Buy the Controls, Not Just the Tank: A Buyer's Case for Spending More on Water Treatment Automation

Four years of writing purchase orders for RO units, dosing systems, and filter presses taught me that negotiating on unit price is the expensive way to buy water treatment equipment. Here's where the money actually leaks, and what I now refuse to sign off on.

The cheapest line item on a water treatment quote is usually the one that decides whether the system works. After four years of writing purchase orders for industrial water purification systems, I've landed on a view that doesn't make me popular with every supplier: procurement should spend its negotiating energy on the controls, the instrumentation, and the commissioning scope — not on shaving six percent off the tank.

Some qualification first, because I don't want this read as an engineering opinion. I'm not an engineer. I'm the office administrator for a 240-person food and beverage plant. I manage capital equipment and MRO ordering — roughly $1.6 million annually across 14 vendors, from a skid mounted RO system down to boxes of filter cartridges — and I report to both operations and finance. My view of water treatment equipment is a buyer's view: invoices, acceptance criteria, warranty language, and the 6 a.m. phone call when something isn't dosing.

Argument one: the pump is cheap, the chemical isn't

When we priced a corrosion inhibitor dosing system for our chilled water loop back in 2022, the industrial metering pump was a four-figure line item. The inhibitor itself was a five-figure annual spend. Everyone in the room negotiated hard over the pump and barely glanced at the dosing strategy.

Then we finally did what we should have done first. For six weeks we logged actual dosing rate against plant flow. Our feed was running above setpoint more often than not — not because the pump was failing, but because it was set to a fixed stroke and our flow isn't fixed. Any chemical dosing setup that can't see flow is, in practice, a system that over-doses on light days and under-doses on heavy ones.

That's the argument in one sentence: a flow-paced dosing system tends to pay for its own instrumentation out of chemical savings, while a fixed-stroke one quietly bills you the difference every month. It's also why I now demand the flow-signal wiring in the quote instead of assuming it's included. It usually isn't.

Argument two: manual is fine until the shift changes

Manual dosing doesn't fail dramatically. It fails at 10:55 p.m., when the outgoing operator tops up a tank by eye and the incoming operator has no idea what was added or when. Same story with biocide dosing on a cooling tower chemical feed — the control that matters isn't the pump, it's the interlock and the log.

Two of our three most expensive water treatment incidents since 2021 were operator-dependent, not equipment-dependent: one wrongly mixed batch, and one dosing pump left on manual after maintenance. Neither was a hardware failure. Both were process failures, and both were cheaper to prevent than to clean up.

This is where I've become a bore at internal meetings. A dosing system that logs its own rate and alarms on deviation is not a luxury item. It's the difference between reading a report and reconstructing an incident from memory.

Argument three: the counterintuitive one

I used to treat the skid premium as padding. Pre-piped, pre-wired, factory-tested — surely that's a markup for convenience, I thought, and we have a maintenance team.

Then we compared two projects. The field-assembled package looked 12% cheaper on paper. Field assembly, piping, wiring, and commissioning ran nine extra days, plus three days of lost production. The skid mounted unit we bought in 2023 was the more expensive quote and the cheaper project. Same logic applies to skid ultrafiltration packages and packaged commercial reverse osmosis units — the equipment price is roughly 60% of the real cost, and the rest hides in labor and downtime.

What I no longer sign off on

These aren't technical requirements so much as things I've learned to demand in writing:

  • No flow or conductivity signal wired to the dosing pump — rejected, unless the duty is genuinely constant.
  • No stated feedwater analysis for an RO unit. Membrane selection without a water analysis is a guess with a warranty attached.
  • No written commissioning and acceptance criteria. If nobody can tell me what 'working' measures, the invoice stays open.
  • No local display and no manual override. Automation without a fallback is how you get the 3 a.m. call.
  • No spare parts list with lead times — filter cartridge part numbers, membrane model, seal kits. Not a catalog, just the list.

On sedimentation and sludge handling, the same instinct applies. When we looked at sedimentation tank types for our wastewater upgrade, the difference between conventional and lamella designs mattered far less than whether anyone had specified sludge blanket monitoring and a realistic desludging schedule. A settling tank isn't a passive box; it's a process with a failure mode.

The objections I hear, and my answers

Our plant is too small for automation.

Then buy a smaller automated unit. Commercial RO units and small dosing systems with basic controllers exist at most sizes. What you shouldn't do is buy a full manual system and plan to upgrade later. The later never gets budgeted.

Our operators prefer manual control.

They can have manual override. That's not the same as manual default, and conflating the two is how the 10:55 p.m. problem happens.

The vendor says it's fully automated.

Ask what measured variable drives the loop, and what the fail position is. Per FTC advertising guidance (ftc.gov), performance claims must be truthful, substantiated, and not misleading — in buyer language, 'fully automated' should come with a sentence about what happens when the analyzer drifts. The good vendors send that sentence unprompted.

Honestly, I'm not sure why skid pricing varies so widely between vendors for what looks like an identical package. My best guess is that a lot of quotes quietly exclude controls and commissioning, so the numbers aren't comparable until you normalize them. Every spreadsheet I built said go with the lowest bid on our 2024 RO replacement. My gut said the slow pre-sale responsiveness was a preview of slow support. We went with the middle bid, and I later learned the low bidder had told another department they couldn't service units more than 300 miles out. So glad we paid for the factory commissioning visit instead of doing it ourselves.

Where this lands

Buy the controls, not just the tank. Not because automation is fashionable, but because the recurring costs — chemical, downtime, operator time, my own time chasing invoices — dwarf the capital line everyone argues about. Efficient process control is a competitive advantage in a plant that runs five days a week and gets paid per case shipped.

That said, I'll defend one traditional habit: get the water analysis, read the commissioning report, and keep a paper copy. No amount of instrumentation replaces knowing what's actually in your feedwater.

Pricing and lead times referenced here reflect quotes we collected in Q1 2025. Equipment markets move, so verify current rates and specifications before budgeting. This is a buyer's perspective, not engineering advice — for sizing and process design, use your engineer.

CB

Cosima Brenner

Cosima Brenner covers logistics equipment procurement across racking, conveying and sortation, AGV and AMR fleets, forklifts and pallet handling, dock lifting, and warehouse control integration. She compares load cases, throughput assumptions, energy and charging windows, building interfaces, spare parts, and total cost of ownership, treating ISO 12100 and ISO 13849-1 as scope checks that must be verified for the target market. She helps capital procurement managers, supply chain engineering directors, and intralogistics engineers build comparable specifications, structure FAT/SAT acceptance, and question vendor claims.