Your contract says 99.9% uptime. That sounds bulletproof — until you read what it actually guarantees, and what it pays you when the line goes down. Here's the math the carriers hope you skip.
By Jonathan Eubanks · June 16, 2026 · 8 min read
⚡ The short version
Pull up your telecom contract and find the number. Most of them say it somewhere: 99.9% uptime guaranteed. Three nines. It reads like a fortress. It's printed in the part of the proposal the salesperson points to when you ask, "But what happens if it goes down?"
Here's what almost nobody does before they sign: the arithmetic. Because 99.9% isn't a promise that you'll stay up. It's a number that quietly tells you exactly how much downtime you've already agreed to live with — and, in the fine print, how little the carrier owes you when it happens.
I've been reading these contracts since 2003. The uptime guarantee is one of the most misunderstood lines in the whole document, because it's doing two jobs at once. It looks like a safety net, and most buyers treat it as one. But what it really is, is a cap on the carrier's liability dressed up as a benefit to you. Let me walk through what the number means, what the "guarantee" actually pays, and the one figure that matters a great deal more — and that your contract almost never mentions.
Start with the math, because it's the part the percentage is designed to hide. There are 8,760 hours in a year. Ninety-nine point nine percent uptime means one-tenth of one percent is allowed to be downtime — and a tenth of a percent of 8,760 hours is about 8 hours and 45 minutes a year. So a "three nines" guarantee is, read plainly, the carrier promising you'll be down no more than roughly a full business day each year. And they're allowed to hit that number and still be fully in compliance with your contract.
Run it down the scale and the picture sharpens:
99.99% (four nines) — about 52 minutes of downtime a year. 99.9% (three nines) — about 8 hours 45 minutes. 99.5% — about 43 hours, nearly two full days. 99% — about 87 hours a year, more than three and a half days. That single decimal place between 99% and 99.9% is the difference between a long weekend offline and a long lunch.
Two things fall out of this. First, the gap between the numbers that sound almost identical is enormous, and it's where carriers compete on price by quietly dialing the guarantee down a notch you didn't notice. Second — and this is the one that stings — a lot of standard business broadband circuits don't carry a meaningful uptime SLA at all. The "99.9%" you think you have may live only on a dedicated fiber or managed circuit, while the cable connection running your second location is sold "best effort," with no guarantee and no credit, full stop. If you've got multiple sites on multiple products, odds are good they're not all protected the same way, and nobody's ever lined them up side by side.
Now to the part that does the real damage. Say your circuit blows past its allowance and goes down for four hours on a Tuesday. You invoke the guarantee. What do you actually get?
An SLA credit. Not a refund of what the outage cost you — a credit against a future bill, calculated as a small slice of your monthly charge for that one circuit. The typical structure prorates the outage against the month and tacks on a modest multiplier, then caps the whole thing. Here's the arithmetic on a $600-a-month circuit: four hours is a sliver of the roughly 730 hours in a month, so even a generous "credit the downtime and double it" formula lands you somewhere around five to ten dollars. On a bad outage. That you have to notice, document, and formally request — usually within 30 days, in writing — or you get nothing.
Read that back. The four-hour outage might have cost you a day's orders, a morning of idle payroll, and a customer who couldn't reach you and called someone else. The "guarantee" answers all of that with a credit that won't cover lunch. That's not an accident or a stingy vendor. That's the entire design. The SLA exists to limit what the carrier owes you, not to make you whole. The number on the proposal points one direction; the liability cap in the fine print points the other.
So if the credit is meaningless, the real question is the one the contract never asks: what does an hour of your business being down actually cost? Because that's the number you're really exposed to, and it's the one you should be managing.
The research here is sobering. Recent industry surveys put the cost of downtime for small and mid-sized businesses commonly in the range of $8,000 to $25,000 per hour, and for many firms higher — not because of one giant line item, but because of how many smaller ones stack up at once. You don't need their average, though. You need yours, and you can rough it out on the back of an envelope: take the number of people who can't do their jobs when the connection drops, multiply by their loaded hourly cost, and add the revenue that simply doesn't happen during those hours — the orders not placed, the calls not answered, the point-of-sale terminal that won't run a card. Then add the part you can't put a clean number on but everyone feels: the customer who couldn't reach you today and isn't sure they'll try again tomorrow.
For a multi-location business the math gets worse, because an outage rarely politely confines itself to one site, and the lost hours multiply across every location that depends on the connection. Run that calculation honestly one time and the $5 SLA credit stops looking like a safety net and starts looking like what it is — a rounding error against your actual risk.
Quick win: Before you do anything else, pick your busiest location and estimate one number — what a single hour fully offline costs it. People who can't work, times their hourly cost, plus the revenue that hour normally produces. Write it on a sticky note. That one figure tells you instantly whether the uptime guarantee in your contract is real protection or theater — and it reframes every renewal conversation you'll have, because now you know what you're actually buying insurance against.
If you'd rather not guess at it, lining up every circuit, every SLA, and every site's real exposure on one page is the first thing we do for a new client. Because the carriers pay us, there's no advisory fee for that look. Ten minutes on the phone usually tells you where you're exposed. Talk to the team if you want a second set of eyes on your contracts.
Here's the figure your contract almost never guarantees and that matters more than the uptime percentage: how fast a human starts fixing the problem, and who that human is.
Uptime is a statistic measured over a year. Response time is what you live through on the actual Tuesday the line is down. And the two have almost nothing to do with each other. A circuit can carry a beautiful 99.9% guarantee and still leave you stranded in a phone tree for forty minutes, bounced between a carrier's tier-one queue and a "we'll open a ticket and someone will call you back" that stretches into the afternoon. The guarantee says nothing about that, because the big carriers aren't built to call you back — they're built to scale, and a single business account with a single dead circuit doesn't scale neatly into their support model. You become a ticket number, and the clock runs while you wait.
This is the whole reason failover and redundancy exist — a second, independent path so a single outage is a non-event instead of a crisis. But redundancy only helps if someone designed it on purpose and someone picks up when it's tested for real. That's the gap a smaller, independent partner actually fills: not a better uptime percentage on paper, but a person who knows your sites, answers the phone, and owns the problem until it's closed. The guarantee is a number in a document. Response is a relationship. When the line is down at 4:45 on a Friday, you find out very quickly which one you actually bought.
A 99.9% uptime guarantee isn't a lie, exactly. It's just not the thing you think it is. It's a precise statement of how much downtime you've pre-agreed to tolerate and a strict ceiling on what the carrier will pay when they hit it — and that payment is calculated against their bill to you, never against the cost to your business. Once you see the line that way, it stops being reassuring and starts being a prompt: what's my real exposure, and who's actually going to fix it fast?
You don't control whether a circuit fails someday; everything fails eventually. What you control is whether you walked into the contract knowing the math, whether you have a redundant path for the sites that can't afford to go dark, and whether the partner on the other end of the line treats your outage like their problem or your ticket. Those are decisions you make calmly at renewal, not frantically at 4:45 on a Friday.
So here's the question I'll leave you with: if your main location went down for four hours this afternoon, do you know what it would cost you — and do you know who would actually pick up the phone? If you can't answer both, the percentage in your contract isn't protecting you. It's just making you feel protected. And those are very different things.
— Jonathan
Jonathan founded Buckeye Telecom in 2003 after years in the Columbus telecom industry — first at 5-Star distributors learning the carrier side, then carrying his own quota in telecom sales. He still works directly with clients — backed by the Buckeye team.
Talk to the Buckeye team — the owner is involved in every engagement, and there’s no advisory fee.
Prefer to talk now? Call or text 614-224-2003.