I Lost $890 on a Small KSB Pump Order — What Two Misread Curves Taught Me About Not Ignoring Small Clients

It was a Thursday morning in March 2022, and two small booster pumps were screaming on our test bench. Not a huge batch — just two units. But they were vibrating hard, the pressure gauge needle was bouncing around, and the client's maintenance supervisor was standing right next to me watching.

Seven hours earlier, I'd thought this job was a no-brainer. A local food-processing plant needed to replace two aging booster pumps. I picked a model, matched the quoted numbers, and sent the quote. Simple. Except the pumps couldn't operate at the duty point their system actually required.

I'd read the pump curve wrong. The best efficiency point (BEP) on the model I picked wasn't anywhere near where I assumed it was. Two tiny pumps, $890 in rework, and a full week of production delay for the client.

I've handled pump and system orders for about nine years now. Personally, I've made (and documented) eleven significant mistakes, worth roughly $14,000 in wasted budget combined. The pre-check list our team uses today? It started with this order.

The order that looked too simple to mess up

Around that time, we were buried in a big mining project order — multiple multistage pumps, submersible sewage pumps, the whole technical review cycle with performance testing and documentation. In that context, a local client asking for two small booster pumps felt, honestly, like background noise.

The inquiry came in as a one-pager. Flow around 12 m³/h. Head maybe 40 meters. Single-phase 220V. DN50 connections. I looked at it, mentally picked a KSB ixo pro 6, checked the price, and quoted it back. Done.

Looking back, I'd assumed three things — and every one of them was wrong:

1. "It's a small order, it doesn't need much checking." Wrong. Simpler orders make it easier to skip the system-level questions.
2. "The client gave me enough data." Wrong. The client gave me their understanding of the system, not the actual operating conditions.
3. "If the curve looks close enough, it'll work." Wrong. Curves don't work that way.

The turn: pumps installed, pumps not working

Pumps arrived. Crew installed them. Wired them up, opened the valve — and by that afternoon, something was clearly off. I got the call around 5 p.m. The client said the pump sounded like it had rocks inside it.

I drove over. One look and I knew we were in trouble. The pumps were running in severe cavitation (that sound — I still hear it sometimes, like someone tapping pebbles inside the casing). Discharge pressure was fluctuating way outside spec. Motor temperature was climbing fast.

We shut them down on the spot. Pulled the suction strainer: fine sand mixed in the water. Then we walked the actual system — seven elbows on the discharge line, a sudden reducer, and the suction-side liquid level sitting roughly two meters below the pump inlet. None of which the client had mentioned.

All the problems were right there. But my first thought wasn't "what did I misread?" It was "why didn't the client tell me?"

That thought was the wrong one. How would a food-processing plant know to tell a pump vendor "oh, by the way, our suction-side liquid level is low"? That's my job to ask.

The damage and the real lesson

End result: both units had to be swapped for a different KSB model with better suction performance, plus freight, plus the cost of renting a temporary standby pump. Total: $890. The client lost a week of production. No formal claim, but the relationship cooled off noticeably for a while.

The bigger cost was trust. The client said something I still remember: "I know our order is small. But when equipment stops, one day of downtime hits us the same as one day hits your big customers."

If you handle small pump orders — on either side of the table — here's the core of the checklist I built after that week:

• Draw the system curve before picking the pump. Don't stop at "flow and head." Find out the actual piping: elbows, reducers, liquid level, valve positions. Standards like ISO 9906 define how pump performance must be verified, but only if you've nailed down the real duty point first.
• Confirm suction conditions separately, every time. NPSH available is the number that kills pumps. If the client says "the water level is around a meter or so," you ask again.
• Don't simplify the process for small orders. Filling out a full inquiry sheet takes twenty minutes. Skipping it costs a week.

I'm not saying we can eliminate every mistake. Nobody can — not in a field where system conditions are messy and pump curves are only one part of the picture. Experienced engineers will tell you there's always some gap between the curve and the real installation, and judgment fills that gap.

But the lesson here was simpler than that. Small-order clients deserve more careful handling, not less. Larger customers usually have their own technical team checking parameters on their end. Small customers don't. Their only check is whether the supplier asks the right questions in those first few minutes.

Now, whenever a new pump inquiry lands, my first move isn't picking a model. It's sketching the system. That habit started on a Thursday morning in March 2022, watching two little pumps shake themselves apart on the test bench.