Building chemistry article
Why I Now Pay 15-25% More for Guaranteed Delivery on Division 10 Products
The Short Answer
On Division 10 construction specialties—expansion joints, wall protection, louvers, corner guards—I now budget 15% to 25% above the cheapest quote to buy a contractually guaranteed delivery date. Not because I've stopped caring about cost. Because in six years of tracking every Division 10 purchase, the three times we chased the low bid instead of the reliable one cost us roughly $47,000 in rework, expedited shipping, and idle site time within 90 days of delivery.
If a Division 10 order sits on the critical path, pay for the date. Get the date in writing. Move on. The rest of this is the case for that rule, the number that changed my mind, and the situations where I'd break it.
Why My Data Is Worth More Than My Opinion
I run procurement at a mid-size commercial general contractor—roughly 340 people, 20–30 active projects a year, about $180,000 of construction specialties spending under my signature. I've kept a running cost-tracking spreadsheet since my second month on the job because I got tired of arguing from memory.
Early on, I believed the standard procurement line: get three quotes, take the lowest. That's what the trade press says, that's what my first manager taught me, and that's what our pre-2022 policy required. It sounds correct until you actually measure what those orders cost beyond the invoice.
So I measured. I audited our 2022–2024 spending across specialties construction—every invoice, every change order, every "we'll work around it" schedule adjustment. The finding: the savings on low-bid Division 10 orders were offset by 30% or more within 90 days. Mostly through four channels:
- Re-rushing the same product after the first vendor missed
- Idle crew time waiting on a gap in the sequence
- Rescheduling crane, hoist, or fixture installation slots
- Occasionally accepting a substitute spec just to stay on schedule
That's the "everything I'd read said X, my experience says Y" moment for me. Everything I'd read said optimize unit price. My data said optimize total landed cost—and the biggest variable in total landed cost is whether it shows up when you need it.
One $650 Decision That Cost $13,100
A specific case, because "pay for certainty" is abstract until you see it with numbers.
In early 2023 we had a ground-up medical office building with a tight exterior envelope schedule. The exterior install crew needed a pallet of expansion joint cover systems on site by a specific Wednesday or the next trade couldn't mobilize.
Two quotes came in:
- Vendor A: $4,200, lead time "about 3 weeks, maybe 4."
- Vendor B: $4,850, delivery date written into the PO.
Difference: $650. A 15.5% premium. I took A.
The upside was obvious—$650 is $650. The risk was that "maybe 4 weeks" was really a 6-week delivery dressed up as optimism. I asked myself whether $650 was worth potentially missing the mobilization window. At the time, I told myself the risk was small.
Vendor A shipped in 5.5 weeks.
Here's what the delay cost:
- 9 working days of idle site supervision and equipment: $4,050
- Crane track and hoist rescheduling: $1,800
- Two-day milestone slippage penalty from the GC: $5,000
- Saturday crew to catch up the sequence: $2,250
Total: $13,100. Against $650 in saved invoice cost.
Looking back, I should have paid for B and never thought about it again. At the time, I was running the calculation I'd been trained to run—lowest unit price wins—without any model for what a missed date actually costs.
There's a second version of this lesson that showed up in wall protection, and it's why I now insist on written dates even for finishes. We tried a new Acrovyn panel supplier to save about $300 on a corridor package. The panels came in with maybe a Delta E 4–5 color shift off the original. Per Pantone's matching guidance, anything above Delta E 4 is "visible to most people." It was visible on the floor. We ate the redo—about $4,100—to keep the client from noticing during punch walks.
Not catastrophic. Not ideal. Expensive in a way that doesn't show up until you're standing in the corridor with the owner.
What Rush Payments Actually Buy
The reason people discount rush fees is that they treat the fee as buying speed. It's not. The premium buys certainty.
Vendor A's "about 3 weeks, maybe 4" was an honest estimate. It just wasn't a commitment. Vendor B's date was a commitment. The $650 bought me the difference between a forecast and a promise.
You can model this on any Division 10 order with two inputs and one question:
- What does a week of delay cost us? (In my data: usually 5–15x the price difference.)
- What's the probability the low bid actually delivers on time?
- Does the expected cost of delay exceed the premium?
Run that on the expansion joint case: 20% chance of a one-week slip, $13,100 cost per slip, expected loss ≈ $2,620. Premium: $650. That's a 4x improvement in expected cost for the same scope. I've seen site supers make that call on instinct and be right, and I've seen them be wrong. I'd rather defend the number in front of a project executive than defend a hunch.
When I Don't Pay the Premium
I want to be careful not to sound like I'm telling you to always pay more. There are four situations where the certainty premium is a waste of money:
- The schedule has real buffer. If the site has three weeks of slack before the next trade needs the material, the premium buys you nothing. Rush on demand, not by default.
- The spec isn't locked. If the design team is still finalizing the expansion joint substrate detail, paying for a certain delivery date just gets a wrong product to you on time.
- The product is genuinely interchangeable. Standard kick plates, base corner guards, generic thresholds. Those are often stocked items, and the premium buys you less reschedule value than you'd think.
- The vendor has an established track record. If a supplier has hit 40+ POs in a row on time, you're paying their reputation, not a specific commitment. That's a different transaction.
The other edge case worth flagging: a rush fee is not a guarantee unless it's in the PO. Verbal assurances from the rep count for nothing when the delay hits. We've had three "yes, absolutely Friday" conversations turn into Monday, Wednesday, and the following Monday. Put the date in writing or don't call it a premium—call it a fee.
What We Changed in the Process
We adjusted our internal approval flow this year: any Division 10 order on the critical path that's due within 10 working days of the next trade's mob date triggers an "expedited spec" review before release. The dollar volume is small. The behavior change is large.
I want to say the change saved us around $9,000 this year in avoided rework, though I'd be guessing—it might be closer to $8,500, and not all the savings map cleanly to the spreadsheet. The direction is right even if the number's fuzzy.
If your construction-specialties procurement still runs on "take the lowest bid" and you've never modeled the cost of a missed delivery, you're not really optimizing. You're just not measuring the loss.