Almost every water treatment procurement I review starts with the wrong number. It starts with RO machine price per cubic meter, or "filter press price" from three supplier emails side by side, or the lowest bid on a skid-mounted RO system. By the time the spreadsheet is closed, the buyer has compared equipment. What they haven't compared is the project.
I should be clear about where I sit. I'm a quality and brand compliance manager at an industrial equipment company. I review incoming equipment packages and the documentation that travels with them — roughly 200+ packages a year, across reverse osmosis trains, ultrafiltration skids, chemical dosing systems, settling tanks, and dewatering lines. In 2023 I rejected close to 30% of first deliveries. Maybe 27%, I'd have to pull the log. The number matters less than the pattern: a third of what passed invoice review failed acceptance review.
That's the opinion I'll spend the rest of this on. The purchase price of water treatment equipment is a weak predictor of total cost, and the quotes that clearly list what is excluded are almost always cheaper in the end — even when the total on the page looks higher.
Where a low quote actually shows up
Commissioning is the line item nobody prices
In 2024 we took delivery of a commercial reverse osmosis system quoted at a genuinely competitive price. The datasheet said 75% recovery. On commissioning the train held 62% and the silt density index on the feed sat close to 5 — above the 3-to-5 ceiling most membrane warranties reference, and measured per ASTM D4189. Normal for their design, the vendor said. "Within industry standard."
It was within the loose edges of the standard. It was not within our spec, and it was not within the membrane manufacturer's warranty terms. We rejected the performance test, they reworked the pretreatment at their cost, and every contract since then has carried explicit SDI, recovery, and permeate quality acceptance thresholds. Learning by rejection. Not ideal, but workable.
Here's the part buyers miss. The delta between that quote and the next one up was around $9,000 on a system of that class. The pretreatment rework, the delayed handover, and three weeks of my team's time came to more. We never invoiced that difference to anyone. It just left the building quietly.
The dosing system is where "cheap" fails quietly
Dosing is the cheapest component in the plant and the most expensive one to get wrong. A coagulant dosing system, a chlorine dosing system in a water treatment plant, a biocide dosing system on a cooling loop, boiler chemical dosing — the pumps themselves are a small fraction of the capex. The failure mode is never a loud failure. It's drift.
One case still bothers me. We specified reciprocating dosing pumps for two effluent treatment plant chemical dosing skids. The lower bid used an oversized pump running at low stroke, the higher bid used a pump sized correctly. We bought the lower one, assuming bigger motor equals more headroom. What we got was a diaphragm pump running far below its stable range, with calibration drifting 8–12% across a shift. Calibration drift in coagulant dosing doesn't announce itself. It shows up as dirty filtrate, then as backwash frequency climbing, then as someone asking why the settling tank in the sewage treatment plant stopped producing clear overflow.
Never expected the cheaper dosing package to hold calibration better on the sister skid. Turns out that unit was oversized and running at about 30% stroke — which is close to where diaphragm metering pumps are most stable. The "premium" pump was running near 85% and hunting. Sizing beat brand. That one surprised me enough that I wrote it into our specification notes.
Settling and sludge: the cost that arrives in year two
Primary sedimentation tanks, secondary clarifiers, final settling tanks — these get bought on surface loading rate and then forgiven for everything else. The specifying question is never "what's the tank price." It's "what's the sludge removal rate, and what happens to the cake."
We once accepted a plate and frame filter press on a low bid. The press worked. The plates didn't. Cake moisture ran high enough that hauling costs rose, and plate cracking started in year two because the feed pressure cycle hadn't been matched to the plate material. Filter press price is a headline number. Filter plate replacement cost is the number that lives in your maintenance budget for ten years. Nobody puts them in the same column.
"Spare parts available" is not a service agreement
I see this clause constantly on industrial reverse osmosis equipment and on water transfer pumps. "Spare parts available upon request." No lead time. No stocking commitment. No named consumable list. It reads reassuring and commits to nothing.
Ask instead for the actual consumable list with replacement intervals — cartridge filters, membrane elements, dosing pump diaphragms and valves, press plates, UV lamps if present — and the quoted lead time on each. A vendor who can answer that in writing has thought about year three. A vendor who can't is selling you a box.
The objection I hear every quarter
"That's fine in theory. Our budget is real and it's fixed."
Fair. I've been on the approving side of a capital request that got cut, and I've defended a spec I couldn't fully fund. But there's a difference between buying less scope and buying less spec. Buy the full specification, then phase the scope — pretreatment first, second RO pass later, the second filter press next fiscal year. What you don't do is accept a lower SDI limit, a thinner press plate, or an unstated electrical scope because the package came in under budget.
The other objection is that I'm arguing for paying more. I'm not. I'm arguing for asking a different question: not "what's the price" but "what's not included." I learned to ask that before I ask for a number, and the answers changed how I read every quote that followed.
A vendor who lists every exclusion, every consumable, and every boundary of scope up front — even when the total looks higher — has usually cost you less by the end of year one. The one who quotes a number and nothing else isn't being generous. They're deferring the rest of the price.
What I put in the document now, before price is discussed
Three things, in this order: acceptance criteria, excluded scope, and consumable schedule.
- Acceptance criteria — measurable, testable at commissioning, tied to payment. Recovery, permeate conductivity, SDI, effluent turbidity, cake moisture. If it can't be measured on site, it isn't a criterion.
- Excluded scope — freight, rigging, electrical, civil, media loading, startup labor, training. Ask each bidder to list exclusions explicitly. Compare those lists, not the totals.
- Consumable schedule — item, interval, unit price, lead time. This turns the five-year number from a guess into arithmetic.
Then normalize. Rebuild all three quotes into one cost-per-cubic-meter-per-year model with the same consumable assumptions, the same electrical scope, and the same installation responsibility. It takes a few hours. It has never once left the ranking unchanged.
Where I still land
Even after all of this, I'll admit the decision is uncomfortable. I went back and forth between a European-built skid-mounted RO system and a domestic one for three weeks. The European unit was roughly 20% more on the equipment line and delivered a month later. I chose it — not for the brand, but because its quote itemized exclusions in writing, down to the membrane cleaning chemicals and the boundary of the electrical scope. The other quote excluded nothing. Which reads as generosity until you ask.
Did I second-guess it? For about five weeks, until the commissioning test ran clean on the first attempt. That's not proof I was right. It's one data point and I'll take it.
So my position stands, and it isn't subtle: the unit price of water treatment equipment is a comparison of invoices, not of projects. Stop ranking quotes by total. Rank them by how honestly they disclose what the total leaves out. The vendor who tells you what you're not getting is the one who's actually told you the price.