The true cost of downtime: how to calculate it and how to cut it

The true cost of downtime is lost sales, plus paid staff who can’t work, plus the work of recovering, plus the customers who quietly don’t come back. Industry averages won’t tell you what an outage costs your business. A fifteen-minute calculation will, and once you have that number you can decide which prevention spending pays for itself and which is just insurance theater.

This post gives you the formula, a worked example with every assumption shown, the costs people usually leave out, the common causes of business downtime, and the handful of fixes that actually reduce it.

How does downtime impact my business?

Through four channels, and only the first one is obvious:

• Revenue that doesn’t happen. Orders, bookings, ad impressions or billable usage that would have flowed through the system while it was down.
• Payroll that buys nothing. Support, sales, fulfillment and finance people who are paid for the hour but can’t do their jobs.
• Recovery work. Overtime, contractors, rush-shipped parts, and the catch-up work after the system is back.
• The long tail. Customers who leave, service credits you owe your own clients, and deals that slip because a prospect saw you go down.

For an online store, the first channel is the loudest. For a subscription business it’s often the smallest, because a monthly subscription still bills during an outage. The bill shows up later, as credits and cancellations.

How to calculate your cost of downtime

Use this for a single outage of length D hours:

Cost of outage = (R × L × D) + (E × W × P × D) + Recovery + Credits + Churn

R = revenue per hour that runs through the affected system, at the time of day it failed
L = share of that revenue that's actually lost, as opposed to delayed
E = employees who can't work normally
W = loaded hourly cost per employee (wages plus benefits and overhead)
P = share of their productivity lost (0 to 1)
Recovery = overtime, contractors, rush parts and catch-up work
Credits  = service credits or penalties you owe your own customers
Churn    = future gross margin from customers who leave because of it

Two variables do most of the damage to estimates. R should reflect when outages actually happen: a failure at 2 p.m. on a weekday hurts more than one at 4 a.m. on Sunday. And L is rarely 100%, because some customers simply come back an hour later. Be honest about both and the number gets smaller, but much more believable.

A worked example

Here’s an online retailer with $5 million a year in web sales and a four-hour outage on a weekday afternoon. Every number below is an assumption. Swap in your own.

Notice what dominates. It isn’t the lost sales. It’s the churn line, which is also the shakiest guess on the sheet. If you track repeat-purchase rates, compare customers who hit an outage against those who didn’t and you’ll get a real number instead of my 40.

Now annualize it. If this business has three outages like that a year, it’s losing roughly $41,600 a year to downtime. That’s your ceiling for prevention spending. Anything that costs less and reliably removes a chunk of those hours is worth buying; anything that costs more needs a better argument than fear.

The same math answers narrower questions, like whether a second internet line pays off. Estimate how many hours a year the single line takes you down, multiply by your hourly cost, and compare that with a year of the second circuit.

The hidden costs most estimates miss

Catch-up work. Orders to re-enter, tickets to answer, reports to rerun. It often costs as much staff time as the outage itself.

Data you lose for good. If your last good backup is from midnight and the failure hits at 3 p.m., fifteen hours of data either gets rebuilt by hand or doesn’t come back. How much data you can afford to lose, your recovery point objective, belongs in the calculation.

Credits you owe. B2B companies often promise uptime to their own customers. Those credits come out of your margin, not your provider’s.

The gap between your loss and your provider’s credit. Hosting SLA credits are typically calculated against the monthly fee for the affected service, not against what the outage cost you. Read the SLAs you depend on with that in mind. Ours is on our SLA page.

Leadership time. A bad outage eats a week of attention from the people who should be running the business. Nobody budgets for it.

Slow is a kind of down. A site that takes eight seconds to load at peak loses buyers too, just less visibly. If that’s your problem, our engineers will look at it through slow website support.

What the surveys say, and why your number is probably smaller

The big surveys mostly ask large organizations, so treat their numbers as a ceiling, not a benchmark. In Uptime Institute’s 2025 annual survey, 57% of respondents said their most recent major outage cost more than $100,000, and one in five said it cost more than $1 million. ITIC’s 2024 Hourly Cost of Downtime survey of more than 1,000 firms found that 90% of mid-size and large enterprises put an hour of downtime above $300,000.

If you’re a 30-person company, you won’t be anywhere near those figures, and that’s fine. Use the formula.

Common causes of business downtime

At the data center level, Uptime Institute’s 2026 outage analysis says power is still the leading cause of impactful outages, mostly failures involving UPS systems, transfer switches and generators. Outages tied to fiber and connectivity problems are rising, and when human error is involved, the leading driver is staff not following established procedures. It also notes that third-party providers (cloud, telecom and colocation companies) account for about two-thirds of publicly reported outages, which is a good reason to ask hard questions of yours.

At the level of one company’s servers, the usual suspects look like this:

• A bad change. A deploy, a config edit, a firewall rule or an OS update pushed without a rollback plan.
• A single point of failure. One server, one disk without RAID, one database, one internet line.
• Hardware failure. Drives and power supplies wear out; what matters is how long the replacement takes.
• Capacity. A disk fills up, a log grows without rotation, a traffic spike exhausts memory.
• Things that expire. TLS certificates, domain registrations and license keys.
• Attacks. DDoS floods and ransomware, which goes after your backups first.

Downtime cost prevention: what actually moves the needle

There are dozens of things you could buy. Four of them do most of the work.

1. Redundancy where your math says it pays

If one server going down costs more than a second server, run two. That might be a replica database, a second web node behind a load balancer, or a full server cluster with failover. Start with whatever your worked example says is most expensive to lose. A second dedicated server with us starts at $126 a month in Chicago or Los Angeles, which is about $1,500 a year. Compare that with your annual downtime number.

2. Monitoring that wakes a human

Ping checks tell you the server is up. They don’t tell you the disk is 97% full, a RAID array is degraded or the database service died. Watch those, check your site from outside your network, and route alerts to someone who’s on call. Our Advanced Monitoring covers load, RAID, swap, services and website uptime; it’s included with a management plan and a $49 a month add-on without one. Tickets and phone calls are answered by people 24/7.

3. Backups you’ve restored, and a recovery plan you’ve tested

A backup you’ve never restored is a hope. Keep at least one copy offsite, out of reach of the credentials an attacker would steal first, and run a real restore on a schedule. Time it. That’s your actual recovery time, not the one in the plan.

Every dedicated server order with us includes 1 TB of R1Soft backup space at no charge. For more than that, or for your own VMware and Hyper-V hosts, there’s Veeam Cloud Connect backup, and if restoring from backup would take too long, Veeam DRaaS keeps a replica ready to fail over.

4. Change control

This one is free, and it’s the one most teams skip. Write down what you’re changing, have a second person read it, know how you’ll roll it back, and don’t push it at 5 p.m. on a Friday. Keep configs in version control so “what changed?” has an answer. Uptime’s finding that not following established procedures is the leading driver of human-error outages is the polite version of this advice.

What doesn’t move the needle much

A higher SLA percentage on a single server. Bigger hardware when the problem is a single point of failure. A monitoring dashboard nobody watches. And on the host side, ask how fast a dead part gets swapped: our managed servers page states that every GigeNET server comes with 24/7 support access and a 2-hour hardware replacement guarantee as standard. That’s worth knowing, but it’s still two hours of outage on a single box, which is why redundancy sits at the top of this list.

Not ready to order yet?

Tell us what you need protected and how much data, and we’ll email you the options and the price. Rather talk it through? Call (800) 561-2656.

FAQ

How do I calculate the cost of downtime?

Add up lost revenue (hourly revenue × the share actually lost × hours down), idle payroll (affected staff × loaded hourly cost × productivity lost × hours), recovery costs, credits you owe customers and the future margin of customers who leave. Then multiply by how many outages you have in a year to get an annual figure you can weigh against prevention costs.

What’s the average cost of downtime?

Surveys of large organizations put it high: ITIC’s 2024 survey found 90% of mid-size and large enterprises put an hour above $300,000. Those averages say little about a small or mid-size business, which is why a calculation with your own numbers is more useful than any benchmark.

What’s the fastest way to reduce downtime cost?

Fix the cheapest causes first: add monitoring with real alerting, put changes through a simple review and rollback process, and test a restore from backup. Then use your annual downtime number to decide whether redundancy, such as a second server or a second internet line, pays for itself.

How do I compare the cost of a DDoS outage with the cost of mitigation?

Use the same formula: hours a DDoS attack would keep you offline, times your hourly cost, times how often you expect to be hit. Every GigeNET server includes always-on baseline DDoS protection, and larger dedicated mitigation is custom quoted, so the comparison is usually between that quote and your annual exposure.

Want help putting numbers on your own setup? Tell us what you run and what an hour down costs you, and we’ll suggest where redundancy, monitoring or backups would earn their keep.