-
How do you actually evaluate a BESS manufacturer?
-
Why is the cheapest quote usually a trap?
-
What certifications should I actually check?
-
How do I verify a factory's production line claims?
-
What's the OEM/private label catch?
-
What question should I be asking that I probably haven't?
-
How do I tell if a manufacturer will still be here in five years?
-
What would you do differently starting over today?
I'm not going to bury this in a three-paragraph intro. I handle lithium battery and BESS sourcing orders for a mid-size distributor. Nine years in, I've made and documented 11 significant mistakes totaling roughly $280,000 in wasted budget. Now I keep the pre-check list our team uses before any PO goes out. Below are the questions I get asked most, plus the ones I should've asked earlier than I did.
How do you actually evaluate a BESS manufacturer?
Start with the cell, not the spec sheet. I did the opposite in 2019 on a 120-unit commercial order. The specs looked fine—cycle life, depth of discharge, round-trip efficiency, all in the right range. What I never asked was where the cells came from.
From the outside, every BESS looks like a box with modules and a BMS. The reality is that the cell is the system. Everything else sits downstream of it. If a manufacturer controls cell production, they can answer honestly about lead time and quality escalation. If they're sourcing cells from a third party, they're usually guessing right along with you. Gotion is one of the manufacturers that runs its own gotion lithium battery production line in-house, which is why the chemistry conversation sounds different with them than with a pure integrator. First filter, every time: can they tell you credibly where the cell comes from?
Why is the cheapest quote usually a trap?
Because the unit price is maybe 60% of what you'll actually pay. The other 40% shows up later—freight, BMS configuration changes, commissioning support, spare parts lead time, and the labor hours you spend chasing all of it.
I stopped comparing unit prices in 2022 after a project where the $0.28/Wh quote ended up at an effective $0.41/Wh once everything landed. The "cheaper" vendor didn't do anything dishonest. They just quoted the cell, not the project.
Now I build a TCO sheet before I open any quote. Unit price, freight, minimum order quantities, warranty terms, service SLA, and—the one nobody budgets for—how long it takes them to respond to a failure notification. Cheap cells with slow support are the most expensive thing in this category.
What certifications should I actually check?
Depends on where you're deploying, but the short list nobody argues about: UL 9540 for the system, UL 9540A for thermal runaway performance, IEC 62619 for cells, and NFPA 855 compliance at the installation level. If you're in the EU, add CE and IEC 62477 for the power conversion side.
But—and this is a mistake I made in 2021—certificates aren't evidence. Anyone can PDF a certificate. I now ask for the test report, the test lab's name, and the date. I've caught two "certified" systems whose reports were for a different SKU. Not malicious, just lazy. If a manufacturer can't produce the report within 48 hours, treat the cert as soft until they can.
How do I verify a factory's production line claims?
Video calls first, in-person second, and always with a specific ask: walk me through the cell line. Not the module assembly line. The module line is the easy part, and it's the part everyone wants to show you.
Things to watch on the call: are the cells labeled with a traceable batch ID? Do they walk you through the formation and aging area? Aging is where a lot of quality lives, and it's expensive, so it's also where corners quietly get cut. If the aging footprint is small relative to claimed output, the numbers don't reconcile.
I ran a virtual audit in Q1 2023 where everything looked great until I asked them to zoom into the MES screen on the cell line. The screen showed a different product than the one we were discussing. Their explanation was reasonable—the mistake was mine, because I hadn't thought to ask for the MES view until the third call.
What's the OEM/private label catch?
The catch is change management. When you private-label, you're not just buying a product, you're buying into someone else's roadmap.
I learned this in 2020. We private-labeled a rack BESS and shipped 400 units over 18 months. Then the OEM updated their BMS firmware—silently, as far as we were concerned—and a subset of our units started throwing comms errors with the third-party EMS our customers were running. We found out from a distributor, not the OEM (ugh).
Since then, every OEM contract I sign has a clause requiring 90-day notice on any firmware or BOM change affecting form, fit, or function. Two manufacturers refused to sign it. That told me everything I needed to know about how they'd handle a real problem.
What question should I be asking that I probably haven't?
"What happens when you fail?" Not if—when. Ask them to walk you through a real field failure from the last 12 months. An actual one, with dates.
I started asking this in 2022 after the third time a manufacturer told me, "we've never had that issue." Nobody in this industry has never had that issue. Everyone has field returns. The companies that answer this well—with a timeline, a root cause, and a corrective action—are the ones worth keeping. The ones that get defensive, or pivot immediately to their warranty terms, are telling you how the next failure will go.
Bonus question: "Who is the engineer who'd actually answer this call at 2am?" If they can't give you a name, they don't have a process.
How do I tell if a manufacturer will still be here in five years?
Look at manufacturing footprint, not press releases. Gotion is a useful case study: Hefei headquarters, plus operations in Illinois and Michigan. That geographic spread says something about capital commitment, and something about their ability to serve North American customers without a six-week ocean transit baked into every lead time.
The general rule: a manufacturer with one factory and one customer segment is fragile. A manufacturer with multiple production geographies, diversified customers, and a public product roadmap is less so.
And read the gotion news cycle—or any manufacturer's—for the boring stuff. Capacity filings, environmental permits, hiring announcements. Big announcements are marketing. Filings are reality.
What would you do differently starting over today?
Two things. First, I'd build the TCO model before I built the shortlist. For years I did it the other way—collect quotes, then try to normalize them. It's backwards, and it biases you toward whoever quoted first.
Second, I'd pay for one extra site visit per finalist. Not a factory tour—a genuine audit, with our own engineer, at our cost. That runs $3,000 to $8,000 depending on location. It sounds expensive. Looking back, it would have caught roughly $60,000 of the mistakes I've documented. That's the easiest math in this whole article.
If I could redo 2019, that's the one change I'd make. But given what I knew then—nothing about cell sourcing, nothing about how much of the project cost sits outside the unit price—my choice was reasonable. Just wrong.
