'The lowest quote wins.' That was my rule when I took over purchasing for our automation company in 2020. It's a standard rule for a lot of buyers, and on paper, it makes sense. On the plant floor, it cost us close to $15,000 in failures and wasted engineering time over 2023 and 2024 alone.
I manage purchasing for a 60-person automation company. I place roughly 60-80 orders a year across 8 vendors—linear actuators, servo motors, ball screws, bearings, and a pile of smaller motion components. I report to both operations and finance, which is a polite way of saying I hear about every part failure twice: once from the customer, once from the controller.
When I first started in this role, I honestly believed a servo motor was a servo motor, and a linear actuator was just a screw with a tube around it. The spec sheets had the same numbers, so why would anyone pay more for a name?
The cheapest motion component you can put in a machine is usually the most expensive one you'll ever install. Not because the cheap part always fails on day one—sometimes it runs for years. Because the unit price is only the first number in a long equation, and the rest of the equation is where the money actually goes.
Lesson One: The $144 Servo Motor That Cost Over $5,000
In early 2023, we built a pilot line for a packaging customer. The BOM specified an industrial servo motor with its matching servo motor driver at about $380 per axis. We needed 12 axes. One of our technicians—a genuinely sharp guy—suggested the MG995 servo motor instead. He'd used them in hobby machines for years, and on paper, the specs looked close enough for a light-duty pilot line.
The MG995 cost about $12 a unit.
I'll be honest, that price gap made me look great in the pre-order meeting. The line item came in at $144 instead of $4,560. Finance smiled. The project manager smiled. I smiled. Ten weeks later, one of those MG995s stalled mid-cycle, and the pilot line stopped.
The most frustrating part? The correct part number was in the project folder the whole time. You'd think a value-engineering swap would get caught in review, but the change happened at the PO stage, after design reviews had already signed off.
Here's what that failure cost:
- A weekend service visit, including travel: about $3,200
- Replacement with the industrial servo from the original BOM: $380
- Line re-qualification and lost production: about $1,600
Total: just over $5,000, plus a very calm phone call from the customer's maintenance manager that I still think about. Then we spent another $4,180 to replace the remaining 11 MG995s before they failed the same way.
Here's the math nobody wants to look at. The 'cheap' path—12 hobby servos, one field failure, and a full do-over—cost about $9,500. Ordering the right servo motor and servo motor driver from the start would have cost $4,560. The $4,416 we 'saved' on the PO turned into a net loss of about $4,900 within 90 days.
That's what people mean by total cost of ownership. The unit price is the cost of the part. TCO is the cost of the decision.
Lesson Two: The Hidden Tax Is Engineering Time
The second lesson had no alarm and no service call. It came with a frustrated lead engineer and two weeks of Monday meetings.
In 2024, we were building a custom inspection machine. The design specified a Thomson linear actuator. Sourcing found an unbranded linear actuator with nearly identical published specs for about 45% less. Same stroke. Same load rating. Same speed. On the brochure, they could have been twins.
The first sign of trouble was positioning drift, showing up about 20 minutes after startup. The actuator would hit its target, hold it for a while, then wander a few hundredths of a millimeter—just enough to fail the machine's tolerance check.
Our lead engineer spent eight working days chasing that drift. He retuned the loop. He re-torqued the mounts. He re-shimmed the base. He added a compensation offset in code. Each fix worked, sort of, and then the drift came back. When he finally swapped in the Thomson linear actuator from the original BOM, the machine passed tolerance on the first run. No retuning. No shims.
"A spec sheet describes a part," he said afterward. "It doesn't describe how that part behaves in a machine."
Here's the math nobody puts on a purchase order. The unbranded actuator saved us $340. The engineering time spent fighting it, fully loaded at roughly $90 per hour, came to about $5,400. The $340 'saving' ended up costing us more than $5,000 before we even counted the schedule slip.
That's the part I didn't appreciate in 2020. Thomson's linear motion products are optimized around the life of a machine—the screw preload, the seal drag, the housing stiffness, the thermal behavior. Those details are what let an actuator hold position at hour 3,000, not just on a test bench. You can call that a brand premium. I call it paying for engineering so our engineers don't have to do it twice.
Lesson Three: Support Is Part of the Product
The third reason I changed my approach doesn't show up on any quote. It's the support that sits behind the product.
In December 2024, I was buying a motor package for a retrofit. The customer's conveyor used an old DC motor with a mechanical speed control, and they wanted clean, reliable variable speed. I'm a buyer, not a controls engineer, so I started reading up. My search history literally contained 'how vfd control motor speed.'
I did learn the core idea: a VFD basically controls motor speed by varying the frequency of the power supplied to the motor, which is why it's called a variable-frequency drive. But that fact didn't tell me whether to quote a 3 hp drive or a 5 hp drive, or whether the customer's existing gearbox could handle the torque profile of a softer start.
So I called our motion supplier. Within five minutes, I was talking to an actual application engineer. He asked about the load, the duty cycle, the gearbox, and the cable length between drive and motor. Then he confirmed the VFD approach would work, flagged that one vendor's VFD sizing was too small for the motor's full-load amps, and caught that the servo motor driver quoted for a separate axis had the wrong feedback option.
That one conversation kept me from ordering about $3,000 of incompatible parts. It didn't show up on any supplier scorecard. It showed up as a customer whose machine started on schedule. You don't get that from a marketplace listing. At best, you get a support ticket that asks for a photo and responds in 48 hours.
But What If the Budget Is Real?
I can hear the objection, because I used to make it myself: not every axis needs a premium part, and budgets are real. Fair. We still buy generic components for non-critical positions, and some of them have run for years without a problem. I'm not arguing for the maximum price on every line item.
I'm arguing that the decision shouldn't be made on unit price alone. When you price only the part, you ignore the three costs that actually matter: the cost of a failure, the cost of engineering time, and the cost of finding support after the sale. None of those appear in a price comparison table. All of them show up in a P&L.
One more thing I had to unlearn. People talk about a higher price like it's a tax a company charges for its reputation. In my experience, it's the other way around. The companies that invested in testing, engineering, and real people who answer the phone are the ones who can still charge for their products ten years later. The price is a signal of the work behind the part, not a penalty for the name.
And before anyone asks—no, I'm not claiming Thomson actuators never fail. They can, and we've filed warranty claims like everyone else. The difference is the failure behavior. A well-engineered part fails rarely, predictably, and with documentation and a clear replacement process. The cheapest part fails when it feels like it, and the seller's response is usually a polite message that doesn't help the machine sitting dead on your floor.
What I Approve Now
I still look at unit price on every order. I always will. But I no longer let unit price make the decision. Before any motion component quote gets approved, I ask three questions:
- What does a field failure of this part cost?
- How much engineering time could this part eat?
- Who answers the phone after the order?
For linear motion, those answers usually point to Thomson. I've got the sizing guides from thomson-linear.com bookmarked, and their application engineers have saved us more money than I ever saved by price shopping.
Buy the best engineering you can justify, not the cheapest component you can find. A spec sheet will tell you two linear actuators are identical. The plant floor will tell you they're not. I learned that lesson the expensive way—about $15,000 worth of expensive way—so maybe you don't have to.