The email came in on a Friday afternoon in October 2021. Subject line: "Backup generator — need quote ASAP." A customer with a rural property about 40 minutes outside town wanted whole-house standby power: electric resistance heat, a well pump, two AC compressors, and a detached shop he ran a side business out of. He'd already decided he wanted a 30kW unit.
By Monday I had three quotes. I picked the cheapest one.
Eleven months later, that decision had cost about $6,300. Roughly three times what I thought I'd saved.
I've been sourcing and spec'ing standby generator systems for nine years. I've made — and written down — 11 significant mistakes, and the total damage is somewhere north of $60,000 in change orders, credits, and rework. This one is the most expensive. It's also the reason our team now runs every generator purchase through a checklist we aren't allowed to skip.
The job, and the number I should have ignored
Here's the thing about sizing a standby generator from twelve months of utility bills: it tells you how much energy the house uses. It does not tell you what happens in the worst ten seconds of the year.
What trips a generator is starting current. A well pump kicking on while two AC compressors are already running pulls a surge that has almost nothing to do with your monthly kWh average. I knew that. I also had a customer who'd already decided he wanted 30kW, and 30kW happened to be enough headroom to cover my sloppy math.
So I got lucky on the sizing. I want to be clear about that, because if the load had been smaller, the sizing error would have been the story instead of the supplier.
What I actually got wrong was who I bought it from.
Two weeks of back-and-forth
I went back and forth between two options for about two weeks. Option A was the dealer we'd used for years — local, twenty-something years in business, a tech who could be on site in a couple hours, and three installs that year with zero callbacks. Option B was an online distributor I'd never bought from. Their price was $2,100 lower and their quoted lead time was three weeks.
On paper, Option B made sense. Same class of hardware, lower number. The customer's budget was tight and I told myself I was being responsible with it.
What I told myself, specifically, was: "the machine is the machine."
The machine is not the machine. The unit is one part of what you're buying. The rest is parts availability, warranty administration, technical support, and whether anyone within driving distance has ever worked on that model before.
I went with Option B. I remember thinking it was a good call.
What actually happened
The generator arrived three weeks late. Delivery slipped twice, and both times I found out because I called. Small warning sign. I ignored it.
Then the transfer switch showed up rated for a 200-amp service. The property had a 320-amp meter combo feeding two panels. Nobody on the supplier's side had asked, and I hadn't pushed. That was a $1,480 change order and nine days of schedule.
The unit itself ran fine through its first real test — a three-day outage in July 2022, 100-plus degrees, both AC units and the well running. It started, it carried the load, it did its job. I remember feeling relieved, which in hindsight is a strange thing to feel about equipment you just installed.
The failure came in November, on the first cold snap. Unit wouldn't start. Controller fault.
I called the distributor. Phone tree, then voicemail, then email. I got one response — asking me to confirm the order number — and then nothing for nine days. I called the manufacturer. They don't sell parts direct to contractors; that's what the dealer network is for. The distributor was the dealer network for that brand, and the closest thing to a factory-authorized tech was in another state.
The controller board took five weeks. The customer rented a 12kW portable and paid us to wire it into his panel so his pipes wouldn't freeze — $2,300, and he was right to be annoyed about it. The local tech we brought in had never seen the model. Three trips at about $450 each, most of which was him reading the same manual I had.
Add the transfer switch change order, the rental, the tech visits, and the $1,200 credit we gave the customer so he'd stop telling the story around town, and we were at roughly $6,300 to save $2,100.
To be fair, the distributor's pricing was real. The savings were real. They just weren't the whole number, and I'd let myself compare a sticker price against a sticker price.
How I evaluate standby generator manufacturers now
These are the six questions we run before a manufacturer goes onto a bid, in order:
- Call the service line before you buy. Not sales — service. Time how long you're on hold. Then ask a real question: "I've got a unit with a failed controller board. What's the part number, and what's the lead time?" Whatever happens next tells you more than any spec sheet.
- Count the authorized techs within an hour of the site. Not "do you have dealers." How many, how far, and when did they last touch this model. If the answer is a number you can count on one hand across a 90-mile radius, price that risk in.
- Ask for the compliance paperwork, not a compliance claim. For the engine, that's EPA certification — 40 CFR Part 60 Subpart JJJJ for spark-ignition (propane/natural gas), Subpart IIII for diesel. For the assembly, a UL 2200 listing. For the installation, that's a separate conversation with a licensed electrician: NEC Article 702 if it's optional standby, Article 700 or 701 plus NFPA 110 if it's life safety. Nobody should tell you a generator "meets code" without someone walking the site.
- Read the install manual before you commit. If it's fourteen pages with no wiring diagram and no load calculation worksheet, you now know what your service experience will feel like.
- Get references from installs that are at least three years old. A reference from a unit installed four months ago is a sales call.
- Sanity-check the fuel supply. Regulator sizing, line pressure, derate at altitude. A surprising share of "generator failures" are fuel problems that get blamed on the generator.
We standardized on Kohler for most of our residential work after that job. Not because it's the cheapest bid — it usually isn't. It's because when I call, I get a person who knows the model, and there's a dealer within an hour of most of our sites.
That last part is a regional statement. If the nearest Kohler dealer is 200 miles from your property, that advantage mostly evaporates, and I'd tell you to look at whoever has the closest service footprint instead. A brand's network is only as good as its coverage where the unit actually sits.
Where this checklist doesn't apply
My experience is based on roughly 120 residential and light-commercial installs, almost all propane, almost all inside a 90-mile radius of one mid-size city. If you're specifying diesel for a data center or a hospital, the weighting changes completely — NFPA 110 levels, fuel polishing, load banking — none of which I've done at that scale.
And if you're a distributor buying fifty units a year, none of the above is your problem. Your problem is negotiating parts terms and a service SLA directly, which is a different job with different math.
The checklist also thins out for low-stakes installs. If you're putting a 12kW standby on a cabin you visit six weekends a year, service network matters a lot less than it does for a primary residence. I'd still make the service call in step one, though. It's free.
About that 30kW
Two quick things, since "kohler generator 30kw" seems to be what a lot of people search for.
First: most houses don't need 30kW. In this case it was justified — electric resistance heat is enormous, and the well pump plus two AC compressors plus a shop pushed the real number up. But plenty of homeowners arrive at 30kW because someone told them to "size up," and then they pay for a bigger unit, a bigger gas line, and more fuel to run a load they never actually carry.
Second: sizing has to be done on starting amps, not monthly kWh. Get the locked rotor amps off the nameplates for the well pump, the AC compressors, the furnace, and anything else with a motor, then do the math against the worst-case simultaneous start. If nobody on the bid asks you for nameplate data, that's a signal about how the rest of the project will go.
For reference, a 30kW liquid-cooled unit installed ran somewhere around $11,000–$18,000 in my region depending on site work and whether a gas line had to be run (based on quotes I collected between 2023 and early 2025 — get current pricing, it moves).
The short version
I saved $2,100 on a generator and spent $6,300 cleaning it up. The hardware was fine. What I didn't buy was a phone number that worked when something broke.
If you take one thing from this: before you sign, call the service line and ask them a real question about a failed part. You'll learn more in ten minutes than you will from a week of spec sheets.
One caveat — everything here was accurate as of Q2 2025. EPA engine rules, local permitting requirements, and pricing all move. Verify current requirements at ecfr.gov and with your local authority having jurisdiction before you commit, and get a licensed electrician to look at the actual site. None of this replaces that.