At 2:37 on a Friday afternoon, the call came in. A client's slurry pump had seized at a mine site, and they needed a replacement on-site by Monday morning. Normal lead time for that pump: six weeks.
The call didn't surprise me. The failure didn't either.
What surprised me was that three different people had already seen the vibration readings pointing to that pump. The data had been sitting in their maintenance system for two weeks. Nobody acted. So instead of a planned replacement, they got a rush order—with air freight, overtime, and a very expensive Sunday afternoon.
The surface problem: emergency orders cost more
Everyone knows emergency orders cost more. Rush fees, freight, overtime, installation time. The premium is not the real problem. The real problem is why the order gets rushed in the first place.
Why does a site end up at 2:37 on a Friday with a dead pump and a Monday deadline? Because of three mistakes that were already in motion weeks before.
It took me four years and about 200 rush orders to understand that. In my role coordinating emergency replacement parts for energy and mining equipment, I've seen the same pattern repeat itself across coal, copper, and lithium projects. The pattern has three pieces. I think of it as the trio of mistakes.
Mistake #1: We treat lead time as a surprise
The conventional wisdom is that emergencies happen because machines fail unpredictably. My experience suggests otherwise.
Most 'sudden' failures showed up somewhere first. A vibration trend. A temperature reading. A note from an operator who heard a noise at 3 a.m. and didn't want to be wrong about it.
I'm not 100% sure why we don't act on these signals. But I know what happens when we don't.
A white-haired reliability engineer once told me, "There are no breakdowns. Only costs you put off paying." I didn't like that at the time. Now I quote it.
Think of it this way: if you were planning to ski at Milano Cortina in 2026, you wouldn't wait until February 2026 to book a hotel. But we do the equivalent with critical spares—because the failure hasn't 'happened yet.'
Mistake #2: We get loose on specs exactly when we need to be tight
When a pump is down, the temptation is to call and say, "I need one of those pumps." The person on the other end of the phone needs more than that. Shaft size. Flange rating. Motor frame. Voltage. Coupling type.
One missing number can turn a two-day order into a six-day rework.
The fix is boring: a locked spec sheet for every critical part. Dimensions, drawings, supplier part numbers, revision dates. Update it when equipment changes.
5 minutes of verification beats 5 days of correction. I've said it so often my team rolls their eyes. And then a rush order shows up with a wrong bolt pattern, and they stop rolling their eyes.
Mistake #3: We treat vendors as transactional
If the only time you call a vendor is when you're desperate, you're not a partner. You're an interruption.
Discount vendors rarely match the responsiveness you need in a crisis. But there are exceptions—and the exceptions are worth finding early.
Based on our internal data from 200+ rush jobs, the vendors that came through consistently weren't the biggest ones. They were the ones we'd already briefed on our equipment, our site conditions, and our approval process. We had a relationship before the fire started.
The surprise wasn't the price difference. It was how much hidden value came with those relationships. Responsiveness. Honest lead times. A willingness to call me and say, "That spec doesn't make sense—can I send a field engineer tomorrow?"
What this really costs
I've seen a missed deadline trigger a $50,000 penalty clause. I've seen a $2,000 part create $120,000 in downtime. I've seen a team pay $800 extra in overnight freight and still lose the project because the motor was wrong.
Here's the part I don't talk about much: I still kick myself for the time I took a vendor's verbal lead time at face value. It wasn't malicious. It just wasn't real. The order shipped three days late. The client's alternative was a full shutdown.
To be fair, some rush orders are genuinely unavoidable. Lightning does strike. But most—and I'm reluctant to put a number on it because every site is different—most are self-inflicted.
In Q3 2024, we reviewed 34 emergency orders across three sites. Twenty-six of them traced back to a known issue that had been documented but not acted on. That number still bothers me.
The fix is boring
Prevention doesn't make a good story. A checklist is not glamorous. But the checklist I created after my third mistake has saved us an estimated $8,000 in potential rework.
- Monitor the data you're already collecting. Vibration, temperature, oil analysis. Set thresholds. Assign owners.
- Lock your critical specs. One source of truth for every part that can stop production.
- Build vendor relationships before the fire starts. Annual audits. Site visits. A conversation that isn't an emergency.
That's it. Simple. Not easy.
Nobody plans to have a failure. But planning for one is the difference between a bad week and a shutdown. The question isn't whether your equipment will fail. It's whether you'll have the right part, the right spec, and the right vendor ready when it does.