The cost of website downtime isn't a single number — it depends heavily on business size, industry, and how much of the impact you count beyond the immediate outage window. This post pulls together the most credible industry benchmarks available, explains what actually makes up "total cost," and shows you how to get a figure specific to your own business rather than relying on an industry average that may not apply to you.
What Do Industry Benchmarks Say Downtime Actually Costs?
Several research firms have measured downtime cost from different angles — large enterprises, data centers, and firms of every size — and the figures vary by an order of magnitude depending on who's being measured.
| Source | Finding | Who was measured |
|---|---|---|
| Splunk (Cisco), The Hidden Costs of Downtime, May 2026 | About $15,000 per minute on average; $600 billion a year across the Global 2000 | 2,000 executives at Global 2000 companies |
| Ponemon Institute and Emerson Network Power, 2016 | $8,851 per minute of unplanned outage, up from $5,617 in 2010 | 63 U.S. data centers that had an outage in the past year |
| ITIC, 2024 Hourly Cost of Downtime | Over $300,000 per hour (about $5,000 a minute) for more than 90% of mid-size and large firms; $1 million to $5 million+ per hour for 41% of enterprises | 1,000+ organizations worldwide |
| ITIC 2024, Part 2 | Worked example: a small business whose downtime costs $25,000 an hour loses about $417 per server, per minute | An illustrative calculation, not a survey average |
The spread is the important takeaway: the enterprise studies land in the thousands of dollars per minute, while ITIC's small-business example lands in the hundreds. If you run a small or mid-sized business, the enterprise headline numbers quoted in the press simply don't describe you, and using them will wildly overstate your real risk.
What Actually Makes Up the "Total Cost" of Downtime?
A lot of downtime cost coverage stops at "lost sales during the outage," but that's an incomplete picture. In Splunk's 2026 survey, 81% of technology leaders said downtime had cost them customers and 90% reported a jump in customer support demand — costs that never show up in the outage-window revenue figure.
Direct costs
These are the costs tied directly to the outage window itself:
- Lost sales or conversions that would have happened while the site was down
- Wasted advertising spend on traffic that arrived but couldn't convert
- SLA penalties or refunds owed to customers or partners, if applicable
Indirect costs
These extend beyond the outage window, and they're harder to measure — studies disagree on how large they are, but they're rarely zero:
- Recovery work — engineering time spent diagnosing and fixing the issue
- Customer support load from affected users
- Customer churn — some visitors who hit an error page simply don't come back, a particularly acute risk for consumer-facing sites
- Reputational damage that can affect future conversion rates even after the site is restored
In short: what does website downtime actually cost?
The cost of website downtime ranges enormously by business size — from hundreds of dollars a minute for a small business (ITIC's worked example puts a $25,000-an-hour outage at about $417 a minute) to roughly $9,000–$15,000 a minute for large enterprises, according to Ponemon (2016) and Splunk (2026). The revenue lost during the outage is only part of it: recovery work, customer support load, churn and reputational damage come on top.
Why Generic Benchmarks Aren't Enough for Your Business
Industry averages are useful for context, but they're built from a mix of businesses with wildly different traffic, average order values, and revenue models — an average across "large organizations" or even "SMBs" doesn't reflect your specific numbers. A high-traffic ecommerce store and a low-traffic B2B services site with the same revenue can have very different per-minute downtime costs, because the calculation depends on how tightly your revenue is tied to continuous site availability.
That's true even within the same size bracket. Two small businesses of the same size could have per-minute costs that differ by 10x depending on order volume and average transaction size — the formula below shows why. For ecommerce specifically, the dynamics are different enough to warrant their own look — see ecommerce downtime cost for why checkout-blocking outages hit differently than downtime on an informational site.
How Do You Calculate Your Own Downtime Cost Per Minute?
Industry benchmarks are a useful sanity check, but they're averages across many different business models. The more reliable approach is calculating your own rate directly from your revenue data.
- Start with your monthly revenue. Use your actual revenue figure from the site or channel in question — not total company revenue if the site is only one part of it.
- Divide by minutes in a month. A 30-day month has 43,200 minutes, so monthly revenue ÷ 43,200 gives you a naive average revenue-per-minute figure.
- Adjust for business hours, if relevant. If your revenue is concentrated during specific hours (a B2B tool used only during the workday, for example), divide by active minutes instead of all 43,200 — this gives a more realistic per-minute figure for outages that happen during those hours.
- Adjust for traffic patterns. An outage during a traffic peak (a promotion, a viral moment, your busiest sales day) costs more than the average minute; an outage at 3 a.m. on a slow day costs less. Weight your estimate toward the time of day the outage actually occurred if you're calculating after the fact.
- Add the indirect costs you can estimate. Recovery time (hours × your team's hourly cost), extra support tickets, and refunds or SLA credits sit on top of lost sales. Churn and reputational damage are real but hard to pin down — in Splunk's 2026 survey, 81% of technology leaders said downtime had cost them customers — so estimate them separately rather than applying a blanket multiplier.
| Approach | Formula | Best for |
|---|---|---|
| Simple average | Monthly revenue ÷ minutes in month | Quick, rough estimate |
| Business-hours adjusted | Monthly revenue ÷ active business minutes | B2B or non-24/7 revenue patterns |
| Traffic-weighted | Adjusted for the specific time period the outage occurred | After-the-fact incident cost analysis |
| Total-cost estimate | Direct revenue loss + recovery labor + support load + refunds/credits | Understanding full business impact, not just direct sales |
How Detection Speed Affects the Numbers Above
Every benchmark in this post assumes the outage runs its full course before anyone notices and responds — but in practice, how quickly you detect an outage has a direct effect on the total cost. A business checking uptime manually or relying on customer complaints might not notice a failure for thirty minutes or more; a business running automated monitoring with SMS or phone alerts might catch it within one to five minutes. Multiply that gap by any of the per-minute figures above, and the value of fast detection becomes obvious. Website uptime monitoring tools compared walks through the free and paid options if this isn't something you have in place yet.
This is also why the "total cost" framing matters more than the headline per-minute number. A business that detects and resolves an outage quickly limits both the direct revenue loss and the indirect costs — less time for customers to hit error pages, less recovery work, less reputational exposure. The benchmarks above describe the raw cost of downtime; how much of that cost you actually incur is partly a function of how fast you respond. Responding well includes keeping customers informed, and our copy-paste outage messages for customers cover each stage from the first notice to the all-clear.
Getting Your Own Number
Rather than relying on an industry average that may be 10x or more off from your actual exposure, the fastest way to get a figure that means something is to calculate it from your own traffic and revenue data. The Website Downtime Cost Calculator does this in under a minute — plug in your monthly traffic and revenue, and get a specific per-minute and per-hour estimate instead of a generic industry figure.
