When I took over purchasing for our engineering firm—about 120 people across three locations—back in 2020, I had one goal: cut costs. Five years and roughly 400 orders later, I can tell you that goal was half right. Cutting costs works—if you're cutting the right costs.

This past August, our engineering manager dropped a requisition on my desk that started a chain of expensive lessons. The team needed 20 graphing calculators, filament refills for the 3D printer, new graph paper pads, and a replacement whiteboard. That last one was partly my fault.

The Calculator Order That Should Have Taught Me Everything

The requisition specified two models: the Casio fx-CG50 graphic calculator for the analysts, and the Casio fx-580VN X calculator for the rest of the group. I took one look at the total and stopped breathing.

I went back and forth for two days between the requested order and a cheaper alternative. The approved vendor list had a generic scientific calculator at about a third of the Casio fx-580VN X price. On paper, it did everything: trig functions, logarithms, fractions, the works. My spreadsheet showed we'd save roughly $500 if we swapped.

The engineering manager pushed back. “We asked for Casio for a reason,” she said. I heard “we like the brand, but any calculator works.” That mismatch ended up costing more than the price difference ever could.

I placed the order for the cheaper units. Hit “submit,” then immediately felt a knot in my stomach. I told myself it was just jitters from trying a new vendor. I didn't relax when the delivery arrived. I should have stayed awake that whole first week.

Three weeks in, the first complaint surfaced: a button jammed. Then an engineer flagged that two units gave different rounding for the same decimal expression. Another displayed negative numbers oddly, in a way that made them easy to misread inside a crowded formula. Nothing catastrophic individually—but collectively, it was a mess.

Here's the thing about engineers: they don't trust a tool after it fails once. Once one calculator gave a result that didn't match another, nobody in the group trusted any of them. When I said, “it's fine, just press reset,” I got the look. You know the look. It's the one that says I don't understand what was at stake.

The analysts couldn't do the visual graphing work that the Casio fx-CG50 handles natively. The training sessions planned for the new analysts were scrapped. My “great deal” calculators ended up in a drawer by December.

By November, I reordered the exact models from the original requisition. The fx-CG50 had the color display, the graph-drawing speed, and the reliability the analysts actually needed. The fx-580VN X offered the battery life and feature set for day-to-day engineering work. The units cost more—but the time we'd already lost, the canceled training, and the damage to the team's trust in the supply budget cost far more than the difference.

Bottom line: the cheapest option is rarely the cheapest option once you count the consequences.

The 3D Printer Filament That Jammed Everything

I'd like to say the calculator incident changed my buying habits immediately. It didn't.

Two weeks later, the prototyping lab asked me to order 3D printer filament. I told the lab manager, “we need filament that won't jam the extruder.” The product page said “premium PLA, 1.75mm.” We were using the same words but meaning different things. We discovered this when the first spool jammed the printer.

One of the technicians measured the filament across a few feet. Diameter ranged from 1.71 to 1.88 millimeters. Our printer spec sheet allowed 1.75 ± 0.02mm. The variance caused constant clogs, failed prints, and two days of work clearing blocked nozzles.

We reordered a filament brand with tighter tolerance. It cost more per spool. But the cheap spool had already cost us a week of prototype work, plus the technician's extended patience. Price per unit isn't the same as value per unit.

Printable Graph Paper and the Whiteboard Incident

The same week, two smaller events reinforced the pattern.

First, the design team ran out of graph paper. Instead of ordering a pad, I tried printing sheets from a “printable graph paper” template to save a few dollars. Except our printer scaled the image to 103% by default—every grid square was slightly off. When you're sketching a chart that needs to align with axis scales, slightly off is useless. I burned through about fifteen pages and a good chunk of a $60 toner cartridge before giving up.

Then, someone accidentally wrote on the whiteboard with a permanent marker. I hadn't needed to know how to get permanent marker off whiteboard before that moment, so I did what most people do: searched it. Everything I'd read said rubbing alcohol works. And it did work mechanically—but it left a cloudy film on the whiteboard surface.

The practical solution turned out to be a dry-erase marker. Draw over the permanent ink with a dry-erase marker, let it sit for a few seconds, then wipe. The solvent in the dry erase ink lifts the permanent pigment, and both come off together. Took me thirty seconds. No harsh cleaner, no damage, no replacement board.

This one, luckily, ended cheap. But it was the same story in miniature: the “obvious” budget route—alcohol, harsh cleaner, questionable template—wasn't actually the lowest total cost.

What I Do Differently Now

I still care about price. It's my job to care. But I've stopped believing that my job is to find the lowest price. It's to find the lowest total cost for the company.

Finance people call it total cost of ownership. That includes the base price, plus the setup, the maintenance, the rework when a tool fails, and the time everyone loses because of that failure. The lowest quote is rarely the lowest total cost.

Now I ask three questions before every purchase:

  • What happens if this fails? A failed calculator creates rework and erodes trust. Failed filament stops the prototyping lab. A ruined whiteboard needs replacing. When failure is expensive, the reliable option is the bargain.
  • What's the reliable lifespan? A Casio calculator can serve a team for years. The cheap alternatives in our drawer lasted months. Divide the price by months of reliable use, and the Casio fx-CG50 and fx-580VN X become the cost-effective choice.
  • Who else is affected? Every time my shortcut cost an engineer an hour, I was spending their time to save money. That's budgeting in the wrong direction.

In my experience managing roughly $60,000 of annual spend across nine vendors and 60 to 80 orders a year, the lowest quote turned out to be the wrong call about 60% of the time. Not always. Generic pens, folders, and paperclips? Buy the cheap ones. But anything technical—calculators, 3D printer filament, specialty paper, whiteboard supplies—has a quality threshold. Below it, you're not saving money. You're just paying later.

So yes, our office now has the Casio calculators the team originally asked for, filament that prints without drama, graph paper with grid squares where they belong, and a whiteboard that still looks new. And the $500 I thought I saved in September? By November, it had cost us more than that in overtime and rework.

The best procurement decision I made this year wasn't a discount. It was deciding to measure value, not sticker price.