IT downtime costs most small and midsize businesses somewhere between 1,500 and 25,000 dollars an hour once you count lost gross profit, idle payroll, recovery labor, and the customers who never come back. The per-minute figures you see quoted online come from enterprise research and are 10 to 100 times too high for a 40-person company. Below is the formula, a worked example with the actual math, and the levers that move the number down.
IT downtime costs a typical small or midsize business 1,500 to 25,000 dollars per hour, covering lost gross profit, paid-but-idle staff, recovery labor, and customer churn. Large enterprises report far higher figures, frequently above 300,000 dollars per hour.
Owners ask us the same question after every serious outage. What did that actually cost? It is the right question, and the answer decides whether managed IT services look expensive or look like insurance.
Most owners run the rough math in their head and stop there. Payroll kept going. Orders didn’t. Somebody upstairs is going to ask what the morning cost, and the honest answer is usually a shrug.
That shrug is expensive. Without a number, downtime stays an annoyance instead of a line item, and nothing that costs an annoyance ever gets budget. Say out loud that an hour of dark systems costs your company 4,200 dollars and the conversation stops being about price. We’ve run this calculation with Texas manufacturers, distributors, law firms, and clinics more times than I can count. The result is almost never what people guess. Owners of 30-person companies usually guess too low. Anyone who read a vendor whitepaper last week guesses wildly too high.
What is the cost of IT downtime?
The cost of IT downtime is the total money a business loses while the systems it runs on are unavailable, not the invoice to fix them. It covers lost gross profit, wages paid to people who can’t work, recovery labor, contract penalties, and revenue from customers who go elsewhere and stay there.
Downtime also isn’t limited to a server on fire. It’s the accounting team locked out of the ERP for 90 minutes. It’s the phone system dropping calls all morning. It’s the shared drive that mounts but reads at dial-up speed while three people give up and go get coffee. Partial degradation counts. If the work can’t happen, the meter is running.
One clarification worth making early. Planned maintenance windows are not free either, but they’re cheap because you choose when they happen. The expensive kind is unplanned, and that’s what every number in this article refers to.
Which costs actually stack up during an outage?
Six buckets. Most owners count the first one, maybe the second, and stop there. In the worked example further down, lost gross profit accounts for only 41 percent of the total damage.
- Lost gross profit. Revenue you never book, valued at your margin rather than your top line. Using top-line revenue is the single most common way businesses overstate this number.
- Idle payroll. Wages and benefits paid to staff who are on the clock and blocked. Weight it by how dependent each team really is on the system that went down.
- Recovery labor. Internal IT hours, emergency vendor rates, after-hours callouts, and any hardware bought in a hurry at retail pricing.
- Catch-up cost. Overtime to clear the backlog, expedited shipping to hit a promised date, and rework on anything that got half-entered when the system dropped.
- Contractual exposure. Service credits, missed delivery windows, and late-filing penalties in regulated work. A law firm that misses a court deadline has a very different downtime cost than a print shop.
- Permanent customer loss. Orders that move to a competitor and never move back. Small as a percentage, large in dollars over a year.
The last bucket is the one people argue about, and fairly so, because it’s the hardest to measure. Splunk’s 2026 research with Oxford Economics found that 81 percent of technology leaders named customer loss as a consequence of downtime, and that 47 percent said customers often noticed the outage before the company did (Splunk and Oxford Economics, 2026). Being told about your own outage by a customer is not a great start to the recovery. When the outage is an attack, the clock works differently, which is why ransomware recovery in Texas gets its own playbook.
How do you calculate your own downtime cost?
Use this formula. It’s deliberately simple, because a number you can defend in a budget meeting beats a model nobody trusts.
Downtime cost = (lost gross profit per hour + idle payroll per hour) × hours down + recovery labor + catch-up cost + permanent customer loss
Four inputs do most of the work. Pull them once and you can rerun the calculation for any outage in about five minutes.
| Input | Where to get it | Common mistake |
|---|---|---|
| Revenue per operating hour | Annual revenue divided by your real operating hours, not 8,760 | Using calendar hours instead of hours you actually trade |
| Gross margin | Last year’s P&L | Applying the full revenue figure instead of margin |
| Fully loaded hourly labor cost | Payroll plus 15 to 20 percent for benefits and taxes | Using base salary only |
| System dependency percentage | Ask each department head what share of their work stops | Assuming 100 percent for everyone in the building |
That last input is where honesty pays off. When the ERP goes down, warehouse staff might be 95 percent blocked while the marketing team is barely affected. Averaging across headcount without weighting produces a number your CFO will pick apart in about nine seconds.
What does a 4-hour outage cost a 40-person company?
Here’s the math on a realistic Texas distributor. 40 employees, 12 million dollars in annual revenue, 30 percent gross margin, 2,080 operating hours a year, average fully loaded labor cost of 42 dollars an hour, and roughly 70 percent system dependency across the staff. The outage runs 4 hours on a Tuesday morning.
| Cost bucket | Math | 4-hour total |
|---|---|---|
| Lost gross profit | 12,000,000 / 2,080 = 5,769 per hour × 30% margin = 1,731 per hour | 6,923 |
| Idle payroll | 40 staff × 42 per hour × 70% dependency = 1,176 per hour | 4,704 |
| Recovery labor | 3 engineers × 8 hours × 150 per hour | 3,600 |
| Catch-up overtime | 25 hours of backlog clearing × 55 per hour | 1,375 |
| Permanent customer loss | 5% of the 23,077 in revenue missed, at 30% margin | 346 |
| Total | Roughly 4,237 per hour | 16,948 |
About 16,900 dollars for one morning. Not 15,000 dollars a minute. Not 300,000 dollars an hour. Around 4,200 an hour, which is a number a 40-person distributor can actually act on, and one that makes a 3,500 dollar monthly IT contract look like a rounding error next to two bad outages a year.
Run the same model at a 12-person insurance agency and the hourly figure lands closer to 900 dollars. Run it at a 180-person clinic with billing deadlines and it clears 30,000. That spread is exactly why a single industry-wide number is useless to you.
What do the published downtime studies actually say?
The research is real and the researchers are credible. The problem is that almost none of it studies companies your size. Here’s what each headline figure is really measuring.
| Source | Year | Headline figure | Who it actually describes |
|---|---|---|---|
| ITIC | 2024 | Over 300,000 dollars per hour for more than 90 percent of respondents; 41 percent report 1 to 5 million | 1,000+ mid-size and large enterprises worldwide |
| Enterprise Management Associates | 2024 | 14,056 dollars per minute on average, 23,750 for large enterprises | 400+ IT professionals, weighted toward enterprise |
| Uptime Institute | 2026 | 57 percent say their last major outage cost over 100,000 dollars; 1 in 5 over 1 million | Data center and infrastructure operators |
| Splunk and Oxford Economics | 2026 | 600 billion dollars a year in aggregate, about 15,000 dollars per minute | Forbes Global 2000 companies |
| Sophos | 2025 | 1.53 million dollars average recovery cost, excluding any ransom; 53 percent recovered within a week | 3,400 organizations hit by ransomware across 17 countries |
| IBM | 2025 | 10.22 million dollars average US breach cost; 241 days to identify and contain | Organizations that suffered a data breach |
Two things in that table are worth pulling out. First, Uptime Institute reports that power is still the leading cause of impactful outages, with IT and networking issues accounting for 23 percent (Uptime Institute, 2025). In Texas, that’s not an abstraction. Second, EMA found a 60 percent jump in per-minute cost for organizations under 10,000 employees while the largest firms actually saw a slight decrease. Smaller companies are getting hit harder, not less.
Why the famous per-minute numbers don’t apply to you
Time for an honest correction, because we’ve quoted these numbers loosely ourselves in the past. The most cited downtime statistic on the internet is 5,600 dollars per minute, attributed to Gartner. It comes from 2014. It described large enterprises even then, and it has been repeated so often that it now shows up in sales decks aimed at 15-person companies where it is off by a factor of roughly 100.
The newer figures have the same problem in the opposite direction. Splunk’s 15,000 dollars per minute is drawn from Forbes Global 2000 companies whose average annual downtime bill is 300 million dollars. Those are useful numbers if you run a bank. They are noise if you run a 60-person fabrication shop in Pasadena.
Here’s the practical test. If a downtime statistic doesn’t state the revenue band or employee count of the companies surveyed, it can’t tell you anything about yours. Use published research to understand direction and cause. Use your own P&L to get the dollar figure. Our breakdown of downtime costs for Houston businesses works through the same exercise with local context.
What does downtime cost by company size?
These ranges come from running the formula above across the client sizes we work with, not from a survey. Treat them as a sanity check on your own calculation rather than a substitute for it.
| Company size | Typical revenue | Realistic cost per hour | What usually dominates |
|---|---|---|---|
| Under 10 employees | Under 2 million | 250 to 1,500 dollars | Owner time and missed sales calls |
| 10 to 50 employees | 2 to 15 million | 1,500 to 8,000 dollars | Idle payroll and deferred orders |
| 50 to 200 employees | 15 to 75 million | 8,000 to 35,000 dollars | Lost gross profit and recovery labor |
| 200 to 1,000 employees | 75 to 500 million | 35,000 to 200,000 dollars | Contractual penalties and customer churn |
Regulated industries sit at the top of their band or above it. A healthcare practice that can’t access records during an outage carries HIPAA exposure on top of the lost revenue. A financial firm under GLBA or FFIEC scrutiny carries reporting obligations that don’t pause because the network is down. If that’s you, add a compliance line to the formula and be generous with it.
Why duration matters more than your hourly rate
Once you have an hourly number, the lever that moves your annual downtime bill isn’t the rate. It’s how long each incident lasts, and how many you have. Cutting a 6-hour recovery to 45 minutes saves more money than any amount of arguing about margin percentages.
That’s the entire point of setting a recovery time objective and a recovery point objective. RTO is how long you can be down before the damage becomes unacceptable. RPO is how much data you can afford to lose. Both should be set per system, because the tolerance for your email server and your production database are not the same. If those terms are new, start with our explainer on what disaster recovery actually involves.
Ransomware is where duration becomes brutal. Sophos found that 53 percent of victim organizations were back within a week, which sounds reassuring until you multiply your hourly figure by 40 working hours. For the distributor in our example, a week of degraded operations is roughly 170,000 dollars before anyone discusses a ransom. The Sophos average recovery cost of 1.53 million dollars excludes the ransom payment entirely.
This is also the argument for testing restores rather than trusting backups. A backup that has never been restored is a hypothesis. We cover the full planning process in our guide to business continuity planning.
What actually lowers the number?
Five things, roughly in order of return on effort. None of them are exotic.
- Proactive monitoring with real alerting. Most outages announce themselves first as a failing disk, a filling volume, or a switch throwing errors. Catching that at 2 a.m. costs nothing. Catching it at 9 a.m. costs a morning.
- Tested, immutable backups. Air-gapped or immutable copies survive ransomware that encrypts everything on the network, including the backup share. Schedule the restore test and document the result.
- Redundancy on the two or three systems that actually stop work. You don’t need failover for everything. You need it for the ERP, the phone system, and the internet circuit. A second circuit from a different carrier is cheap insurance in a state where weather takes out infrastructure regularly.
- Patching and identity hygiene. Sophos found exploited vulnerabilities behind 32 percent of ransomware attacks and compromised credentials behind 23 percent. Multifactor authentication on Microsoft 365 and a current patch cadence close most of that door.
- A support model that responds in minutes. Response time is a direct multiplier on every hour in your calculation. Uprite’s average response time is just over 5 minutes, and our team monitors client environments around the clock.
Worth saying plainly. Not every one of these pays for itself at every company size. Full failover for a 12-person office is usually overkill, and we tell people so. The 12-person office needs tested backups and monitoring, and it needs somebody to answer the phone. If you’re weighing what that support model should look like, our comparison of managed IT versus break-fix versus in-house lays out the cost structures side by side, and the Texas MSP Pricing Index shows what local providers actually charge.
The payoff is real when the work gets done. One Texas real estate firm we support cut more than 100,000 dollars out of its IT costs largely by ending the cycle of emergency fixes that followed every preventable outage.
Get your own number in 30 minutes
Do this before your next budget conversation. It’s a spreadsheet, not a project.
- Pull last year’s revenue and gross margin from the P&L. Divide revenue by your real operating hours to get revenue per hour, then multiply by margin.
- Get total payroll, add 18 percent for benefits and taxes, and divide by the same operating hours to get fully loaded labor cost per hour.
- Ask each department head one question. What share of your team’s work stops if the network goes down? Average the answers, weighted by headcount.
- Add the two hourly figures together. That’s your baseline hourly downtime cost before any recovery spend.
- Look up your last three outages and their durations. Multiply. Add recovery invoices and overtime you actually paid.
- Compare the annual total to what you currently spend on IT. That comparison is the whole argument.
Most companies that run this exercise find their annual downtime cost exceeds their entire IT budget. That’s not a sales line. It’s what the math tends to produce when nobody has been measuring.
Questions we get about downtime costs
How much does one hour of IT downtime cost a small business?
Most small businesses land between 1,500 and 8,000 dollars per hour, and companies with 50 to 200 employees typically fall between 8,000 and 35,000. The exact figure depends on your gross margin, headcount, and how dependent your work is on the systems that fail.
Micro businesses under 10 people usually come in between 250 and 1,500 dollars an hour, where the dominant cost is the owner’s time and missed inbound calls rather than idle payroll.
What counts as downtime?
Any period when the systems your business runs on are unavailable or degraded enough that work stops. That includes email outages, phone system failures, ERP lockouts, and network slowdowns severe enough that people give up.
It doesn’t need to be a total blackout. A file server that responds in 30 seconds instead of half a second stops work just as effectively as one that is offline, and most companies never log it.
Is the 5,600 dollars per minute figure still accurate?
No, and it never described small businesses. That number is attributed to Gartner in 2014, it measured large enterprises, and it is more than a decade old.
Newer research puts the enterprise average far higher, with Enterprise Management Associates reporting 14,056 dollars per minute in 2024 and Splunk reporting about 15,000 in 2026. Both are drawn from enterprise samples. If your company does under 100 million dollars in revenue, calculate your own figure instead of borrowing any of them.
How long does the average IT outage last?
Most incidents resolve in 1 to 4 hours, though the tail is what hurts. Ransomware and major hardware failures routinely run into days.
Sophos found 53 percent of ransomware victims fully recovered within a week in 2025, an improvement over prior years but still a full working week of degraded operations. IBM reported that organizations took an average of 241 days to identify and contain a breach, which is a different measurement but a useful reminder that the incident and the recovery are rarely the same length.
Does cyber insurance cover downtime losses?
Business interruption coverage exists on most cyber policies, but it rarely pays from minute one. Policies commonly include a waiting period before coverage begins and require documented evidence of the loss.
Read your policy for the waiting period, the sublimit on business interruption, and what proof the carrier expects. Having the calculation in this article already documented makes that claim substantially easier to file, which is a side benefit worth the 30 minutes.
What recovery time objective should a small business aim for?
Set it per system rather than company-wide. For most small businesses, 4 hours is a reasonable target for core systems like ERP and phones, and 24 hours is acceptable for secondary systems.
The right way to choose is to multiply your hourly downtime cost by candidate RTO values and compare that against what the technology to hit each target costs. If 4-hour recovery saves you 12,000 dollars per incident and costs 500 dollars a month, the decision makes itself.
Does managed IT actually reduce what downtime costs?
It reduces frequency and duration, which is where the money is. Continuous monitoring catches failures before they become outages, and a fast response time shortens every incident that does happen.
The honest caveat is that no provider eliminates downtime. Power fails, carriers drop circuits, and hardware dies. What a good provider changes is whether you find out at 2 a.m. from an alert or at 9 a.m. from an employee, and whether recovery takes 45 minutes or 2 days.
Find out what an hour of downtime costs your business
We’ll run the calculation with your numbers, review your last three outages, and show you which of the five levers above would have shortened them. No cost, no obligation, and you keep the spreadsheet either way. Uprite has supported Texas businesses for more than 25 years, we’re a seven-time Channel Futures MSP 501 honoree, and every agreement is backed by a 120-day satisfaction guarantee.
Talk to Uprite about managed IT or review our cybersecurity services if ransomware is the outage you’re most worried about.
