-
Why I stopped answering "what UPS should we buy?" with a product name
-
Scenario 1: You're protecting equipment, not keeping the lights on
-
Scenario 2: The power goes out for hours, not seconds
-
Scenario 3: You just need to survive the blink
-
Scenario 4 (the one people don't want to hear): You don't need backup power at all
-
How to figure out which scenario you're in
Why I stopped answering "what UPS should we buy?" with a product name
Nobody trains you for this in office administration. One week I'm reordering toner, the next I'm comparing a 10 kW hybrid inverter against a used diesel generator because the building manager mentioned the power "might get flaky during storm season." Some weeks it's a 2U rack-mount UPS for the server closet. Other weeks it's a $6 spark plug for the Echo weed eater the grounds crew can't start.
So when a department head asks me what backup power to buy, I don't name a product. I ask three questions first: What are you protecting? How long do your outages actually last? And who's going to maintain this thing after I sign the invoice?
Because here's what I've learned the hard way — there's no universal right answer. There are about four situations, and they call for completely different solutions.
Scenario 1: You're protecting equipment, not keeping the lights on
If your concern is a server, a POS system, a security panel, or a couple of workstations that can't take a hard shutdown, you're not solving a power outage problem. You're solving a clean power and graceful shutdown problem. That's a UPS job.
For anything rack-mounted, I spec an APC rack mount UPS. The APC UPS line, specifically the 3000 VA Smart-UPS models, is what I keep coming back to — roughly 2,700 W of real capacity, 2U of rack space, pure sine wave output, and the option to add a network management card so I get an alert before the battery dies instead of after. Those apc smart ups 3000 specs matter more than the number on the box: check the watt rating, not just the VA, look at the runtime curve at your actual load, and find out whether the network card is included or a separate line item.
Here's the part that trips people up. The unit is maybe 60% of the total cost. My first quote for a rack-mount unit came in at $2,100 and looked great — until I read the fine print. Rail kit was extra ($180). Startup and configuration was extra ($450). Battery recycling on the old unit was extra ($75). None of it was hidden, exactly. It just wasn't in the number I was comparing.
The most frustrating part isn't the cost. It's that the quote looked complete. That's why I now ask "what's not included" before I ask "what's the price." A vendor who lists every fee upfront — even when the total looks higher — usually costs less in the end than one who gives you a clean number and invoices the rest later.
One more thing to budget for: UPS batteries have a service life. Manufacturers generally recommend replacing them every 3–5 years whether or not the unit has failed. If a vendor quotes you a UPS without mentioning the replacement schedule, ask.
Rough range for planning: $1,800–$3,000 for a 3000 VA rack-mount unit before installation, based on quotes I collected in late 2024; verify current pricing, since battery costs move.
Scenario 2: The power goes out for hours, not seconds
A UPS buys you minutes. If your outages last four hours because a transformer blew, a UPS is not the answer — it's a very expensive countdown timer.
This is where people start asking about the difference between a generator and an inverter, and honestly, most explanations make it more confusing than it needs to be. A generator burns fuel to make AC power. An inverter takes DC power — from batteries, solar panels, or both — and converts it to clean AC. A hybrid inverter can pull from the grid, charge batteries, and switch over automatically when the grid drops.
My honest take, and some people will disagree: for a small office with a handful of critical circuits, a 10 kW hybrid inverter paired with a decent battery bank is often the better long-term buy than a generator. No fuel storage. No monthly exercise runs. No noise complaints from the office next door. And it switches over fast enough that the servers don't even notice.
But it's not cheaper upfront. Not even close. A generator wins on dollars per watt every time. The trade-off is that you're buying a machine that needs fuel, maintenance, and a place to vent exhaust — plus someone who remembers to test it.
I have mixed feelings about this one. Part of me wants the simple, cheap generator. The other part remembers the year ours sat unused for eight months and then wouldn't start when we needed it.
The "generators are always the practical choice" thinking comes from an era when battery storage was prohibitively expensive. That's changed, at least for small loads. If you go the inverter route, size it by the load you actually need to carry, not by the whole building. Critical circuits only. Lighting, network gear, one HVAC zone. Ten kilowatts carries more than most people think — but not everything.
Scenario 3: You just need to survive the blink
Some offices don't have a real outage problem. They have a "the lights flicker for eight seconds during a storm and the desktop reboots" problem. If that's you, you don't need a rack-mount anything. You need a small UPS at each critical desk and one on the network switch, and you're done.
The trap here is overbuying. I've seen a 40-person office buy a 3000 VA unit to protect a single router. That's like renting a moving truck to carry a suitcase. Match the unit to the load — a 600–1500 VA unit covers most single-workstation setups and costs a fraction of the big rack units.
Scenario 4 (the one people don't want to hear): You don't need backup power at all
This is the scenario that saved us the most money, and it's the least popular answer.
Half the "we need backup power" requests I get turn out to be something else. A server that reboots randomly isn't always a power problem — sometimes it's a dying power supply or a loose cable. A machine that won't start isn't always an electrical issue — sometimes it's a $6 spark plug, like the Echo weed eater that had the grounds crew convinced the building's outdoor outlet was dead.
Before you spend four figures on backup power, spend an hour figuring out what's actually failing. Does the problem happen only during known outages, or randomly? Does it affect one device or the whole room? Has anything changed — new equipment, new circuits, construction nearby?
Sometimes the real fix is surge protection on a specific circuit, or replacing a ten-year-old power supply, or moving a device off a circuit that shares a breaker with the copier. Backup power doesn't fix bad wiring. It just keeps the bad wiring powered.
How to figure out which scenario you're in
Answer these in order. The first one that applies usually decides it for you.
- Do you have equipment that can't take a hard shutdown — servers, POS, medical or security gear? If yes, you need a UPS regardless of anything else. Start there.
- Do your outages routinely last more than 15–20 minutes? If yes, a UPS alone won't cut it. Look at a generator or a hybrid inverter setup.
- Is your only real symptom a flicker or a brief reboot? If yes, small UPS units at the critical points, not a building-wide system.
- Does the problem happen when the power is on and stable? If yes, it's probably not a backup power problem at all. Diagnose first.
The last thing I'll say: whatever you buy, ask for the full cost picture in writing before you sign. Unit, install, rails, network card, battery replacement schedule, disposal. The vendors who give you that without being pushed are the ones I keep. The ones who quote a clean number and get vague about the rest — those are the quotes that come back to bite you six months later, usually in front of your boss.
