I used to fix things by buying parts. It didn't work.
Here's the rule I follow now, and I'll say it upfront because it took me four years to learn it: test before you buy. Not sometimes. Not when it's convenient. Every time—especially when the part looks obviously bad.
I've wasted roughly $3,400 over four years replacing things that were fine. Not because I'm careless. Because I was in a hurry, and buying a part feels like progress. Testing feels like delay.
It's the other way around. Testing is the fastest path. Buying is the gamble.
Exhibit A: The $38 mower repair that fixed nothing
In September 2019, my Bolens walk-behind mower wouldn't start. I did what most people do: ordered a Bolens lawn mower spark plug, a new air filter, and a can of carb cleaner. Total came to $38 and change. In the same order—because I was already paying shipping—I threw in a 2 stroke spark plug for my string trimmer, figuring it was due for a change anyway.
Neither plug was bad. I checked them both with a multimeter after the mower still wouldn't start. The actual problem was a loose wire on the safety switch under the handle. Five minutes of testing would have found it. Instead I waited six days for parts I didn't need.
That $38 wasn't the expensive part. The expensive part was six days of not knowing, and the nagging feeling that I'd bought my way out of a problem instead of solving it.
I want to say that was the last time. It wasn't. But it was the first time I wrote the mistake down, and that turned out to matter more than the money.
Exhibit B: The Leviton switch that was always fine
Fast forward to 2021. I was rewiring the switches in my basement—three Decora-style units, nothing complicated. I pulled up a leviton light switch wiring diagram single pole on my phone, glanced at it for maybe ten seconds, and got to work.
Two hours later the breaker tripped the moment I flipped the switch. I'd reversed the line and load wires. The diagram was right there on my screen the whole time. I just thought I could figure it out faster than reading it.
So I moved to the three-way circuit and made it worse. I skipped the leviton 3 way switch wiring diagrams because I assumed three-way worked the same as single-pole. It doesn't. The light stayed on no matter which switch I flipped—both positions, always on. I bought a replacement switch before I tested the old one. The old one was fine. The wiring was not.
If I remember correctly, the replacement cost me $22 and the wiring diagram would have cost me $0 and about four minutes of reading. I want to say I've stopped doing this. I've mostly stopped. The checklist helps.
Exhibit C: The counterintuitive multimeter test nobody teaches you
Here's the part most people get wrong about how to test voltage with a multimeter. Almost every tutorial teaches you to check whether a wire has power. That's the easy test. The one that actually saves money is the opposite: check whether a wire that should be dead is actually dead.
Because that's where the expensive mistakes hide. Phantom voltage from an induced current. A dropped neutral upstream. A switch that reads "off" at the fixture but is still live at the box. I once spent $90 on a replacement switch because a circuit read dead at the light. The neutral had come loose two boxes upstream. The switch was fine. A multimeter would have told me in thirty seconds—once I knew what to look for.
Leviton publishes free wiring diagrams for every switch model on their site—single-pole, three-way, four-way, dimmers, the works. They're more accurate than any forum post and they take four minutes to read. That's cheaper than the drive to the hardware store.
"But time is money. Just buy the part."
I hear this a lot, and I get it. Budgets are real and schedules are tight. To be fair, there's a version of this that's correct: if you've already tested and you know the part is bad, buy it. Don't overthink it. Don't run a second test to confirm the first one.
But most people—including me for four years—skip the test entirely. Here's the math I ran after the third unnecessary parts order in 2022:
- Average part cost: $18
- Round trip to the store or shipping delay: 40 minutes
- Multimeter test: 5 minutes
- Certainty gained: close to 100%
Even if you value your time at $80 an hour, testing beats guessing. Unless you're certain the part is bad—and if you haven't tested, you're not—you're gambling $18 and 40 minutes on a hunch. That's not efficiency. That's a coin flip with worse odds.
The other hidden cost: parts pile up. I've got a shelf in my garage with maybe 35 unopened returns. 40? I'd have to check the log. Somewhere around $600 in stuff I bought "just in case" and never used, plus the gas and the hours.
What I do now
The rule is short:
- Test with a multimeter before ordering anything.
- Read the actual wiring diagram if it's electrical. Not a forum post. The diagram.
- Check spark, fuel, and compression—in that order—before replacing small-engine parts.
- Only buy after the test confirms the part is the problem.
If I'm under time pressure—say 20 minutes before the store closes—I still test first. The test takes five minutes. The drive takes fifteen. Buying without testing takes both, and sometimes twice if the first part was wrong.
I'm not saying never replace parts. I'm saying the order matters. Test, then buy. The $3,400 I've lost wasn't on expensive components. It was on cheap parts I didn't need, bought in a hurry, because testing felt slower than swapping.
It isn't. Test first. That's the habit that paid for itself faster than anything else I've done in six years of facility and equipment work.
