings at $265, five gallons of oil at $85, pins and bushings and zerks at $140, headgate latch hardware at $95, a rubber floor mat at $120, paint and primer at $210, decals and a serial plate at $40. That is $3,050. At a 15% handling markup the line is $3,507.50.

Machining is small but not free. Fourteen turned pins and bushings on the lathe: 45 minutes of setup plus 12 minutes each is 213 minutes. Two latch plates and a cam on the mill: 30 minutes of setup plus 51 minutes of run is 81 minutes. Together that is 294 minutes, or 4.90 hours, at $95 — $465.50.

Then the floor. Layout and jigging for a machine that has never existed: 6.0 hours. Saw and plasma cutting including nesting the plate: 6.1 hours. Fit and tack the frame: 7.5. Weld out: 11.0. Grind and dress for paint: 4.5. That is 35.1 hours in the fab bay at $85, or $2,983.50. Blast, prime and paint 5.0 hours, plumb the hydraulics and hang the cylinders 6.5, final assembly 4.0, function test and leak check 2.5 — 18.0 hours at $70, or $1,260.00. The two phone calls, the sketch and the drive out to the corral come to 4.5 hours at $85, or $382.50.

Two things in that table are worth stopping on. The first is the first-article allowance. On a machine you have never built, your hour estimate is wrong — not because you are careless but because you do not yet know where the fits will fight you. Fifteen percent on labor only, never on steel or bought parts, is a defensible way to carry that. It is not padding. It is the price of the fact that the number you just built has no history behind it.

The second is the right-hand column. The second chute costs $1,858 less to build and prices $2,416 lower — about 17% off — and not one dollar of that is steel or bought parts. It is 15.5 hours: the jig exists, the cut list exists, the fits are known, the customer conversation already happened. If you quote the second one at the first one's price, you are 17% high against the shop down the road who is on their fourth. If you quote the first one at the second one's price, you eat the difference and call it a bad job.

Corn Belt work does not arrive evenly. It piles up ahead of planting and again ahead of harvest, and it goes quiet in the deep of winter and the middle of summer. Everybody in a shop in Iowa or eastern Nebraska knows this. What most shops do not do is put it in the price.

Here is why it belongs there. Your fixed costs — building, insurance, utilities, equipment notes, the office — do not care what month it is. Say they run $18,500 a month. Those dollars have to be carried by the hours your floor actually bills. In a slow January your six-man floor might bill 520 hours. In the two weeks before planting, with everybody on overtime and a chute waiting behind three planter repairs, it might bill 940.

Same chute, same steel, same 62.5 hours of work. The overhead riding on it swings $993.75 between the slowest month and the busiest — 7% of the sell price, on a machine where you were arguing with yourself about a $200 handling markup.

The wrong lesson is to quote January work at January rates. Nobody buys that, and it punishes the customer for your calendar. The right one is to build your rate off the year — $222,000 of fixed cost across roughly 8,900 billed hours is $24.94 an hour — and then know which side of that line you are standing on when you quote. In April you are absorbing overhead faster than your rate assumes, which is the month to hold the price and quote honest lead times instead of discounting to get the work. In January you are absorbing it slower, which is the month to build your own line for stock, do the jigs and fixtures you keep putting off, and take the odd job at a thin number knowing that a thin number still beats an idle bay.

Ask any shop that does both which side hurts, and it is repair. Not because the rate is wrong. Because on a fabrication job you know what is in front of you and on a repair you do not, and you cannot, until it is open.

A grain cart comes in with a cracked auger housing. You look at it in the yard, you have seen a dozen, and you say $1,400. In the good case you are right: 12 hours at $85 is $1,020, plus $180 in flighting and bearings, is $1,200 of cost and $200 of margin. But four times in ten it is not the good case. The gearbox is chewed up, the tube is cracked the full length under the paint, the flighting is worn thin the whole way. Now it is 22 hours and $520 in parts — $2,390 of cost against a $1,400 quote you already gave. That is a $990 loss.

Run those ten jobs: six at $200 up is $1,200; four at $990 down is $3,960. Net across ten repairs, minus $2,760 — an average of $276 lost every time one comes through the door. And because six of them looked fine, nobody in the shop believes repair is the problem. They think they had four bad jobs.

The fix is structural, not arithmetic. Quote the teardown as its own job — four hours at $85, $340, firm — and tell the customer plainly that at the end of it he gets a real number and a machine he can haul home either way. Then quote the repair from what is actually in front of you, with a not-to-exceed if he wants one. Customers who have been burned by an open-ended shop bill understand this immediately; it is the shops that are shy about proposing it.

Short-line builders love the phrase "available with countless options," and it is usually literally true. Take a pull-type sprayer with three boom widths, three tank sizes, three rate-control packages, two tire packages, two hitches and fenders optional. That is 3 x 3 x 3 x 2 x 2 x 2 — 216 configurations. You have probably built eleven of them. The other 205 are first units, and every phone call is a fresh quote.

Nobody can cost 216 machines. What you can cost is one base machine and each option as a delta — the added steel, the added bought parts, the added hours — and then add. Cost the base once, cost each option once, and any configuration is arithmetic instead of a guess.

A 60-foot boom, 1,000-gallon tank with GPS rate is then $18,400 plus $2,182.50 plus $1,150 plus $4,850 — $26,582.50 of cost, $34,025.60 at a 28% markup. That took thirty seconds and it is defensible line by line if the customer asks what the boom costs him.

One warning: some options change other options. The 90-foot boom needs a bigger pump and a heavier axle, so it is not a clean delta — it drags two other lines with it. Flag those pairs on your option list the first time you find one, because that is exactly the combination somebody orders eight months later when you have forgotten. This is the one part of option pricing that spreadsheets handle badly and that a real costing system — Spindlerate among them — is actually built for.

Everything above depends on one habit, and it is the habit small shops have the least of. When the machine goes out the door, somebody has to write down what it really took. Not the money. The hours and the pounds.

The reason almost nobody does this is not laziness. It is that the person who knows the answer is standing in the fab bay with a hood on and the paperwork lives in the office, and by the time the paperwork catches up, the machine has been gone a week and the number is a shrug. If you want the data, collect it where the work happens — a column on a whiteboard, a clipboard hanging on the jig, a photo texted to the office at the end of the shift. Any of those beats an intention.

The payoff is the right-hand column of that chute table. A shop with records builds the second one for 47 hours and knows it. A shop without them builds it in 47 hours too, and still quotes 62.5, and either loses the job or takes it at a price that made the customer's decision easy. The steel is the same either way. The difference is entirely in what somebody bothered to write down.