I've spent the last six years handling energy storage orders for commercial and medical clients. I've personally made (and documented) four significant mistakes, totaling roughly $47,000 in wasted budget. Now I maintain our team's checklist to prevent others from repeating my errors.

This checklist is for anyone who has to evaluate battery capacity for a commercial energy storage system: EV charging, backup power for critical equipment, or shifting a building's load to off-peak hours. It is not a marketing comparison. It's the order of operations I use after learning the hard way.

There are five checks. If you're comparing manufacturers like BYD, Tesla, or other suppliers, run the numbers in this order.

1. Define the load before you look at a spec sheet

Every battery decision starts with the loads that must run, not with the battery that looks good on paper. In my first year (2019), I sized a backup system for a clinic and forgot to include the patient monitoring system. If you're asking what is patient monitoring system, the practical answer for this article is: it's the equipment that tracks vital signs and alerts the care team when something goes wrong. It doesn't draw a huge load, but it cannot go dark, even for a beat.

The mistake I made that year was skipping the load table and going straight to datasheets. I now build a simple table with two columns: continuous watts and starting surge. Add up everything that has to run for the full outage duration, then separate the loads that can wait. A lot of people miss the surge column until the inverter trips on day one.

2. Compare usable capacity, not the nameplate number

When someone quotes the BYD M6 battery capacity, treat that as the gross pack energy. The same goes for the BYD Atto 2 battery capacity—the headline number on the brochure is not the energy you can pull from the pack every day. The battery management system locks out a portion at the top and bottom of the state of charge to protect cell life and performance.

(Should mention: I don't have hard data on every vendor's locked-out margin. But based on the projects I've reviewed, my sense is the gap between nominal and usable capacity is often 8–20%, depending on chemistry, voltage range, and the manufacturer's BMS policy.)

When I read the Tesla Powerwall 3 capacity kWh and Powerwall 3 specs, I look for the usable figure and the conditions attached to it. Published capacity almost always assumes a specific temperature, discharge rate, and cycle window. That's not unique to Tesla; it's the same physics you'll see in BYD or any other lithium-ion system.

The fix: convert every candidate into usable kWh at your actual site conditions. If one vendor's capacity is 60 kWh with 80% usable depth and another is 55 kWh with 95% usable depth, the 55 kWh pack may actually deliver more daily energy. That kind of comparison is where total cost thinking starts.

3. Make maintenance access non-negotiable

A battery bank is not a one-time purchase. Charger firmware needs updating. Terminal torque needs checking. Air filters and cooling fans collect dust. If you can't safely isolate the battery, every one of those tasks becomes a project.

In September 2022, we installed a battery bank with no practical way to disconnect the charger for service. The specifications looked fine: good capacity, decent price, compatible inverter. But the cables were bolted directly to the terminals. When the charger needed service, an electrician had to work near live lugs, which is slow, hazardous, and embarrassing to explain to a client.

The battery tender quick disconnect cable mount sounds like a minor accessory, but it's actually a service entry point. It creates a clean, fixed location to disconnect the battery from the charger, which turns a four-hour maintenance job into a ten-minute swap. We've caught 47 potential errors using this checklist in the past 18 months, and this one is the most common miss.

4. Calculate the total cost of ownership, not the price per kWh

I used to compare quotes by dollars per kilowatt-hour. That was mistake number one. The lowest price per kWh can end up being the highest-cost system once you add installation, wiring, mounting, disconnects, permits, commissioning, training, maintenance, and the cost of downtime.

Here's a typical example: a quote that looks $2,000 cheaper than an integrated package. After shipping, a second site visit, additional mounting hardware, and a service disconnect, it came out $600 more expensive. The unit price was lower; the total cost was not.

I now calculate TCO before comparing any vendor quotes. That includes the time cost on my side too. If a vendor's documentation is incomplete, that's a line item. If their commissioning support takes three weeks, that's a line item. The steeper upfront quote can look embarrassing in a budget meeting, but the cheapest quote can look far worse after a failure.

If the system is protecting a patient monitoring system, the cost of downtime includes clinical and liability risk, not just lost revenue. Run that math before you sign anything.

5. Verify with documentation and a no-load test

After the third rejection in Q1 2024, I created our pre-check list. Now nothing leaves the warehouse without:

  • A capacity test report matched to the actual discharge rate
  • BMS settings that match the system design, not the default from a different project
  • Temperature derating applied for the site's climate
  • An emergency shut-off that is reachable, labeled, and tested

Then run a no-load test. Simulate a grid failure and watch what the monitoring software reports. I've seen systems that looked perfect on paper drop out in the first five minutes because of a communication cable issue or a ground bond that was never checked.

Ask the installer for the key safety anchors as well: in North America, that usually means UL 9540 for the energy storage system, UL 1973 for the battery, and a design that follows NEC Article 705 and 706. In other markets, ask for IEC 62619 or the local equivalent. A stack of certifications doesn't guarantee performance, but it filters out the suppliers who are building on guesswork.

Even after choosing the final vendor, I usually keep second-guessing. What if the cheaper option would have worked? I don't relax until the no-load test passes and the BMS log shows steady voltages.

What still surprises me

The most frustrating part of battery procurement is that the same spec-sheet mistakes keep appearing, from different people, on different projects. You'd think usable capacity would be the first thing people check, but the marketing number still wins a lot of decisions.

My experience is based on about 40 commercial and medical projects, mostly in North America. If you're working on a smaller residential install, your experience may differ. But the core logic—load, usable capacity, service access, total cost, verification—still applies.

I'm not going to tell you that BYD is better than Tesla or the other way around. I've specified both. For the right load, each is a solid option. What matters is whether the whole system, from battery capacity to the cable mount, works as one integrated product. That's the checklist I use now, and it's the one I wish I'd had in 2019.