"Get three quotes, pick the cheapest one with a reasonable lead time, sign the PO." That was my process for the first four years I handled chemical process equipment sourcing. It worked fine for pumps, valves, gaskets, small skids.

Then we ordered a 12,000-liter jacketed reactor with an agitator assembly. The quote was clean. Delivery was 18 weeks and stayed on schedule. The unit showed up on time, passed visual inspection, got signed off by our receiving team.

It then sat in our yard for eleven weeks because the agitator drive flange didn't clear the mezzanine steelwork over the install pad. Nobody had checked the elevation drawing against the actual site. Nobody had asked.

That wasn't a supplier problem. That was a me problem, and a process problem, and a scope problem — three things I didn't understand were different until the invoice for the rigging crew landed on my desk.

What You Think the Problem Is

When chemical process equipment sourcing goes sideways, most buyers — including me for years — blame the wrong things. "The manufacturer was too slow." "The quote was higher than expected." "The installation caught us off guard."

Those are symptoms. They're also convenient, because blaming the supplier means you don't have to look at what you actually bought.

Here's the question I should have asked on day one: who owns the interface between the equipment and everything it connects to? The nozzles, the foundation loads, the utilities, the control loop, the maintenance access clearance, the commissioning sequence.

The answer at most mid-sized plants, I've since learned, is nobody, specifically. It's assumed to be "the contractor" or "the supplier" or "engineering." It gets picked up by whoever notices first, usually after the equipment is already on site.

What's Actually Going Wrong: Three Purchases With One Nameplate

Chemical process equipment is not a product you buy the way you buy a motor. It's three separate purchases that happen to share a nameplate and often share a vendor name.

Purchase one: the design. Specifying the vessel, the internals, the material of construction, the corrosion allowance, the nozzle schedule, the control philosophy. This is engineering work, usually done against a design basis, not a catalogue.

Purchase two: fabrication. Cutting steel, welding to ASME BPVC Section VIII, applying the required NDE, hydro testing, painting. This is manufacturing work, governed by the code and the approved drawings.

Purchase three: installation. Foundation, piping tie-ins to ASME B31.3 requirements, electrical, instrumentation, commissioning. This is construction work, and it's where most of the surprises live.

Three different disciplines. Often three different companies. Frequently three different contracts, sourced at three different times, by three different people who may not have talked to each other beyond an email chain.

ASME Section VIII tells you how a vessel should be designed and stamped. ASME B31.3 tells you how the piping around it should be built. API 660 tells you what "acceptable" looks like for a shell-and-tube exchanger. TEMA will give you the mechanical standards. What none of them tell you is who is responsible when the design-basis nozzle elevation doesn't match what's actually in the field.

That's called interface risk. It's unglamorous, it doesn't show up on a schedule until it explodes, and in my experience it's the single most expensive category of mistake in chemical process equipment sourcing.

The Assumption That Cost Me $184,000

Let me be specific, because vague lessons don't stick.

In September 2021 we were replacing a shell-and-tube exchanger on a cooling loop. The datasheet had the standard details — duty, area, materials, TEMA type. We sent it out for quotes to four chemical process equipment suppliers and manufacturers. Three came back. We picked the middle one, not because it was cheapest (it wasn't) but because they had a reasonable lead time and a solid references list.

I assumed "same specification" meant identical to what came out. Didn't verify. Turned out our existing unit had a non-standard nozzle orientation — a site-modification from about eight years earlier that was never fully redrawn. The new exchanger arrived exactly to spec. The spec was wrong.

The re-work cascade: piping redesign, new spool pieces, a revised civil load calculation, and a seven-week production deferral on the loop. Rigging alone cost around $18,000 between the removal, temporary storage, and reinstallation. I want to say the total impact was somewhere around $184,000, but I might be misremembering the deferred-production portion — that number is soft.

The hard number I do remember: about 74 hours of my own team's time over four months, cleaning up a mistake that a two-hour site survey would have prevented.

Here's the thing I still find a little embarrassing to admit. That wasn't my first mistake. In my first year (2018) I ordered a small glass-lined reactor without confirming the design house's interpretation of "agitation rate." The vendor built to their standard. Their standard and our process requirement were about 30% apart on impeller tip speed. That one cost us roughly $31,000 in re-work and a three-week delay during a critical campaign. I told myself I'd learned the lesson. What I'd actually learned was to double-check one specific kind of detail, not to change how I thought about sourcing.

Why This Is Getting Harder, Not Easier

If you asked me five years ago, I'd have said the fix was "use a better supplier." I don't think that's true anymore, and I don't think the industry can keep pretending it is.

What was best practice for chemical process equipment sourcing in 2019 doesn't fully apply now. Supply chains are more fragmented. Fabrication shops that were in-house at manufacturers 15 years ago are now subcontracted two tiers down. Design work is often carved out to specialized engineering firms. Installation is a separate competitive tender. Each split is efficient in isolation. Each split also adds another seam where responsibility can fall through.

The fundamentals haven't changed — a vessel still needs to be designed to a code, built to a drawing, and installed to a site survey. But the execution of chemical process equipment design, chemical process equipment installation, and the sourcing that ties them together has fragmented enough that the old assumption of "the supplier handles that" no longer holds. It may never have, honestly. We just got away with it when the equipment was smaller.

I'd argue the buyer's job changed even though the job description didn't. What used to be "choose the right manufacturer" is now closer to "make sure the seams between the manufacturer, the designer, and the installer are owned by someone who isn't me, and then verify that they actually did."

The Checklist That Replaced My Gut

After the third failed handover in Q1 2023, our team wrote a pre-award checklist. It's not elegant, but it's caught 47 potential interface errors in the past 18 months, most of them small, a few of them not.

The core of it:

  • One named owner per package. Design, manufacture, and install can be different companies. They cannot be three separate signatures with no single point of accountability. Whoever signs the sourcing package owns the interfaces, even the ones they don't technically control.
  • Design basis, not datasheet. Before any chemical process equipment manufacturer quotes, confirm they have the design basis document, not just the specification sheet. If they push back on reading it, that's information.
  • Site walk-down before purchase order. Physically verify nozzle orientations, foundation elevations, and tie-in points against the design. Not against the drawing. Against the site.
  • FAT scope agreed upfront. Factory acceptance testing should be in the purchase order, with witness rights, and with criteria that reflect the design basis. Saying "per code" is not criteria.
  • Installation interface matrix. A one-page table: for each connection point, list who supplies, who installs, who inspects, who signs off. It takes an afternoon. It saves months.

None of this is revolutionary. Most of it would be obvious to anyone who's been burned enough times. The point isn't that the checklist is clever — it's that I stopped trusting that somebody would catch things, and started treating chemical process equipment sourcing as one decision instead of three separate ones.

If you're where I was six years ago, ask yourself one question before your next equipment order: if something goes wrong at the interface between the design, the fabrication, and the installation — who is on the hook? If the answer is "we'll figure it out," you already know how much that costs.

I've got the receipts.