The flicker that started it
Last October, a two-second grid flicker took down a shared drive at our main office. I'm the office administrator for a 400-person company—I manage facilities and services purchasing across three locations, roughly $180K annually spread over about a dozen vendors. Power protection wasn't in my job description until it was.
We had a UPS on that circuit. A perfectly good unit—1500VA, rack-mounted, from a brand our IT contractor recommended. When I asked why it didn't hold the load, the answer wasn't about the unit at all. The battery was eight months past its recommended replacement window. Nobody had checked it.
That was the first time I realized our power protection strategy was really just a purchase order and a shrug.
What most of us get wrong
When I took over facilities purchasing in 2020, my UPS checklist had three items: wattage, price, and whether the vendor could deliver by Friday. (Not exactly a rigorous framework.)
Most buyers focus on per-unit cost and capacity—which is reasonable, as far as it goes—and completely miss the two things that actually determine whether a UPS works when it matters: the condition of the circuit it's connected to, and the health of the battery inside it.
Neither of those checks is complicated. But neither is in the sales conversation, either.
The question everyone asks is "what's the wattage and the price?" The question they should ask is "what happens to this unit in year four, and whose job is it to notice?"
The checks nobody schedules
Here's the part that took me three separate incident reports to understand. A UPS is really two systems: the electronics that convert and regulate power, and the battery that carries the load when the grid drops. The electronics tend to be reliable. The battery is a consumable, and it degrades quietly—no alarm, no obvious smell, no change in how the unit looks sitting in the rack.
Industry battery maintenance guidance (as of 2025) generally puts VRLA battery life in UPS applications at three to five years under normal conditions. Temperature, discharge cycles, and charging habits all move that window around. The point is that it's a known range that requires scheduled verification, not a "wait until it breaks" situation.
Two checks cover most of the risk, and both use a multimeter that costs well under $50:
- Checking continuity with a multimeter before installation. This verifies the circuit is actually grounded and the connections are sound. A UPS on a bad circuit will protect the electronics from surges, but it can still fail to pass power when the grid drops—exactly the moment you're relying on it.
- How to check for battery drain with a multimeter after installation, and quarterly thereafter. Testing across the battery terminals under load (following the manufacturer's procedure) tells you whether the battery is holding voltage when it matters—not just sitting at a healthy resting number.
The first time I ran that second test, two units that "looked fine" dropped below their rated voltage within seconds. Both were still passing their built-in self-tests. The self-tests weren't lying—they just weren't testing the thing that fails.
What skipping checks actually costs
The "buy it and forget it" approach to power protection bleeds into everything adjacent. We keep a small RV solar panel kit with battery and inverter for our remote site trailer—same battery chemistry, same degradation curve, same need for periodic testing. The equipment changes; the oversight pattern doesn't.
When that shared drive went down, we lost three days of accounting records that had to be reconstructed from email attachments and two people's memory of the reconciliation. The recovery cost—staff time, an emergency IT visit, and rework—landed around $4,200. The replacement battery cost $180.
The most frustrating part isn't the money. It's that the same problem kept showing up in different offices (like the branch location where nobody had opened the UPS front panel since it was installed in 2021). You'd think a written maintenance schedule would prevent it, but schedules only work when someone actually owns them.
This was true a decade ago when battery monitoring meant waiting for something to break. Today the tools are cheap and the procedures are documented. The "it's too complicated to check" thinking comes from an era when you needed a dedicated technician for basic diagnostics. That's changed.
The prevention check that pays for itself
Five minutes of verification beats five days of correction. That's not a slogan in our office—it's a line item.
A checklist I put together after the third incident (yes, third) now runs quarterly and takes about 90 minutes across all three offices. It covers continuity verification at the panel, battery drain testing per unit, and a visual inspection of connectors and ventilation. We've caught two batteries in the "declining but not dead" range in the past year—both replaced before they failed.
On the purchasing side, the checklist has changed how I buy. For standard office protection, an Eaton 1000VA unit is a sensible starting point—it covers a workstation cluster or a small rack without oversizing. But I no longer spec the unit without also confirming the circuit's continuity test results and putting the first battery check on the calendar before the invoice is filed. A sale UPS from Eaton is only a deal if the installation and maintenance are covered too.
That's the shift I'd recommend to any office manager who owns this responsibility: treat the UPS as a system with a maintenance lifecycle, not a box you buy once. The five-minute check that nobody scheduled is usually the one that would have caught the problem.