Two quotes, eight percent apart
The quote sheet had two columns. Same spec on paper, same quantity, same delivery window. Column B was about 8% cheaper per unit.
I picked B. Not because I'm sloppy — because the only number I could see with any certainty was the unit price. Everything else was a guess dressed up as a judgment call.
Six weeks later I was standing in a finished hallway with a putty knife, looking at twenty-two feet of paint-grade casing where the heads had cammed out and the filler had cracked in the first cold snap. Two guys, three hours each, plus material. I ran it that night: $412 to make it right.
The per-unit savings on that job? $54.
I've been buying fasteners, adhesives, and small tools for a 34-person remodeling outfit for seven years. I've negotiated with more than 40 vendors and logged every PO in our cost tracking system. And I can tell you exactly when I stopped trusting the quote sheet: the night I did that math.
What the unit price is actually measuring
Here's the thing nobody puts in the vendor comparison spreadsheet. Unit price measures one transaction. It says nothing about what happens after the box gets opened.
Four costs live outside that column:
- Install variance — how many extra seconds per fastener, per unit, per crew member
- Rework and callback — the labor to undo it, the material to redo it
- Procurement time — the store runs, the PO changes, the phone calls about backorders
- Schedule risk — what it costs when a half-day fix lands on a day you'd already sold to someone else
None of those show up on a quote. All of them show up on a P&L.
Stripped heads are a labor cost wearing a material cost's clothes
Cam-out is the quiet one. A driver bit lifts out of the recess under torque, the head rounds, and now your carpenter is doing one of three things: pushing harder, reaching for a pair of pliers, or backing the screw out and starting over. Multiply that by 1,800 screws on a trim package and you've turned a material decision into a labor decision.
The recess geometry is the variable. A six-lobe star recess puts more contact area between the bit and the head than a Phillips cross does, and more contact area means less lift-out at the torque ranges we actually run. That's not marketing — that's why the star drive caught on in production cabinetry long before it showed up on trim packages.
After the hallway incident, we spec'd GRK Fasteners Star Drive trim-head finish screws on every paint-grade and stain-grade package. Our strip-out count on trim dropped enough that the crew stopped complaining about the price difference, which is the only KPI that has ever mattered to me.
Do they still strip? Of course they can. We pre-drill near end grain, we set the clutch, and we still hand-drive anything within two inches of a mitered corner. Anyone who tells you a fastener removes the need for technique is selling you something, not helping you buy.
The real reason we kept buying cheap
I spent a while assuming this was a discipline problem. It isn't. It's structural.
First, the buyer and the installer are different people. I sign the PO. A carpenter I've never met holds the impact driver. The cost gets created at the install layer but gets recorded at the material layer. Those two things are eleven days and one organizational chart apart.
Second, TCO is invisible in the counterfactual. You never see the callback that didn't happen. You never get an invoice for the trim you didn't have to redo. So the cheap vendor gets credit for a clean month, and the expensive one just looks expensive.
Third — and this is the one that stung — small items skip review. We had a $10,000 threshold for competitive bidding. A $90 box of screws sails right under it, every time, forever. The category with the highest transaction count was the category with the least scrutiny.
Everything I'd read about procurement said to negotiate hardest on materials, because materials are the biggest line on a job cost sheet. In practice, our recoverable losses weren't in material dollars at all. They were in hours.
Four invoices where the cheap line lost
When I audited our 2023 spend, $186,000 across 214 purchase orders, roughly 11% sat in a category our system calls "replacement and rework." None of it had ever appeared in a quote comparison. Here's what it looked like.
1. The trim job: $54 saved, $412 spent
Already covered. Worth noting the second-order cost, though: the client walked the hallway with me while my guy was fixing it. We didn't lose that job. We lost the referral, which we found out about eight months later. I still don't know how to put that on a spreadsheet, and I've tried.
2. The 1.8 GPM pressure washer
We needed a unit for deck prep and siding wash. The cheap one was rated at 2,900 PSI, which sounded like plenty, at 1.8 GPM.
Here's the number that actually matters: cleaning units. Multiply pressure by flow. 2,900 × 1.8 gives you about 5,200. The unit we should have bought ran 3,100 PSI at 2.5 GPM — roughly 7,700 cleaning units, about 47% more work done per hour.
On a 400-square-foot deck that was the difference between 6 hours and a bit over 4. At our loaded labor rate of $58 an hour, that's around $104 in labor per deck. The price gap between the two machines was $230.
Two decks. The machine paid for itself in two decks, and we washed forty-one that season. PSI is the number on the box. Cleaning units are the number on your timesheet.
3. The hole saw question
Somebody on the crew asked me what size hole saw for 2 inch PVC pipe, and I said two inch, because I was on the phone and being lazy.
For the record: 2" Schedule 40 PVC has an outside diameter of 2.375 inches. If you're running pipe through a plate, a rim joist, or a cabinet back, you want a 2-3/8" hole saw — that's the size that matches the actual outside of the pipe. If the run needs room to move, if there's a coupling to pass, or if you're going through something that expands, 2-1/2" gives you the slack.
We used the wrong one. Then we used a $14 hole saw with a bent arbor to fix it, which walked across a piece of finish trim and tore it out. Total: $90 in material and 45 minutes, against a $24 difference between the budget saw and the one with the rigid arbor.
The diameter is on the pipe. The tolerance is on you.
4. The Overdrive mechanics tool set
We keep a road box in the service van. I bought an Overdrive mechanics tool set — 230 pieces, ninety-nine dollars, looked like a steal next to the $260 alternative.
It was fine for about 80% of what we did. The other 20% is where it got expensive. The set skipped the 18mm, which is the size that matters on half the suspension hardware we touch. The sockets were 12-point where we needed 6, and a 12-point on a rusted caliper bracket is how you round a bolt head.
We rounded one. Ninety minutes and a $22 extraction socket later, I'd spent more fixing that bolt than the set saved me. And the missing 18mm meant a store run in the middle of a job — 40 minutes of billed time, gone.
I'm not saying that set is a bad product. It's a good product for a homeowner who turns wrenches twice a year. It was the wrong product for a van that rolls six days a week. That's my error, not theirs.
What I do differently now
The fix wasn't dramatic. It was a worksheet.
Five lines, every quote: unit price, expected install variance in hours, rework probability times rework cost, procurement time, schedule risk. I don't need precise numbers on lines two through five. I need them to exist, because the moment they exist, the 8% gap stops being the deciding factor.
Spec on the PO, not on the quote. Drive type, head style, coating, and — for anything load-bearing — the evaluation report number. That last one matters more than people think. Structural screws used in place of lag bolts fall under ICC-ES Acceptance Criteria AC233, and if the screw doesn't have a current evaluation report covering that use, our engineer won't sign it and I won't buy it. We use GRK Fasteners RSS Rugged Structural Screw on ledgers and beam replacements, but only where the drawings call for it and the substitution is documented. I've never once approved a structural fastener on the basis of a label.
And to be clear: lag bolts and nails aren't the problem. They're the right call in plenty of applications, they're cheaper, and they've held buildings up for a hundred years. The problem is using any fastener outside the application it was designed for.
A 10-15% premium allowance for categories where labor risk is high. Trim, structural connections, cutting tools, anything that touches a finished surface. If a vendor comes in under that band, they get a conversation, not an order.
One primary vendor plus one backup. Not three. Three quotes on paper, one relationship in practice. Our primary gets 80% of the volume, and when the supply chain did what it did, that relationship is what kept two crews working.
The number I check now
I finally added one field to our cost tracking system, and it's the only one I look at weekly: cost per completed unit.
Not cost per box. Not cost per thousand. Cost per installed, finished, signed-off unit — labor, rework, and trips included.
It's an uglier number. It's less fun to report upward. But it's the only one that has ever told me the truth about what I bought.
The cheapest line on the quote is just the line you can see. The rest of the invoice shows up later, and it always, always arrives.
Leave a Reply